# Thinking Clearly in a Noisy World

Before a consequential choice, make five things explicit: what you want, which alternatives are feasible, what you believe and how uncertain it is, which evidence could change the choice, and when you will review it. Then protect the decision from predictable errors—in your own reactions, other people’s judgments, the data, and the system you are trying to change.

A good process can still produce a bad outcome. A precise estimate can still be biased. A useful rule can fail when its assumptions change. The practical aim of these twelve books is to make those distinctions usable: deciding when to trust experience, when to measure, when to explore, and when to reconsider.

This article integrates the complete local Shortform one-page summaries of twelve books, with their examples and supplementary commentary. It does not claim to cover the full original books. Author arguments, Shortform comparisons and editorial corrections are distinguished in the text; a supplementary exercise is not automatically a tested intervention. The sections below keep each framework together and put its limits beside the mechanism they qualify.

- [When the mind takes a shortcut](#theme-fast-minds-bias-and-emotion)
- [Uncertainty, belief and learning](#theme-uncertainty-probability-and-calibration)
- [Measurement and causal evidence](#theme-evidence-measurement-and-causality)
- [A complete decision process](#theme-decision-process-and-options)
- [Noise in organizations and groups](#theme-noise-judgment-and-groups)
- [Systems, feedback and intervention](#theme-systems-feedback-and-leverage)
- [Mental models across domains](#theme-mental-models-and-second-order-effects)
- [Algorithms under real constraints](#theme-algorithms-and-bounded-optimization)
- [How to judge scientific claims](#theme-science-incentives-and-reliability)
- [Living and planning with uncertainty](#theme-durable-judgment)

<a id="theme-fast-minds-bias-and-emotion"></a>

## When the mind takes a shortcut

<!-- atlas:block-content b-380ddee383979c1e:start -->
### When should you interrupt intuition?

Fast responses are indispensable in familiar situations. The question is whether the pattern is reliable here, whether feedback has trained it, and what an error would cost.

<a id="b-380ddee383979c1e"></a>
<a id="u-thinking-fast-and-slow-38cf94724d24"></a>
<a id="u-thinking-fast-and-slow-2bfc6a9c4a78"></a>
<a id="u-thinking-fast-and-slow-41b7e828e45f"></a>
<a id="u-thinking-fast-and-slow-516142053d8d"></a>
<a id="u-thinking-fast-and-slow-2ccf174e479d"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-71eacdfc7409"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-a2f52c7bce10"></a>
<a id="u-thinking-fast-and-slow-ce1d316c2269"></a>
<a id="u-thinking-fast-and-slow-2adc5eebdaf5"></a>
<a id="u-thinking-fast-and-slow-e19f41a75bdf"></a>
<a id="u-thinking-fast-and-slow-477278e5507d"></a>
<a id="u-thinking-fast-and-slow-b6b51dbdc344"></a>
<a id="u-thinking-fast-and-slow-38786cfe5ba6"></a>
<a id="u-thinking-fast-and-slow-977108fcf6bf"></a>
<a id="u-thinking-fast-and-slow-7d85fea7f02a"></a>
<a id="u-thinking-fast-and-slow-c1e1d308c667"></a>
<a id="u-thinking-fast-and-slow-2b35df22e411"></a>
<a id="u-thinking-fast-and-slow-5932b16bcf55"></a>
<a id="u-thinking-fast-and-slow-05039a8c9fcb"></a>
<a id="u-thinking-fast-and-slow-4add3e31a6f9"></a>
<a id="u-thinking-fast-and-slow-cfc4694b802b"></a>
<a id="u-thinking-fast-and-slow-826065610ab5"></a>
<a id="u-thinking-fast-and-slow-4670ffb7c7a3"></a>
<a id="u-thinking-fast-and-slow-53214a271623"></a>
<a id="u-thinking-fast-and-slow-6e283949c7b2"></a>
<a id="u-thinking-fast-and-slow-abd699220f7e"></a>
<a id="u-thinking-fast-and-slow-c49ce38cccd1"></a>
<a id="u-seeking-wisdom-bevelin-1f807c92a600"></a>
<a id="u-seeking-wisdom-bevelin-7b86ed0d9fe2"></a>
<a id="u-seeking-wisdom-bevelin-2a7c96ee1876"></a>
<a id="u-seeking-wisdom-bevelin-bf5436dea31c"></a>
<a id="u-seeking-wisdom-bevelin-28039a0dfea2"></a>
<a id="u-seeking-wisdom-bevelin-0f075f01c5e4"></a>
<a id="u-seeking-wisdom-bevelin-d6f819b72faf"></a>
<a id="u-clear-thinking-88ebcd5da4c2"></a>
<a id="u-clear-thinking-03d92258a6a7"></a>
<a id="u-thinking-fast-and-slow-8744a321ef1d"></a>
<a id="u-thinking-fast-and-slow-6ee4d6fff0ce"></a>
<a id="u-thinking-fast-and-slow-4a5657c92115"></a>
<a id="u-thinking-fast-and-slow-64cd958e9059"></a>
<a id="u-thinking-fast-and-slow-f23ce6c6a7b4"></a>
<a id="u-thinking-fast-and-slow-58c119856a50"></a>
<a id="u-thinking-fast-and-slow-dd44cdd114f4"></a>
<a id="u-thinking-fast-and-slow-b59bce0f9f61"></a>

<a id="v2-0-0"></a>

**System 1: The Fast, Intuitive Thinking System**

Daniel Kahneman’s dual-system model of thought is the core framework for understanding fast and slow thinking, and it structures the full arc of his research in *Thinking, Fast and Slow*. **System 1** (fast thinking) operates automatically and quickly, with little or no conscious effort and no sense of voluntary control; it generates feelings and intuitions, and suggests courses of action for the conscious mind to consider. **System 2** (slow thinking), by contrast, involves careful, deliberate reasoning and is cognitively taxing. Weinberg and McCann draw directly on this Kahneman model in their discussion of inappropriate intuition in *Super Thinking: The Big Book of Mental Models*.

System 1 produces rapid impressions and judgments through association: it immediately connects what you see or experience to related concepts and patterns stored in memory. For example, seeing someone frown instantly links that facial expression to associated concepts like anger, threat, and negativity, priming you within milliseconds to defend yourself, escape, or defuse the situation. Other everyday examples of System 1 thinking include judging that one object is farther away than another, understanding simple sentences, detecting emotion in someone’s voice, and performing simple or habitual actions such as driving a car on an empty road.

While fast, intuitive thinking is essential for navigating daily life—you cannot consciously deliberate over every sight, sound, or experience you encounter—Kahneman emphasizes that System 1 is impulsive and imprecise, and it regularly produces errors. This creates a practical conundrum: how can people make sound decisions when their default thinking is prone to error and their rational system is effortful to engage? Kahneman’s proposed solution is to learn to recognize situations in which System 1 is especially vulnerable to mistakes, so that deliberate System 2 reasoning can be brought to bear. To that end, he reviews a range of common heuristics—mental shortcuts—and cognitive biases—thinking errors that those shortcuts can produce—that routinely mislead intuitive thinking.

Kahneman’s broader research program, organized around the two systems, covers the regular biases and sloppiness that arise in human thought, prospect theory’s account of reference-dependent gains, losses, and decision weights, and an investigation of happiness and how self-understanding supports well-being.

<a id="v2-0-4"></a>

**System 2: Effortful, Deliberate Thinking**

System 2 is the mode of thought that allocates attention to effortful mental activities that demand it. Common tasks that require System 2 include sustaining focus, making difficult choices, and performing complex calculations. It also converts intuitions into beliefs and impulses into voluntary actions. These are functional descriptions of thinking, not a literal serial route through two brain compartments. Deliberate attention can check a fast answer or be engaged intentionally before acting. For example, driving on an empty road is typically a System 1 task — simple and familiar enough to proceed without conscious thought — but parallel parking in a tight space triggers System 2 to make careful adjustments and avoid surrounding cars.

<a id="v2-0-5"></a>

**The Laziness of System 2 and Mental Resource Conservation**

Daniel Kahneman writes that the brain defaults to System 1 not only because it is faster, but also because it consumes less mental energy, and the brain instinctively tries to conserve mental resources wherever possible. In short, System 2 is lazy: it prefers to let System 1 handle as much thinking as possible. As a result, people often do not question the impulses and intuitions generated by System 1, because rethinking those decisions would expend energy the brain would rather save. This problem worsens when you are tired or stressed, since limited mental resources are already strained, making it even harder to devote conscious effort to solving problems and making decisions.

<a id="b-c46b71e388d73340"></a>
<a id="u-thinking-fast-and-slow-510f0d3d7fa9"></a>

<a id="v2-2-0"></a>

**Daniel Kahneman and the Foundations of Behavioral Decision Theory**

Daniel Kahneman (1934–2024) was a psychologist specializing in human decision-making. His pioneering work on cognitive biases, heuristics, and behavioral economics earned him the 2002 Nobel Memorial Prize in Economic Sciences, making him one of the few non-economists ever to receive the honor. He is best known for co-developing prospect theory with fellow psychologist Amos Tversky.

<a id="b-762a46ee82ce58d1"></a>
<a id="u-thinking-fast-and-slow-30e0dacffec1"></a>

<a id="v2-7-0"></a>

**What the heuristic research tested**

Michael Lewis’s The Undoing Project describes Kahneman and Tversky’s collaboration beginning in the late 1960s. Their misleadingly framed questions explored not simply whether people erred, but which regularities explained the errors. The “law of small numbers”—expecting a small sample to mirror a population closely—and experts’ reliance on stereotypes became important findings. They helped establish the heuristics-and-biases research program; they did not invent every earlier use of the term heuristic.

<a id="b-3fbfb5bd88c673ea"></a>
<a id="u-thinking-fast-and-slow-a48ac927410c"></a>
<a id="u-thinking-fast-and-slow-86bfa3e119ed"></a>
<a id="u-thinking-fast-and-slow-a5a1ec6bdcb1"></a>
<a id="u-thinking-fast-and-slow-9ca02fa8f78b"></a>
<a id="u-thinking-fast-and-slow-b624041b8b67"></a>

<a id="v2-3-0"></a>

**Heuristics: How System 1 Substitutes Easy Questions for Hard Ones**

Kahneman defines heuristics as mental tools through which System 1, when confronted with a complex question, substitutes an easier question and answers that instead. The simple answers System 1 generates are usually imperfect but adequate for most everyday situations. In high-stakes complex situations, however, these simplified answers can be seriously misleading. For instance, a low-stakes question like “What should I order for dinner?” would ideally require weighing many factors—cost, nutrition, personal preference, and more—but people typically replace it with the much simpler heuristic question, “What do I feel like eating?” This relies only on an emotional impulse rather than multiple considerations, and it works well enough for routine choices. By contrast, a high-stakes question such as “Is this company worth investing in?” should be answered by logically evaluating the company’s likelihood of success and thus its stock’s potential to rise. Yet System 1 may substitute a heuristic question like “Do I like this company?” Here, relying on feeling rather than an honest assessment of performance risks substantial financial loss. System 2 has the capacity to recognize and reject heuristic answers, but it often endorses them without further scrutiny. Kahneman therefore urges people to verify, before making important decisions, that they actually answered the question they think they answered—that is, to check whether the conclusion came from deliberate, rational thought or from simplistic, impulsive shortcuts.

<a id="b-a764b8773f592c35"></a>
<a id="u-thinking-fast-and-slow-d2643e5e1680"></a>
<a id="u-thinking-fast-and-slow-af1b5a7d3ac4"></a>
<a id="u-thinking-fast-and-slow-c896e613a644"></a>

<a id="v2-4-0"></a>

**The Fallibility of Fast, Associative Thinking**

Kahneman illustrates System 1's fallibility with several concrete examples. In one social scenario, seeing someone frown can trigger an immediate System 1 assumption that the person is angry and that you are in danger, even though the frown could have many unrelated causes—for instance, they might simply be thinking through a difficult problem. Trick questions provide another sharp demonstration: when asked how many of each type of animal Moses took on the ark according to the Old Testament, most people immediately answer “two.” The correct answer is zero, because it was Noah, not Moses, who took animals onto the ark. System 1's associative processing picks up on familiar words like “animal” and “ark” and answers the question it expects to hear, rather than the actual question being asked.

<a id="b-1d9143b86b8f2fda"></a>
<a id="u-thinking-fast-and-slow-ff5714dc7bb7"></a>
<a id="u-thinking-fast-and-slow-5276f6144131"></a>

<a id="v2-5-0"></a>

**Shortform perspective: System 1, Intuition, and Associative Thinking**

System 1 thinking corresponds to what Malcolm Gladwell calls unconscious thinking in *Blink*—better known as intuition or a “gut feeling.” Gladwell argues that unconscious thinking is often just as effective as conscious, rational thinking, with the added advantages of being faster and less likely to be derailed by stress or anxiety. He explains that the subconscious mind naturally takes in far more information than you consciously process, filters out what is irrelevant, and returns the best answer to your current problem.

The associative thinking Kahneman describes as characteristic of System 1 is not only fast but also creative. In *The Innovator’s DNA*, Hal Gregersen, Jeff Dyer, and Clayton Christensen argue that associative thinking is the very foundation of creativity. They explain that creative ideas and innovations are rarely completely new concepts; instead, they typically emerge when people connect existing concepts in ways no one has connected them before.

<a id="b-1859b8602fb79f2d"></a>
<a id="u-thinking-fast-and-slow-73f02c3a7edf"></a>
<a id="u-thinking-fast-and-slow-fab8dc7603b2"></a>

<a id="v2-6-0"></a>

**Shortform perspective: Mental Fatigue—and When Diffuse Thinking Helps**

Shortform draws on Benjamin Hardy’s Willpower Doesn’t Work to describe how repeated decisions, temptations, stress and fatigue can make the easy default more attractive. The stronger claim that every decision necessarily depletes a single finite reserve is a contested account, not a settled biological mechanism. Treat workload, sleep and context as things to manage, without assuming a fixed daily decision budget.

Barbara Oakley’s A Mind for Numbers distinguishes focused work from a more diffuse mode that may help connect ideas after a break. Shortform compares these modes with Kahneman’s two systems. The comparison is an analogy, not an identity between categories or proof about longer neural routes. The practical suggestion remains: when sustained concentration stalls, take a break and return to test the ideas that emerge.

<a id="b-36c23cca7364bd8b"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-6c158b8d6054"></a>

<a id="v2-38-0"></a>

**When Intuition Fails: The Case for Slow Thinking in Unfamiliar Contexts**

Weinberg and McCann argue that we should avoid relying on intuition in situations that call for slow thinking. They illustrate this with a cross-cultural example: if you are an American visiting a less talkative culture and none of the locals seem willing to talk with you, leaning on intuition could lead you to incorrectly conclude that they are impolite or rude, when different social norms are actually at work. In such cases, slow thinking should be used to consider those unfamiliar norms instead of making a snap judgment.

<a id="b-cf216456c75eedd1"></a>
<a id="u-thinking-fast-and-slow-9738083c8bb0"></a>
<a id="u-seeking-wisdom-bevelin-02c8af0df7b8"></a>
<a id="u-seeking-wisdom-bevelin-c6084a995a86"></a>
<a id="u-clear-thinking-4351e089653d"></a>
<a id="u-clear-thinking-a17868eaed3b"></a>
<a id="u-seeking-wisdom-bevelin-b7bfe8def1f9"></a>
<a id="u-seeking-wisdom-bevelin-29fd27fa49be"></a>
<a id="b-2f887de9e881c432"></a>
<a id="u-thinking-fast-and-slow-f7d94a387dbe"></a>
<a id="u-seeking-wisdom-bevelin-3b0559fb1386"></a>

<a id="v2-41-0"></a>

**Different authors place the correction at different points**

Kahneman’s 2011 Thinking, Fast and Slow examines systematic patterns in judgment. Bevelin, a Swedish investor drawing on mathematics, science, business and philosophy, emphasizes checking assumptions against facts. Parrish stresses habits and procedures that still work when self-control falters. Their practical overlap is fallibility plus correction; it does not require treating emotion as the opposite of all rational choice.

*Sources and further reading:* [Thinking, Fast and Slow](https://short.zaynjarvis.com/books/thinking-fast-and-slow/) · [Daniel Kahneman](https://kahneman.scholar.princeton.edu/) · [Prospect theory](https://thedecisionlab.com/reference-guide/economics/prospect-theory) · [The Undoing Project](https://www.shortform.com/app/book/the-undoing-project) · [Kahneman–Tversky collaboration](https://www.shortform.com/app/book/the-undoing-project#the-collaboration-begins-1969-1973) · [Law of small numbers](http://stats.org.uk/statistical-inference/TverskyKahneman1971.pdf) · [Blink](https://www.shortform.com/app/book/blink) · [Unconscious thinking](https://www.shortform.com/app/book/blink/introduction) · [The Innovator’s DNA](https://www.shortform.com/app/book/the-innovators-dna) · [Associative thinking](https://www.shortform.com/app/book/the-innovators-dna#associative-thinking) · [Willpower Doesn’t Work](https://www.shortform.com/app/book/willpower-doesnt-work/1-page-summary) · [Decision demands](https://www.shortform.com/app/book/willpower-doesnt-work/1-page-summary#the-problem-willpower-is-weaker-than-you-think) · [A Mind for Numbers](https://www.shortform.com/app/book/a-mind-for-numbers) · [Focused and diffuse thinking](https://www.shortform.com/app/book/a-mind-for-numbers/part-1#the-two-basic-modes-of-thinking)

<!-- atlas:block-content b-380ddee383979c1e:end -->
<!-- atlas:block-content b-3ee07485ae5bfe0e:start -->
### Recognize the shortcut before it becomes a conclusion

<a id="b-3ee07485ae5bfe0e"></a>
<a id="u-decisive-140ff6d242df"></a>
<a id="u-decisive-74bad68945d3"></a>
<a id="u-decisive-861e5f49d81d"></a>
<a id="u-thinking-fast-and-slow-ad307f9b7ef9"></a>
<a id="u-thinking-fast-and-slow-bf7b2516c596"></a>
<a id="u-thinking-fast-and-slow-906dab9e722f"></a>
<a id="u-thinking-fast-and-slow-338577b36541"></a>
<a id="u-thinking-fast-and-slow-46d9687c11b7"></a>
<a id="u-thinking-fast-and-slow-588b0bdbb67f"></a>
<a id="u-thinking-fast-and-slow-ecbd6c23678a"></a>
<a id="u-thinking-fast-and-slow-0c85589015b5"></a>
<a id="u-thinking-fast-and-slow-aa92fb772bb4"></a>

<a id="v2-9-0"></a>

**Confirmation Bias and Other Common System 1-Driven Cognitive Biases**

Kahneman classifies a range of cognitive biases and fallacies as products of System 1 thinking, each of which can skew decisions by distorting how we process information—even when we believe we are reasoning rationally.

**Confirmation bias** is the tendency to selectively seek, favor, and interpret information that reinforces existing beliefs, while ignoring or devaluing evidence that contradicts them. In *Decisive*, the Heath brothers identify it as one of the core bad habits that undermine good decision-making: once we have identified our options, confirmation bias prevents us from evaluating them accurately, causing us to overlook information that supports non-preferred options even when those options are objectively better. They illustrate the bias with a neighborhood-choice example: a man who heard a vivid break-in story about Neighborhood A selectively researches safety information that confirms his hunch that A is unsafe, never looks for comparable data on Neighborhood B, and chooses B even though B has much higher crime rates. Kahneman adds that people not only discard disconfirming information but also interpret ambiguous evidence in ways that align with their prior views: for instance, people with conservative and progressive views can look at the same crime data and each interpret it as supporting their opposing positions on gun control.

**Availability bias** leads people to overestimate the frequency of events that are fresh in their working memory and to underestimate statistically more common events they have not thought about lately. For example, city dwellers who frequently hear news stories about crime involving homeless people may feel in great danger every time they leave home, even though such crime is statistically unlikely to affect them, while they largely ignore the more common risk of being hit by a delivery bike.

**Anchoring bias** occurs when an initial piece of information—even one irrelevant to the judgment—pulls later estimates toward it. Kahneman describes respondents being asked how much they would be willing to contribute annually: a $400 anchor was associated with an average stated willingness of $143, while a $5 anchor was associated with $20. These are reported donation intentions, not amounts a nonprofit actually received. See [Kahneman’s anchoring chapter](https://faculty.bard.edu/hhaggard/teaching/sci127Sp20/notes/KahnemanThinkingExcerpt2.pdf).

**Narrative fallacy** is the tendency to weave coherent stories to explain random events, which then breeds unjustified confidence in predicting the future. For instance, sports pundits invent post-hoc stories about why one boxer won and another lost—say, that the loser buckled under pressure or the winner wanted it more—even when the fighters are so evenly matched that the outcome could have gone either way, and they then rely on those stories to predict rematches that are just as unpredictable.

**Narrow framing** is the tendency to make decisions based on a relatively small pool of information rather than considering the whole picture. It appears in many forms: in the **planning fallacy**, people overlook all the ways a project could go wrong and therefore underestimate how much time it will take, factoring in only the timeline for an ideal scenario; in the **sunk cost fallacy**, people narrow their focus to recovering what they have already lost on a failed endeavor, when stepping back to consider all options would often show it is better to cut losses and redirect resources elsewhere.

<a id="b-8b429076a3a0432c"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-3a965c5e47f3"></a>

<a id="v2-8-0"></a>

**Pattern perception and the gambler’s fallacy**

Intuition can detect patterns where none exist. In an independent sequence of roulette spins, several blacks do not make red due on the next spin: treating a compensating outcome as more likely is the gambler’s fallacy. It differs from assuming a streak will continue. Shortform also cites research in which people retained intuitive probabilistic beliefs they knew were incorrect; knowing a rule and having an intuition that follows it are different achievements.

<a id="b-5331398e9793e63a"></a>
<a id="u-decisive-98019269b2ad"></a>
<a id="u-thinking-fast-and-slow-0a899a3a706e"></a>
<a id="u-thinking-fast-and-slow-5619f3ae9b67"></a>
<a id="u-thinking-fast-and-slow-f6cb61170b30"></a>
<a id="u-decisive-2e59d5b86aeb"></a>
<a id="u-thinking-fast-and-slow-cb7dcf95d410"></a>

<a id="v2-10-0"></a>

**Shortform perspective: Why confirmation can feel useful**

Shortform offers three explanations: considering congenial evidence saves effort; argumentative-theory accounts propose that people defending different positions can expose flaws through genuine debate; and beliefs can become tied to identity, so disagreement feels like a personal threat. The group-benefit account is conditional on real challenge and responsiveness. A group that shares a bias or suppresses criticism does not gain accuracy merely because its members are confident.

<a id="b-5781d32fc9a3baec"></a>
<a id="u-thinking-in-bets-82cdc7e4d782"></a>
<a id="u-thinking-in-bets-01b406ddda99"></a>
<a id="u-thinking-in-bets-a64e2f95eb94"></a>

<a id="v2-46-0"></a>

**Beliefs as the foundation of decision-making**

All decisions ultimately flow from beliefs—whether about which job will be more fulfilling, whether city life will suit you better than a small town, whether cardio benefits the heart, or whether a relationship will last. Some beliefs are grounded in direct experience, but false beliefs often take hold and persist without being noticed. Regularly interrogating one’s beliefs is therefore essential to good decision-making.

A series of experiments by a Harvard psychology professor found that people tend to process information as true even when it is explicitly presented as false, especially under time pressure. Unless people are prompted to question their beliefs, vet them, and seek out factual evidence, incorrect beliefs stick—and in some cases they persist even after being debunked.

<a id="b-b4ae4b896c1452b7"></a>
<a id="u-thinking-in-bets-da30e065f60d"></a>
<a id="u-thinking-in-bets-99965bf25ea8"></a>

<a id="v2-47-0"></a>

**Motivated reasoning: how beliefs reinforce themselves**

Once established, old beliefs shape how new information is perceived, reinforcing existing biases and amplifying them in a cycle Duke calls “motivated reasoning.” This is why fake news is so insidious and effective: it rarely converts people to entirely new views, but instead targets and reinforces beliefs they already hold.

The same dynamic appears with recurring viral claims that there is no scientific proof flossing is necessary. Many people share and believe these articles without looking past the headlines, because they dislike flossing and are biased toward anything that reduces their guilt about skipping it. Dentists continue to recommend daily flossing, and the “no scientific support” claims come with significant caveats, yet the articles spread because they reinforce existing feelings and beliefs rather than offering genuine new insight.

<a id="b-bfe33d8b3c422eb5"></a>
<a id="u-thinking-in-bets-7603c5336dc8"></a>
<a id="u-thinking-in-bets-93d478501251"></a>
<a id="u-thinking-in-bets-618aeb5f78ce"></a>
<a id="u-thinking-in-bets-c78e4bff31d0"></a>
<a id="u-thinking-in-bets-3f0e301ebc33"></a>
<a id="u-how-to-measure-anything-39af92782cb7"></a>
<a id="u-how-to-measure-anything-1cc9b12b6cab"></a>
<a id="u-how-to-measure-anything-6d4e19c39a06"></a>
<a id="u-how-to-measure-anything-7c0436c027b8"></a>
<a id="u-how-to-measure-anything-efc7ad77032f"></a>
<a id="u-how-to-measure-anything-f757c7f8f58e"></a>
<a id="u-how-to-measure-anything-5cd689f4ef75"></a>
<a id="u-how-to-measure-anything-0353f75352cc"></a>
<a id="u-how-to-measure-anything-7f28d7dfc55c"></a>

<a id="v2-48-0"></a>

**Two practices for vetting your beliefs**

To make good decisions, people must counter motivated reasoning and ground their beliefs in objectivity. Duke offers two practical strategies for vetting beliefs.

First, imagine you had to place a bet on every belief you hold. If being wrong carried an immediate financial penalty, you would be far more likely to examine the belief, second-guess yourself, and fact-check it. Most misconceptions do not carry a direct cash cost, but many still impose a penalty in the form of forgone futures that better decisions might have produced.

Second, acknowledge that most beliefs carry some degree of uncertainty. The more closely you scrutinize your beliefs, the more you will find there are few things you would confidently bet money on—and that is fine. Even scientific findings are constantly updated or overturned. You do not need to be fully certain of every belief, but it is useful to think about how sure you are; every belief sits somewhere on a spectrum of certainty and confidence.

<a id="b-47e57f13f4fa8328"></a>
<a id="u-seeking-wisdom-bevelin-63757675a79d"></a>
<a id="u-seeking-wisdom-bevelin-351ac1f49d80"></a>
<a id="u-seeking-wisdom-bevelin-4120db32f9a5"></a>
<a id="u-seeking-wisdom-bevelin-4d1b32ea73cc"></a>
<a id="b-d68c965aabcf1e25"></a>
<a id="u-seeking-wisdom-bevelin-34ca1776baae"></a>
<a id="u-seeking-wisdom-bevelin-b0281654b4b0"></a>

<a id="v2-12-0"></a>

**When an association substitutes for evidence**

Bevelin calls attention to judging before the relevant facts are known. A high car price can evoke quality, causing a buyer to skip research and later discover an overheating engine or a failing retractable roof. Tattoos can evoke rebellion, causing a recruiter to overlook a candidate’s relevant skills and interpersonal strengths. The error is substituting an association for evidence about the feature actually being judged.

<a id="b-da2a195edc70414e"></a>
<a id="u-seeking-wisdom-bevelin-b0f05e564307"></a>
<a id="u-thinking-fast-and-slow-daeffc730c83"></a>

<a id="v2-1-0"></a>

**Correcting a First Impression**

Kahneman’s two-system framework becomes practical at the moment you distrust a first impression. System 1 has already supplied the intuitive judgment or feeling; the corrective move is to give System 2 enough time and structure to inspect the assumptions beneath it. Backward thinking is one such technique: instead of merely accepting the first response, work back through what would have to be true for it to hold.

One way to mitigate the focusing illusion—the tendency for a single aspect of a situation to dominate your evaluation—is to reconsider the same issue across different moods or frames of mind. Apocryphally, ancient Persians deliberated over important decisions twice: once while drunk and once while sober, to shift their focus and consider different aspects of a situation; supposedly, only a decision that seemed wise in both states was acceptable. While getting drunk for every decision is not advisable, there is value in revisiting a choice later, especially one made in an emotional moment, because you may reach a different judgment once your feelings have settled and your focus has shifted.

<a id="b-1415503b54c6fa10"></a>
<a id="u-seeking-wisdom-bevelin-fd00adb35eca"></a>
<a id="u-seeking-wisdom-bevelin-b1ce969895b0"></a>
<a id="u-seeking-wisdom-bevelin-97ae9c72cbce"></a>
<a id="u-seeking-wisdom-bevelin-f34f90670048"></a>
<a id="u-seeking-wisdom-bevelin-42dc3b9cc057"></a>
<a id="u-clear-thinking-fe7122092225"></a>
<a id="u-seeking-wisdom-bevelin-a8de679ea3be"></a>
<a id="u-seeking-wisdom-bevelin-333a50386eb5"></a>
<a id="u-seeking-wisdom-bevelin-ad9e505c4efc"></a>
<a id="u-seeking-wisdom-bevelin-2362b5216c0f"></a>
<a id="u-clear-thinking-a8bd5401eb5f"></a>
<a id="u-clear-thinking-f5bba5948e85"></a>

<a id="v2-23-1"></a>

**Countering Hasty Judgments: Question First Impressions and Practice in Low-Stakes Settings**

Bevelin argues that you can counteract the tendency to jump to conclusions by questioning your first impressions, which prevents those often incomplete or wrong initial impressions from guiding your conclusions and decisions. He attributes a specific strategy for doing so to Charles Munger: a rational thinking technique called “backward thinking,” which prompts you to actively search for information that discredits your first impressions and pushes you to base conclusions on factual evidence rather than biased initial reactions. Applied to the tattooed job candidate, backward thinking would mean looking for evidence that undermines the assumption that the candidate is rebellious—for example, reviewing their resume for signs of responsibility and asking references to describe their personality traits. Parrish similarly recommends building expertise through low-stakes practice before facing high-stakes decisions, so that costly errors happen during the learning phase rather than when consequences matter most. For example, if you are asked to deliver a major keynote at a large conference, you could practice the presentation for other teams in your organization and collect their feedback first.

<a id="v2-10-1"></a>

**Shortform perspective: Availability, Salience, Negativity, and Anchoring Biases**

Availability bias is closely linked to the **salience bias**, in which we notice things that stand out because of their vividness or emotional impact, and both correlate with the **negativity bias**, our tendency to pay more attention to negative information than positive. Together, these three biases create a self-reinforcing loop: salience and negativity make shocking events stand out and get talked about, which makes them more available in memory, which in turn makes us focus on and discuss them even more. This loop explains why people often fear rare but dramatic events like shark attacks far more than common but less vivid risks like car accidents.

Anchoring bias may arise in part from **psychological priming**: exposure to an initial idea activates related parts of the brain, which stay active as we process further information and anchor our thinking to that first idea and its associated mental connections. To counter anchoring bias, Shortform advises actively generating counterarguments or alternative options and looking for reasons why they might be better than the anchored information.

<a id="b-c47bcc1ba14b0b54"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-a11f660c51f5"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-26d5e22d59e1"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-5a2a20323a2f"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-ac0fd8df8f9b"></a>

<a id="v2-182-0"></a>

**Frames of Reference and Availability Bias**

False assumptions are not only a risk when reasoning from first principles—we are also vulnerable to false assumptions rooted in our own personal perspective. To reason objectively, we must keep our own frame of reference and the biases that distort it in mind.

Frames of reference were used in physics before Einstein; relativity changed our understanding of their relationships. In an airplane you can be stationary relative to the cabin while moving north at 550 mph relative to the airport. Weinberg and McCann use a personal frame of reference as an analogy: experience shapes what seems salient or normal, so another observer can interpret the same event differently.

Availability bias gives vivid or easily recalled information disproportionate influence; recency is one contributor, not the whole definition. A dramatic shark-attack report can make that danger feel much more frequent than it is. Estimate risk from an appropriately scoped rate—place, period, exposure and outcome—not from a memorable story or an unscoped fatality number.

<a id="b-6ec4935978b62d5c"></a>
<a id="u-same-as-ever-e345279acee0"></a>

<a id="v2-339-0"></a>

**Shortform perspective: Availability Bias and the Difficulty of Predicting Risk**

Shortform explains that one reason risk is hard to predict is the availability bias: the tendency to make decisions based on memorable experiences. For example, after living through a stock market crash, an investor might avoid stocks because she believes another crash is imminent, while overlooking less vivid risks such as unexpected medical expenses. This can leave her unprepared for the true range of possible setbacks. To counteract availability bias and prepare for the unexpected, complement your own information by reading widely and consulting subject experts or people with relevant life experience, use statistical data to inform your conclusions, and make sure to question your assumptions.

<a id="b-3ecfff8dbb0b6524"></a>
<a id="u-thinking-fast-and-slow-ecd66d24e848"></a>

<a id="v2-14-0"></a>

**Shortform perspective: The Narrative Fallacy and the Desire for Predictability**

The narrative fallacy stems from the deep human desire to understand, control, and predict the world. Psychologists argue that conspiracy theories—extreme forms of the narrative fallacy that connect unrelated events—function as self-soothing anxiety responses: if a secret group such as the Illuminati is supposedly orchestrating events, then nothing is truly random, future disasters can in principle be predicted and prepared for, and people have a tangible enemy to focus on rather than feeling helpless before random chance.

<a id="b-22f8565c992586cf"></a>
<a id="u-thinking-fast-and-slow-c4bfe818708a"></a>

<a id="v2-15-0"></a>

**Narrow Framing and Working Memory Limits**

Narrow framing may be an unavoidable consequence of the natural limitations of working memory—the mental workspace where the brain holds ideas currently being used to solve a problem or make a decision. Working memory can only hold a small number of pieces of information at once. Researchers disagree on the exact capacity, and it varies between individuals, but common estimates place the average at somewhere between two and four ideas. As a result, whether a person is using System 1 or System 2 thinking, it is simply not possible to consider everything when making a decision.

<a id="b-dbeb2dd4f673d824"></a>
<a id="u-decisive-1a87b328f6b4"></a>
<a id="u-decisive-980a9feaa888"></a>
<a id="u-decisive-677589947f11"></a>
<a id="u-decisive-ecc311a599df"></a>
<a id="u-decisive-1b42d81f517a"></a>
<a id="u-decisive-190065c43f77"></a>

<a id="v2-13-0"></a>

**Status Quo Bias and the Pull of Familiarity**

In *Decisive*, the Heath brothers identify **status quo bias** as their third bad decision habit and explain that it weakens our judgment when we face difficult choices. Tough decisions stir intense emotions that drive us to seek relief in what is familiar, leading us to resist change even when change is needed. Status quo bias itself rests on two underlying biases: the **mere exposure effect**, which is a preference for things and situations we are accustomed to, and loss aversion, which is a bias against choices that involve loss. Together, these two biases make us fearful of losing what is familiar and what we currently have, so we avoid making significant life changes.

To overcome status quo bias, the Heath brothers argue that we should not dismiss the present emotions that draw us toward the comfort of the status quo. Instead, we should put those emotions into perspective by looking outside ourselves and beyond the present moment.

<a id="v2-10-2"></a>

**Shortform perspective: Status Quo Bias, Certainty, and the Psychology of Choice**

Researchers have identified two unpleasant emotions that drive status quo bias: the discomfort of having too many options and the agony of a tough dilemma. Maintaining the current state of affairs allows us to avoid both feelings.

Uncertainty can consume attention and intensify threat monitoring, helping explain why a familiar option feels relieving. But feeling certain does not establish knowledge or guarantee better judgment. The useful response is to identify what remains uncertain and manage its consequences, not to eliminate doubt by prematurely declaring an answer settled.

*Sources and further reading:* [Pattern perception](https://www.apa.org/monitor/mar05/misfires) · [Gambler’s fallacy](https://thedecisionlab.com/biases/gamblers-fallacy) · [Confirmation bias](https://thedecisionlab.com/biases/confirmation-bias) · [Argumentation and group reasoning](https://www.journals.uchicago.edu/doi/10.1093/bjps/axy079) · [Awaken the Giant Within](https://www.shortform.com/app/book/awaken-the-giant-within) · [Beliefs and identity](https://www.shortform.com/app/book/awaken-the-giant-within/chapter-4#recognize-the-three-categories-of-beliefs) · [Two-system model](https://short.zaynjarvis.com/books/thinking-fast-and-slow/#two-systems-of-thought) · [Persian deliberation anecdote](https://www.forbes.com/sites/chunkamui/2016/03/22/wine-and-sleep-make-for-better-decisions/) · [Salience bias](https://thedecisionlab.com/biases/salience-bias) · [Priming and anchoring](https://thedecisionlab.com/biases/anchoring-bias) · [Availability and experience](https://www.forbes.com/sites/brycehoffman/2024/04/06/availability-heuristic-what-it-is-and-how-to-overcome-it/) · [Conspiracy beliefs and anxiety](https://newrepublic.com/article/150922/pioneer-paranoia) · [Conspiracy scapegoats](https://www.bbc.com/future/article/20170809-the-accidental-invention-of-the-illuminati-conspiracy) · [Randomness and conspiracy beliefs](https://www.bbc.com/future/article/20180124-the-enduring-appeal-of-conspiracy-theories) · [Perceived enemies and control](https://www.independent.co.uk/news/world/europe/bilderberg-group-conspiracy-theories-secret-societies-new-world-order-alex-jones-a8377171.html) · [Working memory](https://www.psychologytoday.com/us/basics/subpage/working-memory) · [Capacity estimate: two items](https://bura.brunel.ac.uk/bitstream/2438/801/1/GobetClarkson.pdf) · [Capacity estimate: four items](https://www.cambridge.org/core/journals/behavioral-and-brain-sciences/article/magical-number-4-in-shortterm-memory-a-reconsideration-of-mental-storage-capacity/44023F1147D4A1D44BDC0AD226838496) · [Choice discomfort](https://thedecisionlab.com/biases/status-quo-bias) · [Uncertainty and brain activity](https://news.mit.edu/2021/how-the-brain-deals-with-uncertainty-1014) · [Threat monitoring](https://nesslabs.com/uncertain-mind)

<!-- atlas:block-content b-3ee07485ae5bfe0e:end -->
<!-- atlas:block-content b-2d139996553997c6:start -->
### What a gain or a loss means to the chooser

Reference points, diminishing sensitivity and loss aversion concern valuation. Expected utility separately explains why the same monetary payoff need not have the same value to everyone.

<a id="v2-0-1"></a>

**Prospect theory: three valuation mechanisms**

Prospect theory describes choices in terms of gains and losses relative to a reference point, diminishing sensitivity as gains or losses grow, and loss aversion. Probability weights can also depart from probabilities. These mechanisms describe systematic choice patterns; expected utility theory is a different model, and ordinary risk aversion is compatible with it. The three valuation mechanisms and the possibility and certainty effects are developed together below.

First, you evaluate situations by comparing them to a neutral reference point, which is usually your status quo—the situation you normally experience—so you make decisions you think will improve that status quo. However, your reference point can also be an outcome you expect or feel entitled to, such as an annual raise; that is why you can feel crushed when you do not get something you expected, even though your objective situation has not changed.

Second, sensitivity to a gain or loss tends to diminish as its distance from the reference point increases. An extra $100 can therefore have a larger effect when added to a small gain than to a large one. The summary illustrates diminishing value with wealth rising from $100 to $200, from $1,000 to $1,100, and a millionaire receiving another $100. These all add $100, but the example mixes changes from a reference point with the value of money at different wealth levels; it does not establish a rule that value equals a percentage of wealth.

Third, loss aversion describes losses generally weighing more heavily than equal-sized gains relative to the reference point. Losing $100 may matter more to a person than gaining $100. This is an equal-amount comparison across losses and gains, distinct from diminishing sensitivity within either side; its strength depends on the setting. These are the three value mechanisms, alongside the separate weighting of probabilities.

Kahneman also identifies practical implications of prospect theory, including the **possibility effect** and the certainty effect. The possibility effect describes how people overvalue the mere possibility of an outcome, even when it remains highly unlikely. For example, most people feel a stronger subjective shift going from a 0% chance of winning one million U.S. dollars ($1 million) to a 5% chance than they do going from a 5% chance of winning $1 million to a 10% chance, even though the objective increase in probability is identical in both cases. This effect explains why people fantasize about small chances of large gains—such as playing the lottery or gambling in casinos—and also why they fixate on worst-case scenarios even when the odds of those scenarios occurring are tiny.

<a id="b-2d139996553997c6"></a>
<a id="u-thinking-fast-and-slow-72a9596075ab"></a>
<a id="u-decisive-710db56a1c2c"></a>

<a id="v2-17-0"></a>

**Loss Aversion: Bias or Rational Ruin Avoidance?**

Shortform contrasts Kahneman’s account of loss aversion with Taleb’s argument in *Skin in the Game* for avoiding ruin—a loss from which recovery is impossible. Repeated exposure to small risks can threaten survival when losses compound or a catastrophic outcome ends participation. That gives a reason to examine downside, resources and repeated exposure; it does not show that every small loss leads to ruin or that every aversion to a loss is optimal. Taleb’s normative argument about survival also does not overturn the descriptive finding that equal gains and losses can receive unequal weight.

Biases such as loss aversion and the mere exposure effect do not exist in a vacuum; social circumstances shape how strongly we experience them and how much they affect us. For example, experimental research suggests that wealthy people and people in positions of power experience less loss aversion, because their social and financial safety nets make losses feel less painful. Research on the mere exposure effect also links it to racial prejudice: one study found that increasing a white person’s exposure to Asian and Black faces increased their liking of new faces from those groups. The researchers conclude that efforts to increase white people’s exposure to faces of color—for instance, in advertising and media—could reduce racial bias and its harmful consequences.

<a id="b-fc0471a1a7e454ba"></a>
<a id="u-thinking-fast-and-slow-118ab0c379c2"></a>
<a id="u-thinking-fast-and-slow-3f3f69875d97"></a>

<a id="v2-16-0"></a>

**Shortform perspective: Diminishing value and expectations**

Diminishing sensitivity in prospect theory is related to, but distinct from, diminishing marginal utility in economics. An additional dollar can matter greatly when it buys a necessary meal and far less when basic needs are secure. A first glass of water is valuable when you are thirsty; later glasses usually add less. This is not evidence that rational people must value every dollar equally.

Shortform connects reference points with the expression “happiness equals reality minus expectations.” Treat it as a reminder that expectations shape satisfaction, not as a psychological equation: dropping expectations to zero does not guarantee contentment under deprivation, pain, or mistreatment. Tara Brach’s radical acceptance likewise invites acknowledging present experience; it need not mean abandoning aspirations or accepting harmful conditions without acting.

<a id="b-d5c1b59f5e952158"></a>
<a id="u-thinking-fast-and-slow-e7e4b6e3e2d2"></a>
<a id="u-thinking-fast-and-slow-375bcd58ead6"></a>
<a id="u-thinking-fast-and-slow-b766dd82ca72"></a>
<a id="u-thinking-fast-and-slow-fe5dc0917b14"></a>
<a id="u-thinking-fast-and-slow-03c89fb0026e"></a>

<a id="v2-18-0"></a>

**The certainty effect**

Kahneman describes a disproportionate change in decision weight when an outcome becomes certain. Moving a recovery chance from 90% to 95%, or from 95% to 100%, adds five percentage points in either case; people may value the second change much more. This illustrates sensitivity to certainty. Whether a particular preference is unreasonable also depends on consequences and utility: a remaining 5% chance of a grave loss can matter greatly. A good prognosis does not make concern about that residual risk inherently irrational.

<a id="b-80c5e0b646b442c8"></a>
<a id="u-thinking-fast-and-slow-be5d131c20b5"></a>

<a id="v2-19-0"></a>

**A small probability still needs a consequence**

The possibility effect should be separated from reasonable concern about catastrophic harm. A risk judgment must consider both probability and severity, as well as whether the loss is survivable. This helps explain why a person can object to a small chance of a Chernobyl-scale accident without simply misunderstanding probability; assessing an energy choice still requires comparing the risks of its alternatives.

A move from 0% to 5% cannot be described as a defined percentage increase because the starting denominator is zero. A move from 5% to 10% doubles the probability. Both changes add five percentage points; in the $1 million lottery example, each adds $50,000 of expected monetary payoff. Their psychological impact can differ, but those feelings do not change the arithmetic.

<a id="b-6650ddc9a6f26e45"></a>
<a id="u-thinking-fast-and-slow-5b79be2bc59b"></a>
<a id="u-thinking-fast-and-slow-9b2511564c98"></a>
<a id="u-thinking-fast-and-slow-42cc60946b0a"></a>
<a id="u-thinking-fast-and-slow-090b4594a605"></a>
<a id="u-thinking-fast-and-slow-3a1ef0eb9241"></a>
<a id="u-thinking-fast-and-slow-d9f59bb4ff79"></a>

<a id="v2-98-0"></a>

**Expected money is not expected utility**

Expected utility theory compares probability-weighted utility: choose the option with the greatest sum of each outcome’s utility multiplied by its probability, given the model and your preferences. It does not require choosing the outcome with the greatest chance of success, nor treating every dollar as equally valuable.

For an 80% chance of $100 and a 20% chance of $10, the expected monetary payoff is $82; a guaranteed $80 has an expected monetary payoff of $80. A risk-neutral person whose utility is linear in these payoffs prefers the gamble. Someone with sufficiently concave utility can rationally prefer the certain $80. The comparison therefore does not, by itself, refute expected utility theory. Kahneman and Tversky’s prospect theory addresses other systematic patterns, including reference dependence, loss aversion and nonlinear decision weights; certainty preference alone is not enough to distinguish the theories.

<a id="b-e8410c1adab9049c"></a>
<a id="u-thinking-fast-and-slow-b7823065ed6c"></a>

<a id="v2-99-0"></a>

**Why the St. Petersburg game does not have an infinite prize**

In the classic St. Petersburg game, a fair coin is tossed until the first head. If that occurs on toss n, the payoff is 2ⁿ dollars. Every finite realized payoff is finite, but its expected monetary value diverges: each possible stopping time contributes one dollar to an infinite sum. There is no positive-probability outcome in which a player receives an infinite sum.

This mismatch between an infinite expectation and the modest amount people will pay helped motivate utility-based evaluation. Diminishing marginal utility, such as logarithmic utility, can give a finite expected utility for this game without assuming that money stops having value above a hard ceiling. Actual willingness to pay also depends on starting wealth and practical limits on the payer’s funds and the duration of play. This is a different issue from prospect theory’s reference-dependent valuation.

*Sources and further reading:* [Skin in the Game](https://www.shortform.com/app/book/skin-in-the-game/) · [Ruin and recovery](https://www.shortform.com/app/book/skin-in-the-game/part-3-chapter-11#loss-aversion-helps-us-avoid-ruin) · [Wealth, power and loss aversion](https://thedecisionlab.com/biases/loss-aversion) · [Face exposure and liking](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2705986/) · [Diminishing marginal utility](https://www.investopedia.com/terms/l/lawofdiminishingutility.asp) · [Expectations and happiness](https://www.theguardian.com/lifeandstyle/2013/oct/12/happiness-reality-expectations-oliver-burkeman) · [Radical Acceptance](https://www.shortform.com/app/book/radical-acceptance) · [Risk probability and impact](https://its.wsu.edu/information-security-services/policies-standards-and-guidelines/how-to-rate-impact-for-risk-assessments/) · [St. Petersburg paradox](http://www.econport.org/econport/request?page=man_ru_basics2)

<!-- atlas:block-content b-2d139996553997c6:end -->
<!-- atlas:block-content b-af3d870fe95fbb32:start -->
### Emotion can inform a goal and still distort an action

<a id="v2-0-2"></a>

**Emotional Decision-Making and Its Impulsive Consequences**

Emotions are a pervasive driver of irrational decision-making across multiple accounts. Bevelin frames emotional decision-making as a key irrational error rooted in evolutionary mismatch: we often make irrational errors because we allow our emotions to influence our decisions, and emotions can push us toward impulsive, unwise choices. He offers three concrete examples: feeling enthusiastic about a new job and taking on too many commitments; letting intense sexual desire lead you to cheat on your spouse; and fearing that a new phone model will sell out and impulsively buying it even though you do not need one. Prospect theory describes systematic patterns in risky choice; it should not be reduced to a claim that all departures from expected monetary value are emotional mistakes.

<a id="b-af3d870fe95fbb32"></a>
<a id="u-seeking-wisdom-bevelin-620fe4f32071"></a>
<a id="u-seeking-wisdom-bevelin-e914d078c85c"></a>

<a id="v2-28-0"></a>

**Shortform perspective: Emotions and Decision-Making: Interference or Aid?**

While Peter Bevelin asserts that emotions interfere with wise, rational thinking, some psychologists offer the counterpoint that emotions can benefit your decision-making process in several ways. First, emotions can motivate you to make important decisions in the first place; for instance, anger about an injustice may drive you to run for public office to address it. Second, some emotions can actually make your decisions less biased and therefore more rational. For example, research suggests that gratitude counteracts discount rate bias—the cognitive bias that causes you to value instant gratification over long-term rewards.

Shortform also notes a related view from psychologist Daniel Goleman in *Emotional Intelligence*: emotions can actually help people make rational decisions by increasing awareness of their own preferences. Knowing your emotional preferences helps you choose among competing priorities; if you ignore or downplay your emotions, you may become paralyzed by indecision or make a choice you later regret. For example, if you receive a bonus and are deciding between spending it on a vacation or saving it for a future purchase, both options have rational justifications, so you must consult your emotional preferences to reach a decision.

<a id="b-1f1df10bdf0ddd5d"></a>
<a id="u-seeking-wisdom-bevelin-e0d9e1cc5fb9"></a>
<a id="u-decisive-fed58089a916"></a>

<a id="v2-29-0"></a>

**Shortform perspective: Regulate the reaction before making the decision**

Shortform adds several ways to create a pause when emotion is intense. Slow, comfortable breathing—for example, four seconds in and eight out if that pace suits you—can reduce arousal. Progressive muscle relaxation works through the hands, arms, chest, legs, and feet, gently tensing and releasing each group; it can interrupt the cycle of muscle tension and perceived stress.

The summary also groups cold showers, immersion and icy water on the face together. These are different exposures: findings about dopamine after prolonged immersion or a slowed heart rate during facial cooling do not establish that all cold exposure is a calming parasympathetic intervention. Cold immersion can itself provoke a strong stress response. It should not be presented as a generally necessary preparation for clear thinking.

The complementary task is to notice feelings rather than merely suppress them. Shortform presents RAIN in four steps:
- **Recognize** the emotion by naming it.
- **Allow** it to be present without immediately trying to change it.
- **Investigate** where you feel it in your body and what may be contributing to it.
- **Non-identification**: the feeling is real, but it is not your whole identity or an instruction you must obey.

These practices create room to choose a response; they do not promise complete control over every emotion.

<a id="v2-0-3"></a>

**Practical Strategies for Countering Irrational Decision-Making**

Beyond Kahneman’s dual-system remedy, authors across the decision-making literature propose concrete strategies for reducing the impact of irrational thinking and improving decision quality. Many of these strategies share a common logic: they create structure or friction that counteracts impulsive, emotion-driven, or biased intuitive judgments.

Bevelin outlines a two-part approach to countering emotional decision-making. First, he argues that you should avoid making decisions under the influence of strong emotions; the best decisions come from taking the time to rationally weigh options, so whenever you feel intense emotion, you should hold off on deciding and focus on calming down first. Second, once you are calm, he recommends consulting **pre-established guidelines**—a list of rules or steps that tell you how to act. These guidelines counteract the influence of your emotional state by reminding you what actions to take or avoid, and should be created in advance so you can reference them whenever you face a decision. Bevelin bases his ideas on pre-established guidelines on the wisdom of Charles Munger and Warren Buffett, both of whom use this rational thinking strategy.

Parrish offers three practical strategies for protecting yourself from the consequences of irrational thinking: make bad decisions more effortful, limit your risks, and learn from every mistake. These approaches acknowledge that irrationality cannot be eliminated entirely, but its harms can be reduced through proactive safeguards.

<a id="b-3184166602567f0f"></a>
<a id="u-seeking-wisdom-bevelin-b57617684b6b"></a>
<a id="u-seeking-wisdom-bevelin-306bef598e6c"></a>

<a id="v2-30-0"></a>

**Pre-Established Guidelines as a Guard Against Emotional Decisions**

Bevelin recommends keeping calm and consulting pre-established guidelines to counteract emotional decision-making. He illustrates the approach with a spending example: if you often splurge on expensive purchases and regret it later, you could adopt a set of rules to consult before any spending decision:

- Don’t buy extra items just to reach the free shipping threshold for online purchases.
- Don’t buy a duplicate of an item you already own if the one you own serves its purpose.
- When you have several options for an important item to buy, opt for the second-least-expensive option.

<a id="b-f798f0d143e9b4da"></a>
<a id="u-seeking-wisdom-bevelin-4700be81fe32"></a>
<a id="u-clear-thinking-099ccccb4c20"></a>

<a id="v2-31-0"></a>

**Shortform perspective: Pre-Established Rules and Making Bad Decisions Harder**

One strategy for countering irrational, emotion-driven decisions is to set pre-established guidelines ahead of time. However, simply having rules does not guarantee you will follow them when temptation arises. Two evidence-based tactics can help you stick to your guidelines by making impulsive choices harder.

First, eliminate reminders of the temptations you want to resist. In *Atomic Habits*, James Clear argues that removing cues makes temptation easier to withstand; for example, to reinforce spending rules, you could unsubscribe from your favorite stores’ marketing lists so you are not constantly reminded of products you might want.

Second, make temptations more inconvenient. For instance, delete your credit card information from online shopping accounts; the extra effort of re-entering it may dissuade you from buying something you do not need.

Shane Parrish’s principle of making bad decisions more effortful also aligns with findings about religious and spiritual practices. According to Roy F. Baumeister and John Tierney in *Willpower*, such practices improve people’s lives in two ways: they provide clear behavioral rules, often framed as divine mandates, that save the mental effort of wrestling with everyday decisions; and they create accountability through community observation and the sense of being watched by a higher power, which raises the perceived cost of bad choices. Non-religious people can gain similar benefits by joining secular groups with guiding rules and peer accountability, such as gyms or martial arts schools.

<a id="b-eacc7f3b3d9b4c84"></a>
<a id="u-clear-thinking-8007425e04bf"></a>
<a id="u-clear-thinking-a9fb18618c37"></a>
<a id="u-clear-thinking-a563f5671c91"></a>
<a id="u-clear-thinking-993f979d0adc"></a>

<a id="v2-32-0"></a>

**Making Bad Decisions More Effortful**

Parrish’s first tip for compensating for irrational thinking is to make bad decisions more effortful by planning ahead for moments of irrationality. The approach has two main forms. First, identify your bad habits and structure your circumstances so that indulging them becomes more difficult; over time, you will naturally do the right thing simply because it is the easiest option. For instance, if you regularly work late and ruin your sleep, you could give your partner permission to walk in and shut off your computer at 7 p.m. no matter what. Because arguing to keep working would be difficult and unpleasant, you will naturally plan around the 7 p.m. cutoff. Second, set clear rules that outright forbid bad decisions. Rules eliminate the emotional struggle of making good choices case by case, so you can act wisely with less willpower. As an example, you could improve your sleep by adopting a rule that you never bring electronic devices into your bedroom.

<a id="b-9f9678c081095f21"></a>
<a id="u-clear-thinking-0f00e81c8da6"></a>
<a id="u-clear-thinking-795fb935f236"></a>

<a id="v2-33-0"></a>

**Shortform perspective: Cautious Planning and Effective Low-Stakes Practice**

Nassim Nicholas Taleb’s *The Black Swan* provides further justification for highly cautious approaches to decision-making. Taleb contends that humans consistently underestimate how surprising the future can be: we build oversimplified models of reality and treat them as far more accurate than they really are. In a fundamentally unpredictable world, he argues, we should stop trying to predict every possible failure and instead design plans robust enough to withstand extreme disasters.

Low-stakes practice environments are one way to build that robustness, but they only work well if you are actually learning from them. Cal Newport argues in *So Good They Can’t Ignore You* that optimal learning experiences meet two conditions. First, they must be appropriately challenging: you learn the most when a task stretches your skills just beyond your current comfort zone. Second, they must deliver precise, immediate feedback about what you are doing right and wrong. For example, a tennis coach might point out that you drop your elbow during a swing, or a code analysis tool might flag a mistake the moment you make it.

<a id="v2-23-2"></a>

**Limiting Risk: Prepare for the Worst and Learn Safely**

Parrish’s second tip is to limit your risks—that is, reduce what you have to lose. Since no one makes perfect decisions all the time, he argues, it is rational to assume things will go wrong. You should make choices that protect you in worst-case scenarios, even if they seem overly cautious. His recommended practice is to picture the worst possible outcome, then prepare for something twice as severe; this is a buffer heuristic, not a guarantee against unprecedented disasters; choose reserves and exposure limits for the actual failure modes. Parrish adds that risk limitation is especially important when you are first learning a new skill. Rather than jumping into high-risk, high-reward environments as soon as possible, seek low-stakes settings where you can experiment freely and build expertise through trial and error. That way, when you eventually face high-stakes decisions, you will have learned from some mistakes while their costs were lower.

<a id="b-209e83eb5f0cf0b0"></a>
<a id="u-seeking-wisdom-bevelin-dff9d981720a"></a>
<a id="u-seeking-wisdom-bevelin-de5a27ad954b"></a>

<a id="v2-34-0"></a>

**Rationality, Emotion, and Success: The Buffett-Munger Example**

Rational thinking is often linked to success, but not all successful decisions are purely rational. For example, one military strategy expert argues that leaders’ emotions can contribute to their success as long as those emotions align with their objectives. He cites Otto von Bismarck, the former German chancellor, as a successful strategist whose fear compelled him to act quickly in battle; far from impairing his judgment, this fear supported his goal of acting offensively rather than defensively.

Warren Buffett and Charlie Munger are widely held up as exemplars of rational, disciplined decision-making. Prominent figures such as Bill Gates praise Buffett for his analytical thinking, and many authors have written books about his investment strategies. Admirers of Munger describe him as a polymath with extraordinary rational thinking skills. Both are also widely praised for their philanthropy: Munger has donated hundreds of millions of dollars for student housing at several universities plus $21 million to Good Samaritan Hospital, while Buffett has given more than $41 billion to various charities.

Critics, however, have chastised the pair for irresponsible investing decisions. One report estimates that billionaires’ investments make them responsible for a million times more greenhouse gas emissions than the average person. Critics of Berkshire Hathaway argue the company is doing less to curb emissions than its peers, despite Buffett’s claims that the company is committed to fighting climate change.

<a id="b-9b435aa9e6b46a9d"></a>
<a id="u-seeking-wisdom-bevelin-fcb4e950e25c"></a>
<a id="u-seeking-wisdom-bevelin-95f517b70e73"></a>

<a id="v2-35-0"></a>

**Learning rational thinking from highly successful thinkers**

Bevelin argues that we should study rational-thinking strategies used by people who achieved unusual financial success or contributed important ideas. He contends that many such figures succeeded partly because disciplined reasoning helped them avoid major mistakes. Throughout *Seeking Wisdom*, he draws on thinkers from many fields, with particular attention to investors Warren Buffett and Charles Munger, whom he presents as unusually committed to rational thought. Munger, who died in 2023, was Berkshire Hathaway's vice chairman—not its chief executive.

<a id="b-8f4cb62dc102b180"></a>
<a id="u-clear-thinking-3c8015847fdb"></a>

<a id="v2-39-0"></a>

**Nudging: Designing Environments to Compensate for Irrationality**

Shortform notes that the core logic of compensating for unavoidable human irrationality underpins the governance framework proposed by Richard Thaler and Cass Sunstein in *Nudge*. They argue that policymakers should design environments that subtly steer people toward the choices they would make if they were thinking clearly, rather than trying to force rationality. For example, a city council could discourage smoking by requiring stores to store tobacco products out of sight in opaque cases, making the decision to buy feel less automatic and obvious.

<a id="b-b98317a679e28361"></a>
<a id="u-seeking-wisdom-bevelin-e80b0c7e8bc6"></a>
<a id="u-seeking-wisdom-bevelin-1e22de7a9e0f"></a>

<a id="v2-36-0"></a>

**The Moral Failure of Following the Crowd**

Bevelin names a third irrational error tied to evolutionary mismatch: failing to make the moral choice. He argues that we often make immoral choices because we irrationally base our decisions on what other people are doing rather than on what we ourselves think is right. One reason for this failure is fear that holding an unpopular opinion will invite criticism. He gives the example of a colleague making a sexist comment in a meeting; you stay silent because you do not want others to criticize you for being overly sensitive or disloyal.

<a id="b-f7805633768ccd18"></a>
<a id="u-seeking-wisdom-bevelin-5676702fc458"></a>

<a id="v2-37-0"></a>

**Shortform perspective: Conformity, Moral Choice, and the Banality of Evil**

Shortform contextualizes the tendency to conform to others’ expectations and make immoral choices by linking it to the concept of the banality of evil—the idea that ordinary people can commit extremely immoral acts when under pressure to conform or to follow authority figures’ orders. Philosopher Hannah Arendt proposed this idea to explain why ordinary Germans became complicit in the atrocities of the Holocaust.

<a id="b-9fa021190033a9ed"></a>
<a id="u-seeking-wisdom-bevelin-948f7da1a6b8"></a>

<a id="v2-116-0"></a>

**Shortform perspective: Using Social Approval to Drive Moral Decisions**

Buffett’s front-page test—imagining how your decision would look if reported on a newspaper’s front page—works, Shortform infers, by tapping into our deep-seated desire for social approval. Bevelin does not explain why the test works, but the mechanism likely involves expanding the imagined audience of your choice to anyone who might read the story, making you fear disapproval of immoral actions and lean toward choices you think others would endorse.

This is one of several strategies that redirect our natural hunger for approval toward moral ends. Because we are susceptible to obeying authority figures and may be pressured into immoral choices, some experts recommend shifting your allegiances: create distance from the authority figure and align yourself instead with potential victims, so that you act morally to gain the victims’ approval rather than acting immorally to please the authority.

*Sources and further reading:* [Emotions in decisions](https://www.annualreviews.org/doi/10.1146/annurev-psych-010213-115043) · [Emotional Intelligence](https://www.shortform.com/app/book/emotional-intelligence/1-page-summary) · [Emotional preferences](https://www.shortform.com/app/book/emotional-intelligence/chapter-2#the-neocortex) · [Parasympathetic function](https://www.hss.edu/article_parasympathetic-nervous-system.asp) · [Calming practices](https://www.psychologytoday.com/us/blog/click-here-happiness/202011/10-science-based-relaxation-techniques) · [Cold exposure and dopamine](https://pubmed.ncbi.nlm.nih.gov/10751106/) · [Facial cooling and heart rate](https://www.jstage.jst.go.jp/article/circj/70/6/70_6_773/_article) · [Paced breathing](https://hbr.org/2020/09/research-why-breathing-is-so-effective-at-reducing-stress) · [Muscle tension and stress](https://myhealth.alberta.ca/Health/aftercareinformation/pages/conditions.aspx?hwid=abs2086#:~:text=The%20body%20responds%20to%20stress,tension%20and%20general%20mental%20anxiety.) · [Progressive muscle relaxation](https://www.sciencedirect.com/topics/medicine-and-dentistry/progressive-muscle-relaxation) · [Muscle-relaxation sequence](https://www.cci.health.wa.gov.au/-/media/CCI/Mental-Health-Professionals/Panic/Panic---Information-Sheets/Panic-Information-Sheet---05---Progressive-Muscle-Relaxation.pdf) · [Emotional suppression](https://www.hcf.com.au/health-agenda/body-mind/mental-health/downsides-to-always-being-positive) · [RAIN practice](https://adaa.org/living-with-anxiety/personal-stories/rain-mindful-framework-addressing-anxious-thoughts) · [Atomic Habits](https://short.zaynjarvis.com/books/atomic-habits/) · [Friction against temptation](https://www.nytimes.com/2019/03/25/smarter-living/why-you-procrastinate-it-has-nothing-to-do-with-self-control.html) · [Willpower](https://www.shortform.com/app/book/willpower/1-page-summary) · [Religion and self-control](https://www.shortform.com/app/book/willpower/1-page-summary#religion-is-an-especially-effective-tool) · [The Black Swan](https://www.shortform.com/app/book/the-black-swan) · [Model simplification](https://www.shortform.com/app/book/the-black-swan/chapter-1#platonicity-and-the-platonic-fold) · [Preparing for extremes](https://www.shortform.com/app/book/the-black-swan/chapter-7#5-prepare-dont-predict) · [So Good They Can’t Ignore You](https://www.shortform.com/app/book/so-good-they-cant-ignore-you) · [Challenge and feedback](https://www.shortform.com/app/book/so-good-they-cant-ignore-you/chapters-4-7#how-to-acquire-career-capital-practice-deliberately) · [Leadership emotions](https://digitalcommons.usf.edu/cgi/viewcontent.cgi?article=1974&context=jss) · [Buffett’s analytical approach](https://hbr.org/1996/01/what-i-learned-from-warren-buffett) · [Buffett’s investing approach](https://www.investopedia.com/articles/investing/092214/best-books-warren-buffett.asp) · [Munger’s thinking](https://fs.blog/intellectual-giants/charlie-munger/) · [Philanthropy](https://www.cnbc.com/2021/06/30/buffett-and-munger-this-is-what-we-really-wanted-more-than-money.html) · [Investment-related emissions](https://www.cnbc.com/2022/11/08/billionaires-emit-a-million-times-more-greenhouse-gases-than-the-average-person-oxfam.html) · [Berkshire climate criticism](https://www.nytimes.com/2022/04/25/business/energy-environment/warren-buffett-climate-change.html) · [Berkshire 2022 Form 10-K](https://www.berkshirehathaway.com/2022ar/202210-k.pdf) · [Berkshire: Munger announcement](https://www.berkshirehathaway.com/news/nov2823.pdf) · [Nudge](https://www.shortform.com/app/book/nudge) · [Designing for fallibility](https://www.shortform.com/app/book/nudge/part-1#common-biases) · [Choice architecture](https://www.shortform.com/app/book/nudge/part-1#libertarian-paternalism-and-choice-design) · [Banality of evil](https://www.themarginalian.org/2017/02/07/hannah-arendt-the-banality-of-evil/) · [Authority and victims](https://effectiviology.com/authority-bias-the-milgram-obedience-experiment/)

<!-- atlas:block-content b-af3d870fe95fbb32:end -->
<!-- atlas:block-content b-afe851fecb16f774:start -->
### Evolutionary explanations are not complete mechanisms

<a id="v2-23-0"></a>

**Bevelin’s evolutionary-mismatch account**

Bevelin proposes two broad sources of error: behavior shaped in earlier environments may fit present demands poorly, and people may misunderstand scientific concepts. His evolutionary-mismatch account moves from ancestral behavior to present errors and then to corrective practices. It is an explanatory framework, not evidence that each specific ancestral scene in the summary occurred as described. Neither ancestral life nor modern life has one uniform set of risks; physical, social and psychological threats can occur in both.

<a id="b-afe851fecb16f774"></a>
<a id="u-seeking-wisdom-bevelin-458ec0fb2661"></a>
<a id="u-seeking-wisdom-bevelin-77c592697f12"></a>
<a id="u-seeking-wisdom-bevelin-0ced48897a09"></a>
<a id="u-thinking-fast-and-slow-37b4dff5ca5f"></a>
<a id="u-seeking-wisdom-bevelin-baef2f1d21c6"></a>

<a id="v2-24-0"></a>

**Evolutionary interpretations of fast responses**

<a id="b-bcc923db1cae377e"></a>
<a id="u-thinking-fast-and-slow-cdbdbddf255f"></a>
<a id="v2-25-0"></a>

Bevelin’s predator scenarios illustrate why the cost of delay can favor a fast response: rustling leaves might prompt hiding or reaching for a weapon, and freezing after movement in the trees might avoid drawing attention. Shortform’s commentary on Thinking, Fast and Slow adds a separate example: eyes in a bush might prompt flight before identification. These are teaching scenarios, not demonstrated comparisons of which ancestors survived or passed on “quick-thinking genes.” They make a trade-off visible: a false alarm and a missed threat can have very different costs. Kahneman similarly offers prioritizing threats over opportunities as an evolutionary interpretation of loss aversion, whose descriptive definition appears in the prospect-theory section. Neither story by itself establishes its genetic mechanism.

Shortform’s phrase “survival of the adequate” emphasizes that evolution does not aim at perfection. It does not mean accurate judgment has never mattered, that careful thinkers left no descendants, or that modern humans have no need or capacity for greater accuracy.










<a id="b-a2c836c6c8dc0c8b"></a>
<a id="u-seeking-wisdom-bevelin-d7131fd67b76"></a>

<a id="v2-26-0"></a>

**Shortform perspective: Stress, Evolutionary Mismatch, and Decision Quality**

Shortform adds John Medina’s *Brain Rules* to distinguish acute from chronic stress. A short stress response can help meet an immediate demand; prolonged stress can interfere with memory and learning, with cortisol among the mechanisms discussed. That distinction is useful without assuming chronic stress never occurred in ancestral life or that every modern stress response is maladaptive. Duration, intensity and context matter to its effect on deliberate judgment.

<a id="b-8dbc1970951e280a"></a>
<a id="u-seeking-wisdom-bevelin-99e3ce106fa5"></a>
<a id="u-seeking-wisdom-bevelin-e704749748ec"></a>
<a id="u-seeking-wisdom-bevelin-4217435d7b81"></a>

<a id="v2-27-0"></a>

**Shortform perspective: Evolutionary Mismatch and Binary Thinking Biases**

The idea of evolutionary mismatch has been widely used by popular science authors to argue for behavioral and lifestyle change, and writers generally fall into two camps. One camp, which includes Peter Bevelin, argues that our hunter-gatherer minds are outdated and that we need to improve our thinking strategies to meet modern demands. The other camp does not see our minds as flawed; instead, it argues that the modern environment is the problem, and advocates reintroducing ancestral lifestyle features—such as a Paleo diet modeled on hunter-gatherer eating patterns—into contemporary life.

Shortform connects quick “good/bad” categorization with Bevelin’s evolutionary account. Whatever its origin, the present error can be examined directly: treating one’s political party as good and therefore every policy it proposes as good replaces scrutiny of individual policies with a group label. The claim that this was the most efficient ancestral survival strategy is an interpretation, not evidence supplied by that example.

Binary, association-based biases can also operate outside conscious awareness. Psychologist Jennifer Eberhardt explains in *Biased* that the culture we grow up in shapes implicit bias—unconscious associations that link groups of people with “good” or “bad” traits. Implicit anti-Black bias is a pervasive problem in the US; for instance, the harmful stereotype that Black girls are not good at math can lead a teacher with such an implicit association to underestimate students’ abilities and hinder their learning, if the bias goes unrecognized.

<a id="b-3ae9219c356da43c"></a>
<a id="u-seeking-wisdom-bevelin-054a6827671f"></a>
<a id="u-seeking-wisdom-bevelin-9143b97026cd"></a>
<a id="u-seeking-wisdom-bevelin-c5c9babae0e8"></a>
<a id="u-seeking-wisdom-bevelin-19b15b1c68a7"></a>
<a id="u-seeking-wisdom-bevelin-0569ca4967de"></a>
<a id="u-seeking-wisdom-bevelin-def019a8b15e"></a>

<a id="v2-303-0"></a>

**Evolutionary mismatch as a root of judgment error**

Bevelin emphasizes the long foraging portion of human history—the summary uses “nearly 99%,” a rough framing dependent on what span counts as human history. His account links seeking food, sex and companionship and avoiding harm with survival and reproduction. He then examines hasty conclusions, emotional choices and conformity in moral decisions. These present-day problems can be studied without treating the source’s broad history as a complete account of every person’s brain or threats.

<a id="b-6bc25f43f1f633ed"></a>
<a id="u-seeking-wisdom-bevelin-c0628b879688"></a>
<a id="u-seeking-wisdom-bevelin-fc92656f7805"></a>

<a id="v2-305-0"></a>

**Shortform perspective: Hunter-Gatherer Brains, Modern Life, and the Neuroscience of Pleasure and Pain**

Although Bevelin argues that our hunter-gatherer brains are ill-suited for the demands of modern life, he discusses the hunter-gatherer lifestyle as a relic of the past. Shortform notes that modern hunter-gatherer societies exist in many regions around the world, and their lifestyles differ in some respects from those of the ancestral hunter-gatherers Bevelin describes. For example, some modern hunter-gatherers supplement their diets with cultivated food—the Pumé people of South America, for instance, grow cassava. Some have also incorporated technology into daily life: Inuit communities in the Arctic use snowmobiles to travel between hunting grounds.

Shortform also cites a study linking particular GABAergic and glutamatergic neurons with approach and avoidance in its experimental setting. These broad neurotransmitter-based categories do not divide the entire brain into pleasure-seeking and pain-avoiding systems. Without the species, brain circuit and task conditions, the result cannot establish a universal mechanism or independently prove an evolutionary account. The defensible point is narrower: reward and aversion can influence behavior through specific interacting circuits.

<a id="b-5fe09545292b4ef9"></a>
<a id="u-clear-thinking-d85bec4d566e"></a>
<a id="u-same-as-ever-1c3a607fb8e3"></a>

<a id="v2-333-0"></a>

**Shortform perspective: A pause does not require a three-layer brain story**

Shortform connects Greene’s advice about fast emotion and slower reflection with the older “triune brain” account: reptilian instinct, mammalian emotion and a later rational neocortex. That picture should not be used as an anatomical explanation. Emotion and cognition arise from interacting systems, rather than three independent evolutionary layers that take turns controlling behavior. Nor does the model justify saying we can never understand a mood’s source.

The practical issue remains: an initial emotional reaction may arrive before a considered response, and expressions do not reveal another person’s entire inner state. Pausing, examining the trigger and asking rather than guessing can improve the opportunity to respond deliberately. A pause is a technique, not a guarantee that one brain region will take over and eliminate mistakes.

<a id="b-61c255d795669104"></a>
<a id="u-seeking-wisdom-bevelin-67cd65b74187"></a>

<a id="v2-291-0"></a>

**Shortform perspective: Scientific Illiteracy and Evolutionary Mismatch**

Although Bevelin treats scientific illiteracy as a separate origin of error from evolutionary mismatch, the two may be connected. In *The Art of Thinking Clearly*, Rolf Dobelli argues that our hunter-gatherer minds did not evolve to handle complex math, which underlies many scientific concepts. Our ancestors mostly dealt with linear math—simple addition and multiplication—whereas modern life demands understanding of exponential math, such as interest rates and statistics like "city crime is rising at 5% per year."

*Sources and further reading:* [Evolutionary account of intuition](https://pmc.ncbi.nlm.nih.gov/articles/PMC11892090/#ejn70024-sec-0001) · [Survival of the adequate](https://www.latimes.com/archives/la-xpm-1985-05-11-me-9726-story.html) · [Brain Rules](https://www.shortform.com/app/book/brain-rules/1-page-summary) · [Chronic stress](https://www.shortform.com/app/book/brain-rules#rule-4-stress-can-hurt-the-brain) · [Evolutionary-mismatch arguments](https://blog.castac.org/2018/07/evolutionary-mismatch/) · [Paleo-diet analogy](https://www.eatingwell.com/article/290400/the-paleo-diet-is-eating-like-our-very-distant-ancestors-really-a-good-idea/#:~:text=What%20Is%20the%20Paleo%20Diet,wild%2Dcaught%20fish%20and%20nuts.) · [Binary-categorization account](https://www.psychologytoday.com/us/blog/overcoming-destructive-anger/202108/the-potential-all-or-nothing-thinking-fuel-anger#:~:text=dichotomous%20thinking%20derives%20from,differentiate%20whether%20or%20not%20they) · [Biased](https://www.shortform.com/app/book/biased/1-page-summary) · [Culture and implicit bias](https://www.shortform.com/app/book/biased/chapters-1-2) · [Stereotypes and mathematics](https://files.eric.ed.gov/fulltext/ED533550.pdf) · [Contemporary hunter-gatherers](http://sarweb.org/media/files/unm_whyforage.pdf) · [Approach and avoidance circuits](https://www.psychologytoday.com/us/blog/the-athletes-way/202001/the-neuroscience-seeking-pleasure-and-avoiding-pain) · [The Laws of Human Nature](https://www.shortform.com/app/book/the-laws-of-human-nature/law-1) · [Emotion and awareness](https://www.shortform.com/app/book/the-laws-of-human-nature/introduction) · [Historical layered-brain account](https://www.shortform.com/app/book/emotional-intelligence/chapter-1#the-science-behind-emotions) · [The Art of Thinking Clearly](https://www.shortform.com/app/book/the-art-of-thinking-clearly/1-page-summary) · [Evolution and mathematical intuition](https://www.shortform.com/app/book/the-art-of-thinking-clearly/1-page-summary#you-struggle-to-understand-complex-math)

<!-- atlas:block-content b-afe851fecb16f774:end -->
<!-- atlas:block-content b-57a1685d3a5302c5:start -->
### Habits that create room for a response

<a id="b-57a1685d3a5302c5"></a>
<a id="u-same-as-ever-5854c9fd96af"></a>
<a id="u-same-as-ever-20451b08b18d"></a>
<a id="u-same-as-ever-977c04ed3195"></a>
<a id="u-same-as-ever-35f8e8088b85"></a>
<a id="u-same-as-ever-c337c475be1b"></a>
<a id="u-same-as-ever-84a1cf600fa5"></a>
<a id="u-same-as-ever-81b9c684a005"></a>
<a id="u-clear-thinking-9345efeedea4"></a>
<a id="u-clear-thinking-dcb9c862d332"></a>
<a id="u-clear-thinking-289243daf905"></a>
<a id="u-clear-thinking-9780c51b3bec"></a>
<a id="u-clear-thinking-6191742adc4f"></a>

<a id="v2-299-habits"></a>

**Parrish’s four habits**

Parrish’s separate account emphasizes habits that create space between a trigger and an action. Fear of upsetting a boss, for example, can lead to accepting an unrealistic deadline and a month of overwork. His four habits are taking responsibility for one’s response, studying oneself objectively, regulating emotion, and drawing motivation from ambitious goals. His evolutionary story is one interpretation of impulsivity; practicing a pause does not require treating that story as a complete neuroscience account.

<a id="b-b90ed790e7bef012"></a>
<a id="u-clear-thinking-0102c4bb233e"></a>
<a id="u-clear-thinking-4f3ab470c502"></a>
<a id="u-clear-thinking-bbeff632b068"></a>

<a id="v2-323-0"></a>

**Taking Total Responsibility for Your Circumstances**

Parrish’s first habit focuses on agency: examine what response remains available when circumstances are unfair or a plan fails. Blaming others can sometimes become a way to avoid learning from one’s own contribution. But responsibility for a response is different from causing the event, and other people or institutions can have real duties to help or repair harm. A person may need support or collective action; acknowledging those constraints need not prevent taking the next feasible step.

<a id="b-150ec480928d9258"></a>
<a id="u-clear-thinking-0fda1aa0ec2e"></a>
<a id="u-same-as-ever-9f9f479ba4e1"></a>
<a id="u-clear-thinking-fccfe2282d50"></a>

<a id="v2-322-0"></a>

**Shortform perspective: Responsibility, Blame, and the Bias Toward Short-Term Thinking**

Two distinct psychological forces shape how we respond to setbacks: the urge to avoid blame, and the brain’s built-in bias toward immediate concerns. People often avoid taking responsibility for their situation because they fear being blamed for causing it. Yet as Mark Manson argues in *The Subtle Art of Not Giving a F∗ck*, you can take responsibility for your situation without blaming yourself—you can look for ways to improve things without feeling guilty about how you got there. The deeper reason people resist responsibility, Carol Tavris and Elliot Aronson explain in *Mistakes Were Made (But Not By Me)*, is cognitive dissonance: when information about ourselves conflicts with our self-image, we feel acute discomfort and are driven to resolve it. Most people believe they are more competent than average, so a mistake that suggests incompetence triggers dissonance; shifting blame or denying the mistake entirely are common ways to relieve that discomfort.

A related bias pulls our attention toward the present rather than the future. Robert Greene argues in *The Laws of Human Nature* that long-term thinking is hard to maintain because our brains evolved to scan for immediate rather than far-off danger. This adaptation served us well when survival depended on noticing dramatic threats and reacting quickly, but in a more complex world, long-term dangers are often larger problems than anything we face right now. To break out of short-term thinking, Greene recommends cultivating a farsighted perspective—simulating hindsight and clearing emotion from your judgment. It has three elements: detachment, or viewing problems without emotion so you can put short-term setbacks in perspective and stay committed to long-term goals; completeness, or looking at all the context of a problem, not just the most attention-grabbing details; and prediction, or weighing the negative consequences, new problems, and long-term goal implications of every possible decision.

<a id="b-61f4d52274817c06"></a>
<a id="u-clear-thinking-df19c8517691"></a>
<a id="u-clear-thinking-80359ccbaecd"></a>
<a id="u-clear-thinking-992be441095c"></a>

<a id="v2-324-0"></a>

**Objective Self-Assessment and the Limits of Self-Awareness**

Parrish’s second core habit of rational thinking is to assess yourself objectively. Accurately judging your own capabilities and limits, he argues, keeps you from taking on challenges you cannot handle. For example, if a project manager hired to oversee the development of a software product admits they do not fully understand the technology, they can avoid launching a flawed product whose issues they lack the technical expertise to spot. Parrish acknowledges that this kind of self-awareness is difficult to cultivate because overestimating your abilities feels good; many people remain ignorant of their flaws because doing so inflates their confidence and self-esteem. He suggests a simple check: reflect on how often you admit to others that you do not know something. If you rarely do, you are likely overestimating yourself.

<a id="b-dff3d3134520df6b"></a>
<a id="u-clear-thinking-a571e259a67f"></a>

<a id="v2-325-0"></a>

**Shortform perspective: Balancing Humility and Confidence in Self-Assessment**

Objective self-assessment requires balancing humility with confidence. While overconfidence undermines rational thinking, excessive self-doubt creates its own problems. Ray Dalio argues in *Principles* that you can overcome challenges beyond what you could achieve alone by drawing on the expertise of people who know more than you—but this strategy requires humility: if you are so overconfident that you think you do not need other people’s ideas, you will never seek them out. At the same time, it also requires a degree of self-confidence. Dalio notes that if you give up on cultivating your own expertise, you will not be able to tell which ideas from others are useful and which are not. The right balance is to never assume you know everything, but never assume you know nothing, either.

<a id="b-7088e2bb110718fe"></a>
<a id="u-clear-thinking-3d60cefcb0c1"></a>
<a id="u-clear-thinking-3e3bdb512101"></a>
<a id="u-clear-thinking-347a75dd4096"></a>
<a id="u-clear-thinking-8305d310666b"></a>
<a id="u-clear-thinking-d1ae56d9b24d"></a>

<a id="v2-326-0"></a>

**Emotional Mastery and the Cost of Impulsive Reactions**

The third habit Parrish emphasizes is mastering your emotions: learning to control your emotions rather than letting them control you. Strong feelings such as fear or anger can push you into decisions you would never make with a cool head. Emotional regulation matters because a single impulsive reaction can destroy years of progress toward your goals. For instance, if a manager is insulted by a subordinate in front of their boss, gets angry, and sends an email badmouthing the employee to the entire company, they may never rebuild their professional reputation there. To avoid such destructive reactions, cultivate the ability to pause and reflect when faced with emotional triggers, so you can rationally decide how to respond. Parrish notes that resisting your own biological signals is difficult, but you will improve with practice.

Parrish also recommends cultivating supportive emotions, particularly confidence grounded in evidence of past effort and recovery. Remembering a difficulty you handled can help counter discouragement and sustain work. It may change the emotional response; it does not guarantee a complete shift of mood or justify ignoring new evidence that the plan is failing.

<a id="b-84968c8ff739c25d"></a>
<a id="u-clear-thinking-0850de49fb93"></a>

<a id="v2-327-0"></a>

**Shortform perspective: Defusion: Separating Yourself From Emotional Triggers**

A key part of getting better at managing emotional triggers is defusion from your thoughts. In *The Happiness Trap*, Russ Harris explains that defusion means recognizing your thoughts and emotions as passing mental events rather than absolute truths that demand immediate action. For example, the manager in the scenario might notice, “I’m having the thought that everyone is judging me.” Instead of immediately accepting that belief as fact, they can gain a clearer view of the situation and pause before making a catastrophic mistake.

<a id="b-b61a7d86448ee0a2"></a>
<a id="u-clear-thinking-6a13a3a6ce93"></a>
<a id="u-clear-thinking-df388a22a9ba"></a>
<a id="u-clear-thinking-c91f26555964"></a>

<a id="v2-328-0"></a>

**Lofty Ambitions and the Motivation to Stand Out**

Shane Parrish argues that setting a high bar for your own work is essential if you want to accomplish great things; without this commitment, you will get mediocre results. Human beings are innately motivated to fit in with those around them, and mediocre work is the social norm, so anyone aiming for something greater must be prepared to stand out as “different” and face negative judgment for it. To inspire yourself to meet higher standards, Parrish recommends identifying people who embody the high standards you want to adopt, and intentionally committing to fitting in with them. He clarifies that this does not necessarily mean cutting “average” people out of your life; rather, spend as much time with your role models as possible, either directly or through media such as books and podcasts. This exposure will inspire you to act like them.

<a id="b-fcd1b7f26aafdec6"></a>
<a id="u-seeking-wisdom-bevelin-aef7e2a86829"></a>
<a id="u-same-as-ever-4137a5c1d366"></a>
<a id="u-clear-thinking-60a7e7dea478"></a>
<a id="u-same-as-ever-c66084f97953"></a>

<a id="v2-304-1"></a>

**Shortform perspective: Role Models, Ambition, and the Need for Solitary Introspection**

Shortform notes that Luke Burgis, in *Wanting*, also recommends spending time with role models who can inspire you to do great things. However, Burgis adds that sometimes the opposite is also necessary: cutting off contact with all your role models and engaging in deep, solitary introspection. This practice can help you distinguish your meaningful aspirations from the misleading ones you may have picked up from flawed role models.

<a id="b-2ee147b2595944a9"></a>
<a id="u-clear-thinking-fe29c48c430b"></a>
<a id="u-clear-thinking-4424e5500dc6"></a>

<a id="v2-119-0"></a>

**Shortform perspective: Meaning, Accomplishment, and the Purpose of Clear Thinking**

Oliver Burkeman (*Four Thousand Weeks*) agrees with Parrish that you should be conscious of your finite lifespan and spend your limited time on pursuits you find meaningful, but he disagrees that meaningful accomplishments are the path to happiness. In Burkeman’s view, you would be happier recognizing that your accomplishments do not matter that much in the grand scheme of things, and that you will not be able to change much about the world in your lifetime. He therefore argues you should spend your time in whatever way is personally meaningful to you—whether that means enjoying the present moment or making small, concrete improvements to others’ lives, such as spending a day keeping your grandparents company. From this perspective, clear thinking is not just thinking that helps you make “good” decisions and reach grand life goals; it can also mean freeing yourself from the pressure to achieve those grand goals in the first place.

Johann Hari (*Lost Connections*) refers to temporarily gratifying goals as “junk values”: like junk food, they feel good at the time, but too much of them will make you sick. Hari adds that people who adopt junk values also tend to have lower-quality relationships and less secure self-esteem.

<a id="b-ee9c4d135044d7cf"></a>
<a id="u-decisive-df86edc45e28"></a>

<a id="v2-118-0"></a>

**Shortform perspective: Social Inequality, Scarcity, and Decision-Making Capacity**

The Heath brothers’ perspective that we can improve our decision-making skills by resisting our bad habits suggests that all individuals have the capacity to make strong decisions, and implies that it is an individual’s responsibility to improve their decision-making skills. By contrast, other scholars who study decision-making in a social and cultural context emphasize how social inequality influences our capacity for deliberating on decisions. Their research suggests that both individuals and policymakers—along with the social and economic conditions the latter create—are responsible for individuals’ decisions.

Scarcity research suggests that demands created by shortages of money, time or space can absorb attention and constrain deliberation. Poverty can make such demands persistent. This does not establish that every poor person has less capacity or makes worse choices: opportunities, burdens and what a choice is trying to achieve differ. An individual technique cannot supply a resource or option that is structurally unavailable.

The commentary proposes reserving deliberation time, such as 30 minutes daily, when feasible; the suggestion itself consumes time that some people lack. It also calls for policy that reduces burdens—for example, nearby health clinics can reduce the transport and coordination required to obtain care. The two levels complement one another, rather than shifting responsibility for structural obstacles entirely onto individuals.

<a id="b-dabdc44920cb1946"></a>
<a id="u-decisive-c3fa08ddc361"></a>

<a id="v2-117-0"></a>

**Shortform perspective: Cultural Context: Why Decision-Making Books Thrive in the US**

Shortform suggests that the popularity of decision books in the United States may reflect cultural emphasis on choice and independence. It connects Adam Smith’s account of self-interest with Ralph Waldo Emerson’s later writing on self-reliance. Emerson was born in 1803, so his influence should not be placed at the nation’s founding. These are historical influences proposed by the commentary, not a complete causal explanation of a publishing genre.

<a id="b-a2c768f0597dc564"></a>
<a id="u-same-as-ever-4a85d4ce7f47"></a>

<a id="v2-334-0"></a>

**Shortform perspective: Accepting Flaws and Vulnerability in Others**

Accepting the flaws and vulnerabilities of role models can feel alien, because Western social conditioning teaches that to be worthy of love and praise we must “fix” our flaws and vulnerable parts rather than accept and show them. As Tara Brach notes in *Radical Acceptance*, one path toward accepting people’s flaws and vulnerabilities is to recognize them as important, inevitable parts of existence. We are all imperfect; having flaws and vulnerabilities does not mean something is “wrong” with a person—it simply means they are human, like everyone else.

<a id="b-bccf9a58e705e328"></a>
<a id="u-clear-thinking-0893680bb71d"></a>

<a id="v2-337-0"></a>

**Shortform perspective: Two Views of Confidence: Past Success vs. Value Integrity**

In *Dare to Lead*, Brené Brown offers a counterargument to confidence built on past performance: true confidence is not about believing you are likely to succeed. Instead, it comes from trusting in the integrity of your values and principles. That integrity gives you the determination to act in the face of uncertainty or failure, because you know it is the right thing to do regardless of the outcome. By contrast, Parrish’s approach represents a more conditional form of confidence, which may fail to motivate you if you feel your past successes are not convincing enough proof that you can overcome the obstacles ahead.

<a id="b-9072ebe5302cc651"></a>
<a id="u-clear-thinking-b573d73e2a32"></a>
<a id="u-same-as-ever-46d9f8e2661a"></a>

<a id="v2-335-0"></a>

**Shortform perspective: Habit Formation, Bad Habits, and the Compounding Power of Small Changes**

Shortform suggests using social accountability to interrupt a habit—for example, an agreed daily check-in with a partner about soda consumption. The mechanism is making behavior visible and keeping the intention salient. Embarrassment may motivate some people but can also produce secrecy or shame; choose a supportive arrangement rather than treating humiliation as necessary.

Drawing on Atomic Habits, Shortform uses “1% improvements” and a plateau of latent potential to explain why small efforts may precede visible gains. Repetition can build skills, routines and accumulated advantages. It is a compounding metaphor, not a guarantee that every habit improves by a measurable daily percentage or that every plateau eventually turns sharply upward. Track whether the practice is working, use feedback, and adjust the method when it is not.

*Sources and further reading:* [The Subtle Art of Not Giving a F*ck](https://short.zaynjarvis.com/books/the-subtle-art-of-not-giving-a-f-ck/) · [Mistakes Were Made (But Not By Me)](https://www.shortform.com/app/book/mistakes-were-made-but-not-by-me) · [Cognitive dissonance](https://www.shortform.com/app/book/mistakes-were-made-but-not-by-me#what-is-cognitive-dissonance) · [Above-average beliefs](https://www.shortform.com/app/book/mistakes-were-made-but-not-by-me#belief-2-i-am-a-competent-person) · [The Laws of Human Nature](https://www.shortform.com/app/book/the-laws-of-human-nature) · [Short-term threat attention](https://www.shortform.com/app/book/the-laws-of-human-nature/law-12) · [Principles](https://www.shortform.com/app/book/principles-life-and-work) · [Humility and expertise](https://www.shortform.com/app/book/principles-life-and-work/chapter-5#4-balance-humility-with-self-knowledge) · [The Happiness Trap](https://short.zaynjarvis.com/books/the-happiness-trap/) · [Wanting](https://www.shortform.com/app/book/wanting) · [Solitary reflection](https://www.shortform.com/app/book/wanting/1-page-summary#step-1-identify-meaningful-desires) · [Role models and desire](https://www.shortform.com/app/book/wanting#mimetic-desires-negative-effects) · [Four Thousand Weeks](https://www.shortform.com/app/book/four-thousand-weeks) · [Limits of achievement](https://www.shortform.com/app/book/four-thousand-weeks/1-page-summary#fact-2-your-life-is-finite) · [Lost Connections](https://www.shortform.com/app/book/lost-connections) · [Junk values](https://www.shortform.com/app/book/lost-connections/chapter-5#disconnection-from-intrinsic-motivation) · [Scarcity and attention](https://www.apa.org/monitor/2014/02/scarcity) · [Decision-book culture](https://www.shortform.com/best-books/genre/best-decision-making-books-of-all-time) · [American choice traditions](https://www.npr.org/transcripts/519266687) · [Individualism in public documents](https://www.state.gov/courses/answeringdifficultquestions/assets/m/resources/DifficultQuestions-AmericanValues.pdf) · [Acceptance of imperfection](https://www.shortform.com/app/book/radical-acceptance/chapter-1) · [Dare to Lead](https://www.shortform.com/app/book/dare-to-lead) · [Values-based confidence](https://www.shortform.com/app/book/dare-to-lead/chapter-1-5) · [Habit persistence](https://short.zaynjarvis.com/books/the-power-of-habit/)

<!-- atlas:block-content b-57a1685d3a5302c5:end -->
<a id="theme-uncertainty-probability-and-calibration"></a>

## Uncertainty, belief and learning

<!-- atlas:block-content b-06dc0d72caac6463:start -->
### A decision is a bet; an outcome is evidence

Record what was knowable before the choice, then examine the result without pretending it reveals its own cause. Skill, circumstances and luck can interact.

<a id="b-06dc0d72caac6463"></a>
<a id="u-thinking-in-bets-c39bcceafeea"></a>
<a id="u-thinking-in-bets-691c11486cee"></a>
<a id="u-thinking-in-bets-a56e936ebdc7"></a>
<a id="u-thinking-in-bets-a886b53e594c"></a>
<a id="u-thinking-in-bets-6a141db94502"></a>
<a id="u-thinking-in-bets-18fc4bcb6ab3"></a>
<a id="u-thinking-in-bets-e87f3ac49f22"></a>

<a id="v2-42-0"></a>

**Decisions as bets: the core of Duke’s framework**

In *Thinking in Bets*, Annie Duke draws on decades of professional poker experience to refine a decision-making framework applicable to everyday life and business. Poker provided a vivid laboratory: every choice leads to one of two clear outcomes—winning or losing money—and those outcomes act as feedback on the quality of a decision. Her approach aims to help people become more consistently rational and intellectually flexible.

Every decision, no matter how large or small, carries risk. Yet people do not always keep risk at the front of their minds when choosing, and this inattention is one reason they tend to decide in irrational ways: acting impulsively without thinking decisions through, picking the easy or immediately satisfying option even when it harms them in the long run, and letting biases shape their choices. One way to keep risk salient is to treat every decision as a bet.

Duke identifies three core building blocks of “thinking in bets”: well-calibrated beliefs, the ability to learn from past decision outcomes, and a community of people who can help guard against ineffective habits. Together, beliefs, outcomes, and groups form the foundation of her philosophy.

<a id="b-7c4b2fbdfe841da9"></a>
<a id="u-thinking-in-bets-75d75335dd11"></a>
<a id="u-thinking-in-bets-463bac708d19"></a>
<a id="u-thinking-in-bets-ef5584c7a025"></a>
<a id="u-thinking-in-bets-44e214411047"></a>
<a id="u-thinking-in-bets-2d4cbccafd7c"></a>
<a id="u-thinking-in-bets-cad0698bb5ae"></a>
<a id="u-thinking-in-bets-28c54d739431"></a>

<a id="v2-43-0"></a>

**Embracing uncertainty: moving beyond black-and-white judgment**

Improving decision-making starts with getting better at learning from past decisions, and that requires confronting uncertainty head-on. In a bet, you know that even favorable odds do not guarantee a win; betting means acknowledging you do not have all the information and must exercise your best judgment in the face of risk and uncertainty. Decisions work the same way: you weigh the odds and hope for the best result.

Avoid treating decisions as entirely right or wrong. Compare feasible options against your goals, the probabilities and consequences of outcomes, your preferences, and risks you can bear. The option with the highest chance of a narrowly defined success need not have the greatest expected value or utility. A bad outcome can reflect any of five possibilities Duke asks us to consider: bad luck; no good options; missing key information; a potential reward judged worth the risk; or a choice between the best and worst alternatives that seemed reasonable at the time. The result alone cannot tell which explanation is right. A review may still uncover avoidable information gaps, poor reasoning or an unsuitable trade-off. Recording how uncertain you were—and what you knew before choosing—makes that review more informative.

<a id="b-1d12e6d527ab7530"></a>
<a id="u-thinking-in-bets-8eb7d8a87581"></a>
<a id="u-thinking-in-bets-a21460f9314f"></a>
<a id="u-thinking-in-bets-cf2f31e86e0b"></a>
<a id="u-thinking-in-bets-f2cd6d0be736"></a>
<a id="u-thinking-in-bets-cf72519e11e1"></a>
<a id="u-thinking-in-bets-46cd9e815c6c"></a>

<a id="v2-44-0"></a>

**Avoiding resulting: judge the process, not just the outcome**

A common trap when judging past choices is “resulting”—a poker term for the habit of evaluating a decision solely by the outcome it produced. Resulting is dangerous because it can lead you to conclude you must change your strategy after a single bad result, even if that bad outcome came from luck rather than the quality of your decision. In that case, changing strategy will not help over the long run; it will only make you confused and erratic. Good decision-making is not just about getting the best outcome: it is about having a sound decision process that holds up regardless of the final result.

Before you blame a negative outcome on poor decision-making, analyze the factors that led you to your choice. You cannot control your luck, but you can control your skill and the soundness of your thinking. To check whether your process was solid, ask yourself whether you considered alternatives, took steps to lower your risk, and thought through all possible outcomes. Being a good decision-maker means staying rational in the face of losses. You will not always be right, but you can always work toward objectivity and away from emotional or biased decision-making.

<a id="b-4a8f1ae9f3401f5d"></a>
<a id="u-thinking-in-bets-a574dcbf21bf"></a>
<a id="u-thinking-in-bets-d71232154651"></a>
<a id="u-thinking-in-bets-69c0cd129cf9"></a>
<a id="u-thinking-in-bets-16e10b46711c"></a>

<a id="v2-45-0"></a>

**Outcome fielding: separating luck from skill on a spectrum**

Before you can learn from an outcome, you have to determine how much of it is due to skill and how much to luck. If an outcome comes entirely from forces outside your control, such as luck, it may have nothing to teach you. Most outcomes, however, are the product of a mix of factors, some of which you can control, like skill. Outcomes can be thought of as lying on a spectrum between pure luck and pure skill, with the majority falling somewhere in the middle.

Annie Duke calls this sorting process “outcome fielding.” For example, if an actor cannot land any jobs, he has to figure out whether the problem is a lack of training—a skill issue—or simply not having found the right role yet—a luck issue—or a little of both. To advance his career, he must learn to tell those factors apart and address them accordingly. How you sort an outcome is itself a kind of bet, just like your decisions and beliefs. If you attribute an outcome to skill, you are betting you can learn from it and adjust future decisions—other bets—in a way that will shape future results, setting off a chain reaction. That is why it matters to get better at outcome fielding and get it right as often as you can.

<a id="b-3de3c0d13919a15f"></a>
<a id="u-clear-thinking-fd769d3bdbdd"></a>

<a id="v2-330-0"></a>

**Learning From Results: Disentangling Skill From Luck**

Annie Duke’s *Thinking in Bets* offers a complementary perspective on learning from results, clarifying that it is not impossible to learn from outcomes—the key skill is determining which results you can learn from and which you cannot. When evaluating a decision’s outcome, people construct a narrative: they attribute the result primarily either to their own choices and actions, or to luck. The reason many fail to improve their rational decision-making is that they confuse one category with the other. For example, if the CEO of an online clothing retailer adds a new line of retro goth clothing and then sees record profits driven by an unrelated industry-wide boom in online fashion, and they credit those profits to their shrewd bet on goth nostalgia rather than luck, they will never realize their understanding of customer demand is distorted. Parrish’s advice to keep a written record of your thinking is one way to more accurately assess how much of an outcome came from your actions versus luck. Duke recommends a second approach: learn from the results of *other people’s* decisions. This is difficult because ego biases you to attribute others’ successes to luck and their failures to bad decisions; you can counteract this by consciously accounting for the bias.

<a id="b-641a40f6dcc4006d"></a>
<a id="u-clear-thinking-679311f5dd95"></a>
<a id="u-clear-thinking-a75880fe8f2b"></a>

<a id="v2-329-0"></a>

**Learning From Mistakes by Auditing Your Decision Process**

To learn effectively from mistakes, you first need to accurately identify when you have made one. Parrish warns that most people judge their decisions purely by their results, which makes it impossible to tell whether a decision was wise or merely lucky. Instead of focusing on outcomes, you should look for mistakes in *how* you made the decision—for instance, whether your logic was flawed or your emotions impaired your thinking. The goal is to build a consistent, well-informed, rational decision-making process, even if it does not succeed every time. The best way to identify flaws in your decision-making process, according to Parrish, is to keep a tangible record of your thoughts while making a decision, rather than relying on memory. This prevents your ego from retroactively convincing you that you were smarter or more rational than you actually were at the time.

<a id="b-400cb155e4878eb0"></a>
<a id="u-same-as-ever-4bb3b2e72078"></a>

<a id="v2-332-0"></a>

**Cognitive Biases and the Role of Experience in Judgment**

The mismatch between how we think we will act and how we actually react often stems from mental shortcuts that produce cognitive biases. In *Thinking, Fast and Slow*, Kahneman explains that our minds frequently rely on these shortcuts, and his findings support the view that people’s judgments are deeply shaped by personal experience and emotional state. For example, the planning fallacy affects thinking before you confront actual experience: we underestimate how long tasks will take and overestimate our ability to handle future situations because we rely on best-case scenarios we construct in our minds. To avoid this fallacy, Kahneman suggests studying past outcomes of similar situations—for instance, you might think it takes 10 minutes to get ready in the morning, but that assumes you laid out your outfit the night before and no obstacles arise. By contrast, anchoring affects thinking after you encounter a significant experience: when you meet new information, your mind often latches onto it as a reference point even if it is not relevant. Memories and past experiences anchor how you interpret and respond to new information, shaping how you judge risks, assess value, and react emotionally. As a result, major events from your past can bias current decisions even when those experiences have little to do with the situation at hand. Since we are all prone to such misjudgment, this lends support to the suggestion to have empathy for ourselves and others: seemingly irrational or different behaviors often grow out of unique experiences and cognitive biases, not mere misinformation or irrationality.

*Sources and further reading:* [Thinking in Bets](https://short.zaynjarvis.com/books/thinking-in-bets/) · [Experience and cognitive bias](https://short.zaynjarvis.com/books/thinking-fast-and-slow/#cognitive-biases-and-heuristics)

<!-- atlas:block-content b-06dc0d72caac6463:end -->
<!-- atlas:block-content b-254f5635f7f1f132:start -->
### State uncertainty in a form that can be checked

A subjective range, a confidence procedure and a probability forecast are different objects. Keep the event, time period and uncertainty being described explicit.

<a id="v2-48-1"></a>

**Subjective ranges and calibration**

Hubbard asks estimators to express uncertainty as a range with a stated confidence level: for example, “I assign a 90% probability to this project costing between $40 million and $60 million.” This is an elicited subjective interval, even though the source calls it a confidence interval. It communicates a probability judgment and bounds, not new measured evidence by itself. A statistical confidence interval is different: its stated coverage concerns how often the interval-producing procedure would contain the fixed parameter in repeated sampling under its assumptions. Neither requires every outcome to follow a bell curve.

Calibration training compares confidence with eventual correctness. A trainer asks questions, collects probability estimates or intervals, reveals the answers, and uses feedback over many questions to improve judgment. In a 20-question exercise with 90% intervals, 12 containing the answer means 60% observed coverage; 16 means 80%, still below the target of about 18. One small round is noisy, and merely widening everything is not the goal: useful calibration combines realistic coverage with informative ranges.

The summary’s examples include placing Microsoft’s founding between 1970 and 1980—the answer is 1975—and estimating Amazon’s workforce at 500,000–800,000 when the dated comparison used more than 1.5 million. The latter must be tied to the date and workforce definition being tested. Hundreds of practice questions can reveal recurring overconfidence, but they do not guarantee 90% coverage in every new domain.

<a id="b-254f5635f7f1f132"></a>
<a id="u-how-to-measure-anything-66a2db7ded36"></a>
<a id="u-how-to-measure-anything-1eb77e131e20"></a>

<a id="v2-49-0"></a>

**Uncertainty Is Inherent in All Measurement, Even Everyday Ones**

Even seemingly straightforward measurements, such as the length of a couch, contain some degree of uncertainty, however minimal. When you measure a couch with a tape measure, small margins of error come from the precision of the tool, slight variations in how you position the tape, or even temperature effects on the materials. For practical purposes this uncertainty is often negligible—you might be confident the couch is between 84.8 and 85.2 inches long rather than exactly 85 inches—but the principle remains: even “certain” measurements deal in probabilities and ranges, just very narrow ones.

A range is only as credible as the model or judgment behind it. Normal distributions can be useful, but neither uncertainty nor interval estimation requires normality. Shortform uses investor Jamie Mai to illustrate model mismatch: a tobacco company facing major litigation might have sharply different outcomes if it wins or loses, rather than a concentration around one middle result. In the account, options looked cheap because the market underpriced large moves. This is a claim about a particular distribution and price, not a general proof that options markets always assume bell curves. The litigation example also simplifies possible settlements and intermediate outcomes; identify those before applying it to a real decision.

<a id="b-fa8994dfd235b2f3"></a>
<a id="u-how-to-measure-anything-e829b3719f31"></a>

<a id="v2-50-0"></a>

**Shortform perspective: Calibration Training Has Conditional Benefits and Clear Limits**

The study discussed by Shortform reports that feedback comparing answers with correct responses improved some trainees’ self-assessment, and some benefits persisted on new material after feedback was withdrawn. That sequence suggests transfer after training; it does not show that the earlier feedback was unnecessary or that novices learn equally well without it.

Furthermore, the research revealed that when trainees were asked to predict their performance before completing tasks rather than judging it afterward, the training sometimes backfired: medium and high performers became overly cautious and underestimated their abilities, suggesting different types of self-judgment respond differently to this kind of training.

Within those tasks, the results favor clearer comparison standards and attention to trainees’ initial self-assessment ability; judging a completed performance differs from predicting a future one. These findings do not establish the universally most effective training method or automatic transfer to unrelated decisions.

<a id="b-01593c849fc33f87"></a>
<a id="u-same-as-ever-daeabd984d36"></a>

<a id="v2-338-0"></a>

**Shortform perspective: Probabilistic Thinking and the “Wrong-Side-of-Maybe” Fallacy**

Gardner and Tetlock’s “wrong-side-of-maybe” fallacy treats above 50% as a guarantee and below 50% as impossibility. A well-calibrated collection of comparable 70% rain forecasts should be followed by rain about 70% of the time in their specified places and periods; one dry outcome does not by itself refute such a forecast. Calibration requires many resolved predictions, while the quality of a single forecast also depends on the information available when it was made.

<a id="b-f1c7d765d14214d7"></a>
<a id="u-algorithms-to-live-by-b3b9fa4237ef"></a>
<a id="u-algorithms-to-live-by-d7813eda2842"></a>
<a id="u-algorithms-to-live-by-e4c60fd4c14b"></a>

<a id="v2-215-0"></a>

**Bayesian prediction: prior odds times the likelihood ratio**

Christian and Griffiths describe prediction as combining prior knowledge with new evidence. Bayes’s rule supplies the missing operation: posterior probability is proportional to prior probability multiplied by the likelihood of the evidence under each hypothesis, then normalized. For two hypotheses, posterior odds equal prior odds times the likelihood ratio.

For an illustrative calculation, let a hypothesis have prior probability 20%, or odds 1:4. If a piece of evidence is three times as likely under that hypothesis as under its alternative, posterior odds become 3:4 and probability becomes 3/7, about 43%. This is an editorial numerical illustration, not a result reported by the books.

In the source’s pay-raise example, a colleague’s waiting time provides a starting comparison, and credible information about your performance and the employer’s policy can update it. One colleague is a limited reference class, and your impression of your boss’s view is not a measured likelihood. Repeated reporting of the same news must not be counted as fresh independent evidence.

<a id="b-987733591bc07ce3"></a>
<a id="u-algorithms-to-live-by-73fc90fdd6cd"></a>

<a id="v2-225-0"></a>

**Shortform perspective: Update in proportion to the evidence**

Shortform draws on Tetlock and Gardner’s Superforecasting to warn against overreacting to weak new information. Strong forecasters often make measured updates; a good-performance story, for example, may not justify expecting an immediate raise when pay policy is unchanged. But Bayesian updating does not require always making tiny changes or ignoring most evidence. Discard irrelevant and redundant signals, weigh evidence by its diagnostic value, and make a large revision when strong evidence warrants one. The prior and likelihood model can both be wrong, so calibration and later outcomes remain important.

<a id="b-d35af523f9b8f613"></a>
<a id="u-science-fictions-3ad946bb6a63"></a>

<a id="v2-288-0"></a>

**Shortform perspective: Applying Bayesian Reasoning to Everyday Choices**

You can apply Bayesian reasoning—updating beliefs as new evidence arrives—to everyday decisions such as financial planning. Before investing, assess the initial probability of success based on current evidence about the market or a specific stock. As new information arrives—such as quarterly reports or market shifts—update your probability estimate to reflect the latest likelihood of success. This approach helps you make more informed decisions that account for uncertainty and change.

<a id="b-fe618f4a08459da7"></a>
<a id="u-same-as-ever-0110141c6070"></a>

<a id="v2-340-0"></a>

**Shortform perspective: How Superforecasters Approach Predictions**

Shortform notes that a separate group of experts known as superforecasters take a different approach to making confident future predictions. In *Superforecasting*, Dan Gardner and Philip Tetlock explain that rather than relying on fixed rules about human behavior and society, superforecasters use analytical skills. For instance, they break seemingly intractable questions into progressively smaller sub-questions, which lets them separate truly unknown matters from those on which they can make educated guesses. They also cultivate objectivity by gathering information from every possible source, letting the data shape their interpretation, and remaining willing to revise their predictions when new information emerges.

*Sources and further reading:* [Measurement uncertainty](https://www.webassign.net/question_assets/unccolphysmechl1/measurements/manual.html) · [Normal-distribution model](https://www.investopedia.com/terms/b/bell-curve.asp) · [Non-normal outcomes](https://macro-ops.com/jamie-mais-trading-strategy-explained/) · [Options-market example](https://www.investopedia.com/options-basics-tutorial-4583012) · [Calibration-training limits](https://onlinelibrary.wiley.com/doi/full/10.1002/acp.3548) · [Superforecasting](https://www.shortform.com/app/book/superforecasting/) · [Wrong-side-of-maybe fallacy](https://www.shortform.com/app/book/superforecasting/chapter-3#probabilities-are-useful-estimates-not-facts) · [Bayesian prediction](https://www.shortform.com/app/book/superforecasting/chapter-7) · [Superforecasting](https://www.shortform.com/app/book/superforecasting) · [Analytical forecasting skills](https://www.shortform.com/app/book/superforecasting#part-2-traits-of-superforecasters)

<!-- atlas:block-content b-254f5635f7f1f132:end -->
<a id="theme-evidence-measurement-and-causality"></a>

## Measurement and causal evidence

<!-- atlas:block-content b-323612a4aebd14fe:start -->
### Measure what could change the choice

Choose the decision first, identify its threshold, and spend measurement effort on uncertainty capable of moving you across that threshold.

<a id="b-323612a4aebd14fe"></a>
<a id="u-how-to-measure-anything-3ec7f733a336"></a>
<a id="u-how-to-measure-anything-5ee83a564696"></a>
<a id="u-how-to-measure-anything-6811131260f9"></a>
<a id="u-how-to-measure-anything-b3e03ce19a61"></a>
<a id="u-how-to-measure-anything-e318fa32a6d0"></a>
<a id="u-how-to-measure-anything-50449189383a"></a>
<a id="u-how-to-measure-anything-d1439a8f0946"></a>
<a id="u-how-to-measure-anything-bc8c76b81eed"></a>
<a id="u-how-to-measure-anything-e5513559b537"></a>
<a id="u-how-to-measure-anything-f3ac99c892f5"></a>
<a id="u-how-to-measure-anything-f8f8153e6863"></a>
<a id="u-how-to-measure-anything-79cf99d60566"></a>
<a id="u-how-to-measure-anything-428bac9593a3"></a>
<a id="u-how-to-measure-anything-30cd624d5ce8"></a>
<a id="u-how-to-measure-anything-3b012c8fafc1"></a>
<a id="u-how-to-measure-anything-7eba7c5ada19"></a>
<a id="u-how-to-measure-anything-5fc8dd8d1225"></a>
<a id="u-how-to-measure-anything-b05be811f61d"></a>
<a id="u-seeking-wisdom-bevelin-f083760e5d96"></a>
<a id="u-seeking-wisdom-bevelin-7c8d91d472a9"></a>
<a id="u-seeking-wisdom-bevelin-2662bd515dc4"></a>
<a id="u-seeking-wisdom-bevelin-bc2147f02fd3"></a>
<a id="u-science-fictions-04ad6c553517"></a>
<a id="u-science-fictions-8055521d8b7a"></a>
<a id="u-seeking-wisdom-bevelin-f04a55b4e81a"></a>
<a id="u-seeking-wisdom-bevelin-da38ecea8a28"></a>
<a id="u-seeking-wisdom-bevelin-2920cdc48afd"></a>
<a id="b-3db4e6890a38959f"></a>
<a id="u-how-to-measure-anything-acb9b92056de"></a>
<a id="u-how-to-measure-anything-1fec1a76398f"></a>
<a id="u-how-to-measure-anything-32f7a2a109e1"></a>
<a id="u-how-to-measure-anything-06d3d40cbf94"></a>
<a id="b-2b06eaefee342a0d"></a>
<a id="u-seeking-wisdom-bevelin-559bb675a3d8"></a>
<a id="u-seeking-wisdom-bevelin-aaf82c19b2a4"></a>
<a id="b-6cf4dbd5ef7058a9"></a>
<a id="u-seeking-wisdom-bevelin-b2873a213361"></a>
<a id="u-seeking-wisdom-bevelin-2a8640385258"></a>

<a id="v2-52-0"></a>

**Measurement reduces decision-relevant uncertainty**

Hubbard defines measurement as observations that reduce uncertainty, expressed through values, ranges or probabilities rather than labels such as “high risk” or “good quality.” The purpose is to improve a specific choice, not to know everything. Even modest information can be valuable when it changes customer-satisfaction, market-risk or investment decisions. More measurement is not automatically better: irrelevant measures, biased data and collection costs can outweigh the benefit. Although the principles apply beyond business, Hubbard’s main audience is business decision-makers.

<a id="b-a08a5f680c1d2670"></a>
<a id="u-how-to-measure-anything-e67cefd0d649"></a>
<a id="u-how-to-measure-anything-4dfa58a91293"></a>
<a id="u-how-to-measure-anything-7afb453dcec6"></a>
<a id="u-how-to-measure-anything-65289fe12d18"></a>

<a id="v2-53-0"></a>

**Defining Measurement Scope and Its Three Determining Factors**

Defining measurement scope—what will and won’t be included—is critical to measurement success. A scope that is too broad leads to excessive data that yields no meaningful insights; a scope that is too narrow risks missing important factors and failing to address the problem at hand. Hubbard illustrates this with a customer satisfaction example: if a company measuring satisfaction to reduce churn defines scope too broadly, it surveys customers about everything and cannot identify what drives attrition; if it defines scope too narrowly by using only product ratings, it may miss that poor service is the cause. The right scope focuses on the most relevant touchpoints—support interactions, billing processes, and core product functionality—providing actionable insights without irrelevant data.

According to Hubbard, three factors shape measurement scope: 1) your current state of knowledge, 2) what you plan to do with the information, and 3) the decision’s impact potential.

<a id="b-42c9a3eac7baacda"></a>
<a id="u-how-to-measure-anything-bd209b603f5e"></a>
<a id="u-how-to-measure-anything-5c9df6d4fdc5"></a>
<a id="u-how-to-measure-anything-95b72154d173"></a>
<a id="u-how-to-measure-anything-af16bbb7d3d1"></a>
<a id="u-how-to-measure-anything-93280a1ffa35"></a>
<a id="u-how-to-measure-anything-abf62f5161f5"></a>
<a id="u-how-to-measure-anything-4deac1c5a6ba"></a>

<a id="v2-54-0"></a>

**Current Knowledge and Decomposition as Scope-Framing Tools**

Assessing your current state of knowledge is a critical first step in defining measurement scope: it reveals where uncertainty lies and which knowledge gaps matter most for your decisions, so you can direct measurement efforts toward where they will deliver the greatest value. For example, a restaurant owner trying to improve profits might separate what she already knows (food costs, staffing expenses) from what she is uncertain about (optimal pricing), then focus on targeted measurements such as customer feedback and repeat visits rather than trying to measure everything at once.

Hubbard identifies decomposition as a tool for assessing what you already know: you break complex, hard-to-measure concepts into smaller components. This helps refine measurement scope by showing which components truly require new measurements and which can be handled adequately with existing knowledge. Decomposition also makes measurement more feasible, because smaller components can usually be measured more easily or estimated more precisely than the complex whole. A marketing director overwhelmed by predicting a new campaign’s success can split “campaign success” into parts: reach, interaction, purchase rate, and long-term customer spend. This process immediately shows her that she already has reliable historical data on reach and solid benchmarks for click and engagement rates across channels, while her highest uncertainty lies in purchase rate for this specific product and long-term customer value. She can therefore narrow her measurement scope to those high-uncertainty components, which are also more concrete and actionable to track than a vague “overall campaign success” metric.

<a id="b-483d3036bed7d901"></a>
<a id="u-how-to-measure-anything-5bb799f41f9f"></a>
<a id="u-how-to-measure-anything-66190ac2a8dd"></a>
<a id="u-how-to-measure-anything-8e19e45fd86f"></a>

<a id="v2-55-0"></a>

**Decision-Driven Measurement and the Inflection Point Principle**

The second scope factor—what you plan to do with your measurement—is central to Hubbard’s approach. Measurements should be taken to inform a specific decision with multiple options, and every measurement should have a clear inflection point: a specific tipping point at which the collected data would cause you to choose one option over another. Without grounding measurements in specific decisions, you risk spending significant time and money collecting purposeless data. For example, a product team redesigning a website defines its decision clearly: “Should we implement a single-page checkout or maintain our multi-page process?” This establishes an inflection point—if the single-page checkout shows a 15% or greater relative improvement in completion rates (for example, 40% to 46%, not 40% to 55%), they will switch—and focuses their measurement exclusively on checkout-specific metrics such as abandonment rates, completion times, error rates, and checkout-related satisfaction scores. This relative-change convention clarifies the teaching example; the source did not specify which meaning of “15%” it intended. A real team should specify relative change versus percentage points in advance and retain necessary cost, revenue and experience constraints.

<a id="b-859fcaf0c04623b1"></a>
<a id="u-how-to-measure-anything-910987caa8e0"></a>

<a id="v2-56-0"></a>

**Shortform perspective: Defining Measurement Scope With OKRs**

In *Measure What Matters*, venture capitalist John Doerr presents a system for identifying what matters at a company and measuring its success, beginning with defining measurement scope by clarifying OKRs—objectives and key results. Objectives are what you aim to achieve; good objectives are tangible, concrete, objective, and action-oriented. To identify your organization’s highest objectives, ask: “What are the most important tasks we need to accomplish in the next three months, the next six months, and the next year?” Objectives alone, however, are often too broad and vague to measure effectively. For example, an objective like “Hire new employees to meet the needs of the expanding organization” does not define what success looks like or how to track progress.

Key results turn vague objectives into measurable benchmarks. Good key results state specific, quantifiable sub-goals that, if achieved, will deliver the objective. They provide concrete metrics to track, deadlines to meet, and clear definitions of success, typically including specific timeframes and numbers such as revenue, market share, active users, growth, and hiring targets. For the hiring objective, instead of the vague goal of “hire new employees,” key results might define specific benchmarks such as “Meet with three candidates this month and hire one.” This quantifies exactly how many interviews are needed, sets a clear deadline, and enables progress tracking—how many interviews have been completed, whether the hire was made within the month. Without measurable terms, there is no objective way to assess progress toward a goal, or even to define what the goal truly means.

<a id="v2-52-1"></a>

**Decision Impact Potential and the Unequal Value of Uncertainties**

The third scope factor is decision impact potential: the degree to which reducing uncertainty about a specific factor would change your decision. Hubbard’s core insight is that not all uncertainties are created equal: some gaps in knowledge have enormous impact on your decisions, while others do not. High decision impact potential means better data about that factor could significantly shift you toward one choice or another; low decision impact potential means even perfect information about that factor would not meaningfully alter your decision. You are not looking for perfect knowledge; you are looking for enough information to move from indecision to well-informed confidence. For example, a health care administrator deciding whether to invest $200,000 in a new patient scheduling system is already 85% confident about patient satisfaction problems and 90% certain about staff preferences, but only 40% confident that the system would reduce no-shows and wait times—the core benefits justifying the investment. Since this uncertainty has the biggest impact on her decision, she focuses her research budget on that specific question rather than on areas where she is already reasonably confident.

<a id="b-82f9c86e83292d5a"></a>
<a id="u-how-to-measure-anything-71624b41c40e"></a>

<a id="v2-57-0"></a>

**Shortform perspective: Decision Impact Potential: Finding Signals in Noise**

Identifying which data points will most reduce uncertainty is not always intuitive—particularly in financial markets, where the decision impact potential of seemingly unrelated pieces of information can be extraordinarily high, but buried within vast oceans of market noise. In *The Man Who Solved the Market*, business journalist Gregory Zuckerman tells the story of Jim Simons, a former mathematician who became one of the most successful hedge fund managers in history. The key to the success of Simons’s fund, Renaissance Technologies, was his insight that price fluctuations in financial markets follow recognizable and predictable patterns; once identified, these patterns could be used to strategically buy and sell the right stocks, bonds, currencies, and other financial instruments at the right time.

Where most investors focused on obvious financial metrics and company fundamentals, Simons recognized that the highest decision impact potential might lie in subtle correlations hidden deep within historical data patterns. Simons and his colleagues analyzed historical data sets to see how prices of different financial instruments moved in the past, and based on the patterns they found, they could identify which price movements were correlated with others. In other words, within the maze of data, they could find market “signals” indicating what the market would do next. For example, their model might show that under certain conditions, a 0.25% increase in the price of wheat is historically correlated with a 3% surge in the price of sesame seeds three months later. In this case, the market signal—the increase in wheat prices—indicates a specific action: buying sesame seed futures in anticipation of a price surge. This approach mirrors Hubbard’s core insight: you are not looking for all possible information, but for the specific data that moves you from uncertainty to informed confidence.

<a id="b-98ac0fae4184ac49"></a>
<a id="u-how-to-measure-anything-666ec81ef534"></a>

<a id="v2-74-0"></a>

**Shortform perspective: Basic Research as a Counterexample to the Inflection-Point View of Measurement**

Shortform pushes back on Hubbard’s claim that all measurements should have an inflection point tied to a choice between alternatives. This is not true of basic research, which involves the investigation of phenomena simply to understand how things work. Researchers may study newly identified biological structures, unexplained natural occurrences, or poorly comprehended mechanisms without any clear endpoint in mind. They are not trying to solve a specific problem or choose between alternatives—they are exploring questions driven by curiosity about the natural world. Any value from such research, if it materializes at all, emerges unpredictably and often much later.

<a id="b-90752f605792e573"></a>
<a id="u-how-to-measure-anything-3594909187e1"></a>
<a id="u-how-to-measure-anything-58d85aa3c049"></a>

<a id="v2-70-0"></a>

**Shortform perspective: The Epistemology of Measurement: Uncertainty, Objectivity, and Intuition**

Shortform contextualizes Hubbard’s position against broader philosophical debates. Whereas Hubbard writes that absolute certainty is impossible, some philosophical traditions go further: in *Cynical Theories*, Helen Pluckrose and James Lindsay describe postmodernism as the rejection of the very idea that objective truth exists. Hubbard’s approach to measurement offers a practical middle ground between blind faith in objectivity and total skepticism toward truth: while we may never achieve perfect objectivity, we can still use observation to make better decisions than we would by relying on intuition alone. Even if our measurements contain some bias or uncertainty, they often provide more reliable information than subjective assessments.

Although Hubbard argues a measurement-based approach leads to more rational decisions, some experts contend the most effective decision-making combines quantitative analysis with human intuition. Successful executives recognize that raw numbers cannot capture every nuance of complex business situations, particularly when dealing with ambiguous or abstract challenges. Human intuition functions as a sophisticated pattern-recognition system that becomes especially valuable in high-stakes, uncertain scenarios where additional data collection will not meaningfully influence the outcome. Research on surgeons making emergency decisions and investors allocating startup capital shows that “gut feelings” often provide the catalyst needed to move forward when extensive analysis creates indecision. Rather than viewing intuition as unscientific, expert decision-makers understand it as a form of unconscious intelligence that complements analytical thinking.

<a id="b-4d78848b8b85c8ed"></a>
<a id="u-how-to-measure-anything-ab8c8af0d6cf"></a>
<a id="u-seeking-wisdom-bevelin-280962fb2cac"></a>

<a id="v2-72-0"></a>

**Shortform perspective: Cognitive Biases in Assessing What You Know**

One practical challenge in applying Hubbard’s advice to assess your current state of knowledge is the Dunning-Kruger effect—a cognitive bias that leads people to be irrationally overconfident in their skills and competence. In a measurement context, this bias can cause you to mistakenly believe you already understand what you are trying to measure well enough. Instead of honestly evaluating gaps in your knowledge—such as uncertainty about market conditions, employee performance, or customer satisfaction—you may assume your current understanding is adequate and jump straight to measurement or decision-making. As a result, you might measure the wrong things or ask the wrong questions.

Marketers also deliberately exploit people's tendency to rely on case evidence rather than stronger representative evidence that might contradict their message. Marketing experts point out that sharing testimonials from satisfied customers can boost sales, because new customers are swayed by others' success stories—even when those testimonials are not representative of the average customer experience.

*Sources and further reading:* [Bitcoin investor-loss estimate](https://gizmodo.com/bitcoin-crypto-bank-of-international-settlements-1849784466) · [Measure What Matters](https://www.shortform.com/app/book/measure-what-matters/1-page-summary) · [Objectives and key results](https://www.shortform.com/app/book/measure-what-matters/1-page-summary#what-are-okrs) · [The Man Who Solved the Market](https://www.shortform.com/app/book/the-man-who-solved-the-market/1-page-summary) · [Historical market patterns](https://www.shortform.com/app/book/the-man-who-solved-the-market/1-page-summary#finding-signals-in-the-noise) · [Basic research](https://nap.nationalacademies.org/read/10733/chapter/9) · [Cynical Theories](https://www.shortform.com/app/book/cynical-theories) · [Debates about objective truth](https://www.shortform.com/app/book/cynical-theories#phase-2-postmodern-deconstruction) · [Analysis and intuition](https://www.harvardbusiness.org/insight/data-and-intuition-good-decisions-need-both/) · [Dunning–Kruger effect](https://www.psychologytoday.com/us/basics/dunning-kruger-effect) · [Testimonials and case evidence](https://cxl.com/blog/is-social-proof-really-that-important/#h-what-is-social-proof)

<!-- atlas:block-content b-323612a4aebd14fe:end -->
<!-- atlas:block-content b-0593b6b3db95ce08:start -->
### Sampling precision, representation and significance

A useful small sample is not a substitute for every larger sample. Precision, bias and the cost of being wrong answer different questions.

<a id="v2-52-2"></a>

**Random Sampling, Small Samples, and Evidence Representativeness**

Random sampling is a measurement technique that can significantly reduce uncertainty: you examine a small, randomly selected portion of a larger group to draw conclusions about the whole. It is especially useful for measuring ongoing behaviors, activities, or characteristics across large populations, such as the share of work time employees spend on different tasks, how often customers encounter specific problems, or what proportion of inventory has quality issues. For example, a retail manager who wants to measure satisfaction among 10,000 monthly customers can assign each customer a number and use a random number generator to select 200 customers for a survey, giving every customer an equal chance of being chosen and avoiding biases such as only surveying weekend shoppers. With this random sample of just 2% of customers, the manager can estimate overall satisfaction with reasonable confidence at a fraction of the cost of a full survey.

Hubbard challenges the belief that only huge samples can teach us anything. In the summary’s illustration, a chain with 10,000 customers draws seven random responses with scores between six and nine out of ten and a mean of 7.5; that mean requires a scale permitting fractional responses. A later hypothetical sample of 700 has a mean of 7.6 and the same range. Those stipulated results suggest similar average satisfaction; seven answers alone do not establish that all customers are generally satisfied or that the next 693 will add little. Under independent, identically distributed sampling with finite variance, 100 times as many observations reduce the standard error of the mean to one tenth. Whether that extra precision changes the action is a separate value-of-information question.

Bevelin distinguishes a vivid individual case from evidence representing the population of interest. A colleague’s profitable bitcoin investment tells us what happened to that colleague, not the odds facing a new investor. Shortform contrasts it with a cited estimate that three in four new bitcoin investors lost money on their initial investment. That estimate belongs to its particular period, sample and assumptions; it is not a timeless probability. A larger study is more informative only insofar as its selection, measurements and target population suit the question. This reconciles Bevelin’s warning about anecdotes with Hubbard’s point that a small well-chosen sample can still be useful.

<a id="b-0593b6b3db95ce08"></a>
<a id="u-how-to-measure-anything-a1baaa5daf3b"></a>
<a id="u-how-to-measure-anything-9fdca1659af4"></a>
<a id="u-how-to-measure-anything-98926709c906"></a>
<a id="u-how-to-measure-anything-819f0f2929ae"></a>

<a id="v2-60-0"></a>

**Small samples can be useful; larger samples can still matter greatly**

Simple random sampling gives each member of the population an equal chance of selection. Probability sampling can also use unequal, known selection probabilities with appropriate weighting. The problem is not unequal probabilities by themselves, but ignoring the selection mechanism, nonresponse, coverage gaps, or other systematic distortions. Increasing sample size does not automatically repair those biases.

Hubbard’s decision-useful point is that six or seven observations can be worth collecting: moving from no evidence to some evidence may already change an action. That does not mean moving from seven to 700 adds little precision. For independent observations from the same distribution with finite variance, the standard error of a sample mean scales as 1/√n; increasing n one hundredfold reduces that sampling error by a factor of ten. Larger samples also help estimate small effects, rare events, subgroups, and heterogeneous populations. Whether that improvement is worth its cost depends on the decision and its threshold, not on a universal sample-size rule.

<a id="b-089dd31b7b217282"></a>
<a id="u-how-to-measure-anything-033a3de7c357"></a>

<a id="v2-61-0"></a>

**Stratification protects relevant subgroups**

Darrell Huff, in Shortform’s supplementary discussion, stresses the practical difficulties of obtaining representative samples. Those difficulties do not make true random selection impossible. Stratified random sampling is itself a probability method: divide the population into nonoverlapping groups covering the target population, then sample randomly within each. Age and race categories must be crossed or otherwise defined so they do not count the same person twice.

For three factory lines producing 5,000, 2,000 and 1,000 units a month, proportional samples of 50, 20 and 10 ensure that all lines are represented. An unstratified sample may include too few observations from a small line to study it well. Stratification is useful for customers, employees, products or transactions when subgroup differences matter; the sampling and weighting must match the overall or subgroup quantity being estimated.

<a id="b-1d100fe281e6ba68"></a>
<a id="u-how-to-measure-anything-ff05c3f5e877"></a>
<a id="u-how-to-measure-anything-af92db7cc009"></a>

<a id="v2-58-0"></a>

**Statistical significance and its limited role in practical decision-making**

A significance test asks how compatible the observed data are with a specified null model. A p-value below 0.05 means that, **if the null hypothesis and the model's assumptions were true**, results at least as extreme as those observed would occur less than 5% of the time. It is not the probability that the finding is a fluke, is not the probability that the hypothesis is true, and cannot prove a theory. This corrects an inverse-probability formulation inherited from the frozen summary, following the [American Statistical Association's statement on p-values](https://www.amstat.org/asa/files/pdfs/p-valuestatement.pdf).

Hubbard's practical point remains useful: business measurement aims to reduce decision-relevant uncertainty, not merely to clear an academic threshold. If you do not know whether a product task takes five minutes or two hours, and a small test puts it near 30 minutes, that narrower estimate can improve the decision relative to complete ignorance even when it does not meet a conventional significance threshold.

<a id="b-6e689a89d151c5d4"></a>
<a id="u-how-to-measure-anything-2c21223a5eb3"></a>
<a id="u-how-to-measure-anything-a3c53a8c2489"></a>
<a id="u-how-to-measure-anything-3c9032862ab2"></a>

<a id="v2-59-0"></a>

**Statistical Significance and Sample Size: From Fisher’s Tea Test to Business Applications**

Fisher’s famous tea-tasting example makes a small experiment concrete. A taster identifies which four of eight randomly ordered cups had milk poured first. Under uniform random guessing among the 70 possible groups of four, all eight labels are correct with probability 1/70. Getting at least six labels right has probability 17/70, roughly one quarter; one swapped pair changes two labels, not one. These are probabilities under the guessing model, not probabilities that the taster has no ability. The example illustrates randomization and evidence strength, rather than establishing that statistical significance originated in this single event.

Some statistics writers take a similar view that larger samples are better for reducing uncertainty. In *Naked Statistics*, Charles Wheelan writes that the larger a sample, the more likely it is to represent the underlying population, both because a larger sample has more chance of capturing diversity and because it reduces the influence of outliers. As a general rule, bigger samples produce more reliable statistics—but this rule only holds for a truly random sample; a biased sample yields biased results no matter its size.

Kahneman’s kidney-cancer example illustrates the “law of small numbers”: small populations can produce more variable rates. Some rural US counties appeared among both the lowest and highest recorded rates. That invites opposing stories—fresh air and unadulterated food protect rural residents, or poverty and fatty diets endanger them—but the shared small denominator is another explanation to examine. A few additional cases can move a small county’s rate substantially, so extremes alone cannot identify a rural lifestyle effect. This sampling mechanism does not rule out real regional risk differences or guarantee that every large city lies near one common “true average.”

Shortform also notes that, despite Hubbard's skepticism about the value of statistical significance for most businesses, some experts argue it is important for sound business decision-making. Knowing whether observed results reflect genuine effects or random chance prevents companies from making costly mistakes by acting on misleading patterns that do not represent true underlying relationships. Businesses that can confidently identify real patterns in their data can forecast customer behavior, market trends, and operational outcomes more accurately, and allocate resources more effectively. These experts add that statistical significance is especially relevant today, as modern businesses generate vast amounts of data from many sources; organizations that understand statistical significance can better evaluate the reliability of their analytics and communicate findings more clearly to stakeholders.

*Sources and further reading:* [How to Lie With Statistics](https://www.shortform.com/app/book/how-to-lie-with-statistics) · [Stratified random sampling](https://www.shortform.com/app/book/how-to-lie-with-statistics/chapter-1#good-sampling) · [Fisher’s tea experiment](https://www.sciencehistory.org/stories/magazine/ronald-fisher-a-bad-cup-of-tea-and-the-birth-of-modern-statistics/) · [Significance in business](https://mailchimp.com/resources/statistical-significance/) · [Naked Statistics](https://www.shortform.com/app/book/naked-statistics) · [Sample size and representation](https://www.shortform.com/app/book/naked-statistics/reducing-bias-in-data)

<!-- atlas:block-content b-0593b6b3db95ce08:end -->
<!-- atlas:block-content b-a792dd240eb44750:start -->
### What would have happened without the intervention?

<a id="v2-52-3"></a>

**Association and causal interpretation**

Ritchie warns that press releases and news often turn associations into statements that one factor “causes” another. Bevelin explains why that matters for action: a variable that accompanies success may not produce success when changed. Observational associations alone cannot identify causation. A causal interpretation needs a defensible design and assumptions about confounding, timing, selection and the intervention being considered. Randomized experiments are one strong approach, not the only possible route; an effect estimated in one setting also needs justification before transfer to another.

<a id="b-a792dd240eb44750"></a>
<a id="u-seeking-wisdom-bevelin-70986b433939"></a>
<a id="u-seeking-wisdom-bevelin-83dcde4d64fe"></a>
<a id="u-seeking-wisdom-bevelin-2a3a1112523c"></a>

<a id="v2-62-0"></a>

**Confusing correlation with causation leads to flawed decisions**

Bevelin’s sweater-shop example begins with low summer sales, extra advertising, and a sales rise from October through January. Crediting the rise entirely to advertising ignores a plausible common driver: colder weather increases demand for sweaters. Comparing other shops’ autumn sales can strengthen the seasonal explanation, but does not prove advertising had no effect; shops may differ in customers, prices or campaigns. The decision is how much advertising changes sales beyond the seasonal change, which calls for a comparable counterfactual.

<a id="b-c618bf4c74a9f3be"></a>
<a id="u-seeking-wisdom-bevelin-36dd3ef6f4c8"></a>
<a id="u-how-to-measure-anything-19f42bbe3de2"></a>

<a id="v2-63-0"></a>

**Generating Causal Evidence Through Trials and Experiments**

When existing causal evidence does not fit your situation, a small trial can generate useful firsthand observations. But running sweater advertisements one fall and none the next does not isolate advertising: weather, prices, competition, customer mix and economic conditions can all change between years. Treat that comparison as suggestive evidence. A randomized concurrent comparison, or a defensible quasi-experimental design with its assumptions examined, provides a stronger basis for estimating the campaign’s effect. The Heaths’ trial-run advice remains useful for testing feasibility and experience, even when a trial cannot establish a clean causal estimate or long-term outcome.

Despite Hubbard’s argument that experiments are easy to run and deliver real benefits, companies often resist them. One objection is that experiments feel unfair in the moment. When businesses experiment with pricing, employee benefits, or product features, they inevitably create situations where some customers or workers receive a better deal than others. Many executives find this morally uncomfortable and worry it may backfire, even if the long-term goal is to improve outcomes for everyone. The visible unfairness of the present often seems more important than the possibility of better results later. A second issue is that businesses generally crave certainty and clear direction. Hiring expensive consultants or running focus groups gives leaders definitive answers they can act on immediately; these approaches feel productive precisely because they produce confident recommendations. Experiments, by contrast, generate questions and require ongoing analysis. They slow decision-making and introduce uncertainty into an environment that rewards confident leadership. Experimentation forces leaders to admit they do not already know the best path forward, which feels uncomfortable and indecisive.

<a id="b-8102edf0b960ac3c"></a>
<a id="u-how-to-measure-anything-4716fa857b14"></a>
<a id="u-how-to-measure-anything-1312ce5e4a49"></a>
<a id="u-how-to-measure-anything-2e1e5488a9fa"></a>
<a id="u-how-to-measure-anything-d35c57cd7781"></a>

<a id="v2-65-0"></a>

**Controlled experiments: intervention, comparison, measurement**

Hubbard emphasizes three elements: deliberately change something, use a comparison condition, and measure the outcome. A business experiment need not be a costly laboratory project. Random assignment helps make groups comparable in expectation; it does not literally hold every characteristic constant or guarantee balance in a particular sample. Execution, attrition, spillovers and measurement still matter.

A restaurant can randomly offer complimentary appetizers to selected tables and track return visits. Return rate measures repeat behavior, a possible proxy for satisfaction rather than satisfaction itself; define the follow-up period and track both groups comparably. An email team can randomly send alternative subject lines while holding the message and timing comparable. These interventions answer a narrower question than whether the whole business strategy works everywhere.

<a id="b-a13155ee049439ce"></a>
<a id="u-science-fictions-9fc7b49db7e2"></a>

<a id="v2-64-0"></a>

**Shortform perspective: Practical Habits for Avoiding Mistaken Causal Inference**

A Shortform editorial sidebar adds practical habits for avoiding mistaken causation in everyday life. First, practice “reverse predictions”: notice an outcome, brainstorm possible causes, then research afterward to see whether your proposed correlations hold up or whether other explanations fit better. For example, if you see unusually heavy local traffic, list potential factors such as weather, events, or construction, then check traffic reports, weather updates, and local news to test your guesses. Second, use the “five whys” technique from root cause analysis: when you encounter a causal claim, ask “Why?” repeatedly to dig deeper into the reasoning behind it. For instance, if a news story claims eating chocolate causes happiness, keep asking how and why until you reach a more fundamental understanding of the relationship.

<a id="b-2e5563a4c6091ecf"></a>
<a id="u-how-to-measure-anything-33fe15f24841"></a>
<a id="u-seeking-wisdom-bevelin-2e0cb95dd82e"></a>

<a id="v2-71-0"></a>

**Shortform perspective: Decision Theory, Causal Evidence, and the Limits of Good Decisions**

Hubbard contends that the purpose of measurement is to inform decisions, and that while perfect information may never be achievable, reducing uncertainty through measurement should at least guide us toward better decisions. However, research into decision-making reveals a hard truth: a good decision does not necessarily result in a good outcome. Researchers also note that perfectly rational decisions are not always possible, due to our finite cognitive capacity, the complexity of the situations we face, and practical limits on our time. Many experts have converged on “decision theory” as a response to these limitations: we should ask ourselves what we value most, then make the decision that maximizes that outcome. The goal is to weigh options and choose the one that best aligns with your values.

Shortform turns to Emily Oster’s Cribsheet for guidance on causal evidence. Random assignment can help separate an intervention from pre-existing group differences. It is different from random recruitment, which concerns who enters the study. A paper’s statement that it used random assignment is a starting point for checking the design, not a reliability certificate: examine implementation, missing outcomes, comparability, measurement and uncertainty. Observational studies can sometimes support causal claims through justified identification strategies, although uncontrolled associations alone cannot do so. Parenting evidence also needs a relevant population and outcome before transfer to other decisions.

*Sources and further reading:* [Decisive](https://short.zaynjarvis.com/books/decisive/) · [Testing an option](https://short.zaynjarvis.com/books/decisive/#strategy-3-conduct-a-trial-run) · [Resistance to experiments](https://www.linkedin.com/pulse/why-businesses-dont-experiment-denis-vanuyshkin) · [Decision research](https://hbr.org/2006/01/a-brief-history-of-decision-making) · [Values in decisions](https://www.newyorker.com/magazine/2019/01/21/the-art-of-decision-making) · [Cribsheet](https://www.shortform.com/app/book/cribsheet/1-page-summary) · [Randomized-trial evidence](https://www.shortform.com/app/book/cribsheet#goal-1-gather-actionable-data)

<!-- atlas:block-content b-a792dd240eb44750:end -->
<!-- atlas:block-content b-31e7e08342edae82:start -->
### Simulate combinations without mistaking the model for the world

<a id="b-31e7e08342edae82"></a>
<a id="u-how-to-measure-anything-935eee089dd6"></a>
<a id="u-how-to-measure-anything-a624b863477f"></a>
<a id="u-how-to-measure-anything-2b7b08e901db"></a>
<a id="u-how-to-measure-anything-cf07e6a4ff01"></a>
<a id="u-how-to-measure-anything-17e16deb8220"></a>
<a id="u-how-to-measure-anything-1a29fb386cc4"></a>
<a id="b-63042b844522971b"></a>
<a id="u-how-to-measure-anything-a65416d6c0ce"></a>
<a id="u-how-to-measure-anything-582e989eb85b"></a>

<a id="v2-66-0"></a>

**Monte Carlo: make assumptions generate a distribution**

A Monte Carlo simulation repeatedly samples from specified input distributions and runs a model. It helps examine combinations too numerous to calculate manually, using anything from spreadsheet tools to specialized software. Its output describes possibilities under those distributions, dependencies and model equations; it is not a complete inventory of the future.

The summary illustrates the method with Sarah considering three coffee shops. Per store, construction is set at $180,000–$280,000, daily customers at 300–500, and monthly operating expenses at $12,000–$18,000. A midpoint plan uses $230,000, 400 customers and $15,000. The teaching example stipulates an 18% return, then 10,000 simulated scenarios with the same mean, a 25% chance of returns above 30%, and a 15% chance of a loss. These outputs cannot be derived from those ranges alone: revenue per customer, time horizon, return denominator, distributions and correlations are not fully specified.

Within the example, construction overruns emerge as an important exposure. Sarah negotiates fixed-price contracts, selects locations with more predictable demand and stages the rollout. A fixed-price contract transfers some cost risk to the contractor; it can leave scope changes, counterparty failure and other exposures. The useful result is identifying which assumptions drive the choice and which actions change them.

<a id="b-91728efaeca85eab"></a>
<a id="u-how-to-measure-anything-4138d4999658"></a>

<a id="v2-67-0"></a>

**Shortform perspective: Limitations of Monte Carlo Simulations for Measurement**

Shortform emphasizes that simulation inherits defects in its inputs and model: missing dependencies, biased data and implausible distributions can make precise-looking results misleading. Computational cost depends on the model and number of runs; many ordinary business simulations fit on a laptop, while demanding models need more resources. Interpreting the outputs is a separate skill. Check sensitivity, distinguish model uncertainty from simulated variability, and connect the results to an explicit action threshold.

<a id="b-4398a0625c719309"></a>
<a id="u-how-to-measure-anything-19b520c56901"></a>

<a id="v2-69-0"></a>

**Shortform perspective: Decomposition has limits, but there is no universal ten-day forecasting barrier**

Shortform contrasts Hubbard’s decomposition method with computational irreducibility: for some processes, no general shortcut predicts the state more cheaply than following the process’s evolution. Breaking a system into pieces therefore does not guarantee an accurate or inexpensive forecast.

Weather illustrates a related but distinct issue—sensitivity to initial conditions. Measurement uncertainty and model imperfections grow during a forecast, limiting reliable detailed predictions. “Beyond ten days, even unlimited computing could not predict the weather” is too strong: skill depends on the variable, spatial scale, initial information, model and forecast horizon. More computation can improve some predictions without removing all uncertainty. The practical lesson is to test whether decomposition preserves the interactions that matter, and to represent the uncertainty left over.

*Sources and further reading:* [Simulation limitations](https://projectmanagers.net/top-10-cons-disadvantages-of-monte-carlo-analysis/) · [Computational irreducibility](https://www.wolframscience.com/nks/p737--computational-irreducibility/) · [Weather-prediction limits](https://www.nesdis.noaa.gov/about/k-12-education/weather-forecasting/how-reliable-are-weather-forecasts)

<!-- atlas:block-content b-31e7e08342edae82:end -->
<a id="theme-decision-process-and-options"></a>

## A complete decision process

<!-- atlas:block-content b-f3672afde52f7449:start -->
### Two processes, kept distinct

Parrish asks: clarify the aim, generate alternatives, evaluate them. The Heaths’ WRAP targets four vulnerabilities across that work: narrow options, confirmation, short-term emotion and overconfidence about outcomes. They are related frameworks with different steps.

<a id="b-f3672afde52f7449"></a>
<a id="u-decisive-a3f352962b12"></a>
<a id="u-decisive-a4aa56431af3"></a>
<a id="u-decisive-160c005c450f"></a>
<a id="u-decisive-284179a025f0"></a>
<a id="u-decisive-057d7a9bbab8"></a>
<a id="u-clear-thinking-f3f5df748a5c"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-2357d034bf26"></a>
<a id="u-decisive-0384fc6c1abd"></a>
<a id="u-decisive-e4dea54cfc8e"></a>

<a id="v2-75-0"></a>

**Why structured decision-making matters**

Most people struggle to think clearly. Unconsciously, knee-jerk emotional reactions or the desire to feel “smart” push them toward poor choices that harm their careers, relationships, and overall life satisfaction. Shane Parrish argues in *Clear Thinking* that the world’s most successful people have found ways to overcome their innate irrational tendencies and consistently make decisions aligned with their goals; by cultivating certain mental habits and acting more systematically, anyone can engage in clearer, more purposeful thinking and maximize their potential for success.

Dan and Chip Heath, in *Decisive*, similarly frame flawed, biased thinking as the central barrier to good decisions. They observe that while research has extensively documented how flawed human decisions are, there is far less practical guidance on how to improve them. Drawing on psychology research and interviews, and coming from business backgrounds themselves, they present a general decision-making process usable for major choices across every area of life. They insist that mere awareness of our bad decision habits is not enough to improve our decisions; we need a process that actively resists each habit.

Our default decision-making process—identifying options, weighing them, and picking one—is not inherently flawed. According to the Heaths, the real problem is that we passively allow cognitive biases and irrational thinking to influence every step of that process. Shortform organizes these distorting influences into four “bad habits” that compromise the quality of our choices.

Gabriel Weinberg and Lauren McCann note in *Super Thinking: The Big Book of Mental Models* that decision-making is also inherently difficult because no one has access to perfectly accurate, comprehensive information; our information is often messy and flawed, so we cannot always predict how decisions will turn out. Because mere awareness of bad decision habits alone is not enough to improve decisions, structured processes and tools are needed to actively counter those habits and help select the best choice based on the complexity of the decision.

<a id="b-70c6808c732e8af4"></a>
<a id="u-clear-thinking-ce2edfa9a9e0"></a>
<a id="u-clear-thinking-5616c8e47c9a"></a>
<a id="u-clear-thinking-b68dfe7654eb"></a>
<a id="u-clear-thinking-4bef65010c67"></a>

<a id="v2-76-0"></a>

**Parrish’s three-step process for important choices**

A second way to improve rational thinking is to follow a decision-making system. Parrish argues that much of daily behavior runs on autopilot, driven by habits or emotional reactions. While this automatic decision-making is usually helpful and efficient, it also increases the chance of making mistakes. Therefore, when facing a decision with important consequences, he recommends deliberating through a step-by-step process to maximize the chance of reaching the best possible outcome. Three steps can be followed to procedurally make effective decisions in high-stakes situations: first, clarify what you want to accomplish; second, brainstorm potential solutions; third, assess which solution is best.

<a id="b-f9c6ddb7eff37328"></a>
<a id="u-clear-thinking-0ddabe7d5a13"></a>
<a id="u-clear-thinking-9e6c805c1074"></a>
<a id="u-clear-thinking-6e8de5888ea2"></a>
<a id="u-clear-thinking-f2bef2c2ab1b"></a>

<a id="v2-80-0"></a>

**Step 1: Clarify the problem and its root cause**

When making a decision, your first step should be clarifying the problem your decision is intended to solve, Parrish says. Reflect on what you are trying to accomplish and what is standing in your way.

Specifically, try to identify the problem’s underlying cause before taking action. Often, people waste time and effort pursuing solutions that feel productive but end up creating the conditions for the problem to return. For example, a company might repeatedly hire last-minute contractors to hit important deadlines at significant expense instead of realizing that training a new full-time team would make it consistently easier to meet deadlines.

Additionally, if you separate the act of clarifying the problem you are trying to solve from the act of solving it, you are less likely to jump to misguided solutions: think deeply about what you are really trying to accomplish, then step away for some time before considering courses of action.

<a id="b-984d1be057253f95"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-847c0f2514e8"></a>
<a id="u-clear-thinking-85cb20f200de"></a>
<a id="u-clear-thinking-e079f5d0c038"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-5536b483a2e7"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-a54cf011946c"></a>
<a id="u-decisive-c321efa83de5"></a>
<a id="u-decisive-6632a0122c8d"></a>
<a id="u-decisive-dae1636e20e9"></a>

<a id="v2-81-0"></a>

**Shortform perspective: Clarifying Core Goals and Root Problems Before Deciding**

Weinberg and McCann do not offer actionable advice for finding your “north star” to guide decisions. As a supplementary strategy, Simon Sinek’s *Start With Why* proposes stepping back to reflect on your motivations—for instance by asking, “What do I particularly enjoy about my job?” A bakery owner might realize through this exercise that baking serves the larger goal of bringing the community together in a warm environment.

To dig down to the root of a specific problem, the “Five Whys” exercise from Eric Ries’s *The Lean Startup* works well: repeatedly ask “Why did this problem occur?” to trace a chain of causes backward. If weekly team meetings run too long, you might find they lack an agenda, which in turn happens because the team has no process for building agendas; by the fifth “why,” you may uncover that the real issue is a company culture that discourages process improvements.

Stepping away after you have clarified the problem does more than prevent misguided solutions—it can also spark better ones. Barbara Oakley explains in *A Mind for Numbers* that when you stop actively focusing on a problem, you shift into diffuse mode: a relaxed, big-picture state in which your brain makes intuitive connections across ideas. When you return, you may notice solutions you would have missed otherwise.

<a id="b-ee550fa92bfc15ab"></a>
<a id="u-clear-thinking-872187204220"></a>
<a id="u-clear-thinking-781a8ca98f18"></a>
<a id="u-clear-thinking-aa1ead50ed99"></a>
<a id="u-decisive-b2eeff0a03e3"></a>
<a id="u-decisive-401ba4c955fe"></a>
<a id="u-decisive-cfae13584940"></a>
<a id="u-decisive-d456f95423dc"></a>

<a id="v2-82-parrish"></a>

**Parrish: generate at least three options**

Parrish echoes this emphasis on moving past your first couple of ideas. People frequently trap themselves in false dichotomies, assuming they must choose between Option A and Option B when the best solution is actually a hidden Option C. To avoid this, he recommends always considering at least three options for important decisions.

Parrish also advises trying to construct solutions that combine the benefits of multiple options, a strategy called integrative thinking. Challenging your initial rigid understanding of your options and uncovering their hidden nuances takes effort, he says, but it is almost always worthwhile. For example, a marketing professional choosing between a prestigious but expensive high-production campaign and an affordable but amateurish social-media campaign might realize they could produce one polished video to establish prestige, then rely on social media clips to keep the overall budget low.

<a id="b-d0c175442484910d"></a>
<a id="u-decisive-37940aea58b3"></a>
<a id="u-decisive-de4ea493d2a9"></a>
<a id="u-clear-thinking-15a9c859c99c"></a>
<a id="u-clear-thinking-6722c28544c5"></a>
<a id="u-clear-thinking-5998f4aa9b51"></a>
<a id="u-decisive-78658b19743f"></a>
<a id="u-decisive-65af85c7bf98"></a>
<a id="u-clear-thinking-a9d41fe8b3a9"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-e5fbc2ff36e7"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-93b9b81a4a62"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-ec09078578f5"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-534a13ff0e86"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-34e0d1878891"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-5dad3d4e310e"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-133539258381"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-3b1b82cdd57a"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-36b687a764d9"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-666ded804fce"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-fd7101641f32"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-044d64350b79"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-e4a6698152ad"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-e92144336728"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-affe31ea0972"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-e80c54e9887f"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-cd4d48e47f61"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-d7c932fc2d91"></a>
<a id="b-07c785a196b52ac4"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-aa3071755e23"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-c866ea17e9f1"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-c46d262de464"></a>

<a id="v2-91-parrish"></a>

**Parrish: criteria and information**

Parrish frames good evaluation as resting on two pillars: appropriate decision criteria and effective information sourcing. Decision criteria are the measures you use to judge whether a decision advances your long-term goals. He advises you to generate a list of criteria and rank each one by its impact on your goals; for example, if you are deciding what kind of business to start, you might prioritize skill development opportunities first, then profit potential, and finally networking opportunities. Clearly articulating your criteria makes it easier to accurately assess the quality of your options.

Information sourcing is the process of gathering enough context to understand which option best fulfills your criteria. Parrish argues that effective information sourcing means learning from people who have extensive experience making the kinds of decisions you want to make, and he recommends communicating with these experts directly whenever possible. Anyone who summarizes an expert’s advice filters it through their own assumptions about what matters, which may not align with yours. Returning to the business-startup example: as part of deciding what kind of business to launch, you might interview an entrepreneur who has started multiple companies about their experience.

<a id="b-abfe5fc8e11821ac"></a>
<a id="u-clear-thinking-24ead268aef8"></a>

<a id="v2-77-0"></a>

**Why a Step-by-Step Process Interrupts Autopilot**

Shortform uses Kahneman’s two-system model to explain why Parrish’s framework can help. System 1 handles most everyday tasks quickly and intuitively, relying on heuristics that are usually efficient but sometimes fail. The availability heuristic, for example, makes people estimate frequency by how easily examples come to mind, helping explain why they overestimate death in plane crashes while underestimating heart-disease risk. Deliberately following a framework for an important decision can prompt a shift toward slower System 2 reasoning instead of leaving the choice to these automatic shortcuts.

<a id="b-f27d4d1cbaf92490"></a>
<a id="u-clear-thinking-c7352549ad0d"></a>
<a id="u-clear-thinking-995301bf5f6e"></a>
<a id="u-clear-thinking-9507f855f730"></a>
<a id="u-clear-thinking-23f4ae1aed68"></a>
<a id="u-clear-thinking-796a1b1a87cc"></a>
<a id="u-clear-thinking-94785be7cb7f"></a>
<a id="u-clear-thinking-796e722d8d77"></a>
<a id="u-clear-thinking-d275078aa953"></a>

<a id="v2-78-0"></a>

**What makes a decision “good”: aligning with long-term life goals**

Parrish defines a “good decision” not by how clever it looks on paper, but by whether it helps you achieve your big-picture life goals; decisions that leave you miserable and unfulfilled cannot count as good, no matter how rational they seem. He warns that many people chase goals that feel rewarding in the moment but are ultimately empty—such as amassing more wealth and social status than the people around them—a never-ending race that leaves them unsatisfied.

To avoid this trap, he recommends clarifying what truly matters to you by vividly imagining how you will feel just before death, when you are very old. Picturing the accomplishments you are proud of, the decisions you regret, and the experiences that brought you joy helps you view your present circumstances from a long-term perspective and motivates you to build a satisfying life. The purpose of clear thinking, in this view, is to help you live a fulfilling life. The *Clear Thinking* guide outlines two main ways to strengthen clear thinking: building better mental habits and following a decision-making system. It also addresses how to compensate for the times when you inevitably fail to think clearly.

*Sources and further reading:* [Start With Why](https://www.shortform.com/app/book/start-with-why) · [Clarifying motives](https://www.shortform.com/app/book/start-with-why#why-should-affect-everything-you-do) · [The Lean Startup](https://short.zaynjarvis.com/books/the-lean-startup/) · [Diffuse thinking and incubation](https://www.shortform.com/app/book/a-mind-for-numbers/part-1#diffuse-mode) · [Biases and fallacies](https://short.zaynjarvis.com/books/thinking-fast-and-slow/#common-cognitive-biases-and-fallacies)

<!-- atlas:block-content b-f3672afde52f7449:end -->
<!-- atlas:block-content b-808a24d52ff8f615:start -->
### WRAP: widen the options

<a id="b-808a24d52ff8f615"></a>
<a id="u-seeking-wisdom-bevelin-767ae2fc2c28"></a>
<a id="u-decisive-d4addf09179c"></a>
<a id="u-decisive-063f29b391f3"></a>
<a id="b-1c00c53cfebff720"></a>
<a id="u-clear-thinking-3f4f4481ea99"></a>
<a id="u-decisive-79f1f309ad56"></a>

<a id="v2-79-0"></a>

**The WRAP Decision Framework and Its Intellectual Context**

The Heath brothers organize their decision-making process into a four-step mnemonic, “WRAP”: Widen Your Options, Reality-Test Your Assumptions, Attain Distance Before Deciding, and Prepare to be Wrong. Shortform rephrases each step to clarify its exact goal but preserves the original sequence because each step builds on the one before it.

Scholarly approaches to decision-making fall into several traditions. The Heath brothers belong to a cognitive-flaw camp that studies how mental biases shape choices; this line of work includes Daniel Kahneman (*Thinking, Fast and Slow*) on irrational decision-making and Malcolm Gladwell (*Blink*) on intuitive decision-making. A separate tradition focuses on cultural and societal influences: Sheena Iyengar (*The Art of Choosing*) explores how culture shapes the value we place on choice and the choices we make, while Barry Schwartz (*The Paradox of Choice*) studies how the abundance of choice in market democracies affects decision-making and well-being. *Decisive* contributes to the cognitive-flaw tradition by offering a structured process designed to make decisions more rational, and it spends little time on cultural or societal factors.

Shortform adds Gladwell’s *Talking to Strangers*: suspicion can help detect exploitation, while default trust makes cooperation possible. Gladwell presents an evolutionary argument for the value of trust. It is not a universal claim that trusting always beats caution; the practical tension is how to retain cooperation while responding to credible signs of danger.

<a id="v2-82-0"></a>

**W — Widen your options**

The Heaths’ W step widens the choice set beyond an either/or frame. Their practical target is a few strong, genuinely distinct options—often three or four—not a universal optimum or a guarantee of usable backups. Three methods follow: combine compatible benefits with both/and thinking, imagine the current options unavailable, and learn what experienced people have tried.

One strategy the Heaths recommend for expanding options is to imagine eliminating your current options, which can push you to discover alternatives you would otherwise overlook.

<a id="b-7f7b8f595ff052f6"></a>
<a id="u-decisive-faf65265859d"></a>
<a id="u-decisive-f1d63202898a"></a>
<a id="u-decisive-9fab02f2ab4a"></a>
<a id="u-seeking-wisdom-bevelin-d462121fb885"></a>
<a id="u-seeking-wisdom-bevelin-3ccba2f2126e"></a>
<a id="u-seeking-wisdom-bevelin-284c70c7f699"></a>
<a id="u-seeking-wisdom-bevelin-df3a6b9047fb"></a>
<a id="u-decisive-a4283585d969"></a>
<a id="u-decisive-39363de5ecaf"></a>
<a id="u-decisive-e416bf396ea7"></a>

<a id="v2-83-0"></a>

**Binary Thinking and the Evolutionary Roots of Narrow Choice**

A core flaw in intuitive decision-making is binary thinking: when faced with a choice, people tend to identify and consider only two options. In *Decisive*, the Heath brothers argue that this either/or framing prevents us from developing multiple, potentially better alternatives. For example, a retiree weighing whether to downsize to a smaller home near family or stay in their beloved current house may frame the choice as “stay or sell,” missing viable options such as downsizing while renting out the original home so they can spend time in both places.

Bevelin connects moral conformity with social cooperation in his evolutionary account. Hunting together, building shelters and tending wounds illustrate benefits of group membership; exclusion could remove those benefits. This offers a possible explanation for seeking acceptance and fearing rejection, not proof that people are genetically destined to rank approval above independent moral judgment. In a present decision, the relevant question is whether fear of social costs has excluded an option you otherwise consider right.

Bevelin argues that although our evolutionary past biases us toward social inclusion over independent moral judgment, we can counteract this tendency with rational thinking. He offers a strategy from Warren Buffett to motivate moral choices even when they are unpopular: when facing a moral decision, ask yourself whether you could live with a critical journalist writing a front-page story about it. For instance, consider whether you could bear having everyone you know read the headline: “Worker Fails to Call Out Colleague for Sexist Comment, Claiming She ‘Hoped Someone Else Would Do It.’”

<a id="b-ff6828a8930fb336"></a>
<a id="u-decisive-2fe419e158b4"></a>
<a id="u-decisive-cd003601fefa"></a>
<a id="u-decisive-737d88d9c042"></a>

<a id="v2-84-0"></a>

**Both/And Thinking: Combining Options to Escape Binary Frames**

Both/and thinking is a strategy for uncovering more options by moving beyond binary A-or-B choices. Instead of picking one side, you create an option that combines the best parts of both alternatives. One way to build such an option is to pair an optimistic choice with a cautious one; the result is a cautiously optimistic option that prepares you for a range of outcomes—both positive and negative—and acts as a middle ground between a high-risk path and one that is overly prudent. For example, consider a postal worker weighing whether to quit his job to become an artist. Within the binary of “quit or stay,” the optimistic option is to quit and try to earn a living from art, while the cautious option is to keep working at the post office. A cautiously optimistic alternative would be to reduce his hours at the post office while he builds his art career.

<a id="b-77e56b3de0ea027b"></a>
<a id="u-decisive-5134e28efc23"></a>
<a id="u-decisive-c6e826842e89"></a>

<a id="v2-85-0"></a>

**Imagining Eliminated Options to Spark Creative Alternatives**

Another strategy for generating more options is counterintuitive: imagine that one or more of your current options have been eliminated. The Heath brothers argue that when we are forced to rule out an option, the constraint drives us to creatively develop other alternatives. For illustration, consider someone who wants to sign up for a full marathon but is new to running and unsure whether they can be ready in time. After imagining that the full marathon option is off the table, they research other possibilities and add “run a half-marathon” to their list—an option that is both ambitious and realistic.

<a id="v2-83-1"></a>

**Learning From Experts to Expand Your Option Set**

A further way to develop additional options is to learn from experts. The Heaths define an “expert” as anyone who has more experience than you with the topic of your decision; experts may be able to share tried-and-tested options that you had not even thought of. For example, consider a teenager who asks his parents for a cell phone. His parents want to please him but are reluctant to increase his social media access. To produce an extra option beyond the binary of “phone or no phone,” they could seek advice from other parents who have bought their children cell phones about how to supervise and limit children’s social media use.

<a id="b-41ef303edc01c722"></a>
<a id="u-clear-thinking-3479967027c3"></a>
<a id="u-decisive-a158ad0dfaaf"></a>
<a id="u-decisive-dabc159b9251"></a>

<a id="v2-86-0"></a>

**Shortform perspective: How Many Options? Evidence and Limits**

Recent psychological research explains why we default to binary thinking: it functions as a cognitive shortcut that lets us efficiently process large volumes of information, since weighing just two options is faster than weighing many.

Many researchers and authors investigate how many options are optimal for high-stakes decisions, and there is no clear consensus, suggesting the ideal number depends on the nature of the decision. Like the Heath brothers, researchers have observed clear negative consequences of having too many options: for example, when people are offered too many healthcare plan options, they become stressed and make careless choices—or in some cases opt out of enrollment altogether. One study found that online daters who chose from among six people were more satisfied with their decisions than people who considered 24 options. Although the Heath brothers warn against binary thinking, some research illuminates the benefits of binary choices: for example, research on retirement planning found that people made wiser investment choices when presented with two options versus five options. Furthermore, the optimal number of choices may depend on how different those choices are from each other. Research supports the claim that it is important to consider distinct options: when doctors were presented with three patient-care options including two highly similar ones, they usually opted for the most dissimilar option to avoid having to choose between the two similar options.

Taken together, the evidence does not yield one universal number. Shortform therefore treats the Heath brothers’ three-or-four-options rule as a practical default: enough variety to escape a false binary and retain backups, without creating an unmanageable choice set.

<a id="b-e9c46908947a8278"></a>
<a id="u-decisive-50dbb03ecd03"></a>

<a id="v2-87-0"></a>

**Shortform perspective: Constraints That Boost Creative Option Generation**

Experimental research supports the Heath brothers’ claim that constraints bolster creativity: limitations during the creative process drive creative problem-solving. When people brainstorm without any limits, they tend to follow their intuition rather than innovate; by contrast, brainstorming within limits—such as time or budget constraints—tends to yield more creative solutions.

Three additional ways to apply creativity-inducing constraints during option brainstorming are:

- **Ditch the cellphone.** Keep your phone in another room while brainstorming, as research shows that having your phone within reach—whether or not you actually use it—limits your cognitive competence.
- **Reduce online distractions.** If you are researching options online, download a browser website-blocker so you do not waste time web-surfing.
- **Limit the time you spend developing options.** Cap your brainstorming time so you do not expend excessive mental energy and time generating options you will later reject.

<a id="b-94a714cb527ade20"></a>
<a id="u-clear-thinking-fe4afbb471c0"></a>

<a id="v2-88-0"></a>

**Shortform perspective: Integrative Thinking: Combining Conflicting Perspectives**

Parrish credits the concept of integrative thinking to business researcher Roger Martin, who explores it more fully in *The Opposable Mind*. For Martin, integrative thinking is harder than simply combining benefits across options—it means understanding the distinct, often conflicting perspectives behind each option and integrating insights from those perspectives to generate truly innovative solutions. As an example, when designing the iPhone, Steve Jobs integrated his team’s focus on feature-rich functionality with his own conviction that technology should be intuitive to use, producing a device that felt revolutionary.

*Sources and further reading:* [Talking to Strangers](https://www.shortform.com/app/book/talking-to-strangers/1-page-summary) · [Default trust](https://www.shortform.com/app/book/talking-to-strangers/part-2-2) · [The Art of Choosing](https://www.twelvebooks.com/titles/sheena-iyengar/the-art-of-choosing/9780446504119/) · [The Paradox of Choice](https://www.shortform.com/app/book/the-paradox-of-choice/1-page-summary) · [Blink](https://www.shortform.com/app/book/blink/1-page-summary) · [Binary choices](https://www.psychologicalscience.org/news/releases/binary-bias-distorts-how-we-integrate-information.html) · [Health-plan choice overload](https://advocacy.consumerreports.org/wp-content/uploads/2012/11/Too_Much_Choice_Nov_2012.pdf) · [Dating-option experiment](https://www.tandfonline.com/doi/abs/10.1080/15213269.2015.1121827?journalCode=hmep20) · [Retirement-option experiment](https://www.uxmatters.com/mt/archives/2011/02/the-nature-of-choice-sets-and-their-affect-on-decision-making.php) · [Similarity among options](https://pubmed.ncbi.nlm.nih.gov/7815657/) · [Constraints and creativity](https://hbr.org/2019/11/why-constraints-are-good-for-innovation) · [Phone proximity and cognition](https://www.journals.uchicago.edu/doi/full/10.1086/691462) · [Website blockers](https://medium.com/@kiwiandroiddev/the-case-for-constraint-why-fewer-options-mean-less-procrastination-and-more-creativity-d11dc0f4e741) · [Brainstorming time limits](https://www.bloomberg.com/news/articles/2006-02-12/creativity-loves-constraints#xj4y7vzkg?sref=mIZ1p2Gz) · [The Opposable Mind](https://www.shortform.com/app/book/the-opposable-mind) · [Conflicting perspectives](https://www.shortform.com/app/book/the-opposable-mind#useful-beliefs)

<!-- atlas:block-content b-808a24d52ff8f615:end -->
<!-- atlas:block-content b-aa5b59634d7d265e:start -->
### WRAP: reality-test the assumptions

<a id="v2-91-0"></a>

**R — Reality-test your assumptions**

The Heaths’ R step—Reality-Test Your Assumptions—addresses confirmation bias. Test the leading options by seeking a contrary case, examining objective information and comparable base rates, and running a small trial where feasible. Parrish’s evaluation criteria serve a related purpose but belong to his separate three-step framework.

The Heaths similarly argue that confirmation bias leads people to trust their own biased instincts over more accurate data about what others in a similar situation have experienced; actively seeking objective information—especially base rates and real outcomes from comparable cases—helps counter this tendency.

<a id="b-aa5b59634d7d265e"></a>
<a id="u-decisive-e6c0950d58ce"></a>
<a id="u-decisive-2547db1d6fac"></a>
<a id="u-decisive-c0b8030bace3"></a>

<a id="v2-92-0"></a>

**Stress-test an option by eliciting an opposing view**

The Heaths propose a distinct stress test: actively seek an opposing viewpoint. Research they cite shows that considering a contrary view forces attention onto high-quality information that would otherwise be overlooked. A practical technique is to find an expert on the topic of your decision and ask questions designed to elicit the contrary case. For example, imagine a high school graduate choosing among colleges, with one school at the top of their list because they believe it has a strong art program. In a meeting with an admissions counselor, they might ask only questions that support that belief. Instead, they can stress-test the option by asking, “What are some issues with the art program?”

<a id="b-46cdaa8abeb839eb"></a>
<a id="u-decisive-57b582352726"></a>
<a id="u-decisive-83b5f448729b"></a>
<a id="u-decisive-5b11e7183fae"></a>

<a id="v2-93-0"></a>

**Gathering Objective Information: Base Rates and Targeted Expert Questions**

To evaluate options more reliably, you can seek objective information in several ways. One concrete method is to pay attention to base rates—statistics that describe what share of a population a given statement applies to. For instance, a college applicant choosing between schools could compare the base rates of each institution’s art majors who find jobs in the arts after graduation.

The Heaths also recommend asking experts questions that specifically prompt them to share factual information, rather than opinions or predictions. They advise avoiding questions about the future, because such questions may fall prey to the expert’s own biases. Instead, you should ask questions about the past and the present, on which an expert can provide a more objective perspective. For example, a college applicant should not ask a counselor, “Do you think this program will prepare me for an art career?” Better options include a present-focused question such as “What are ways you provide current students with job experience?” and a past-oriented question such as “What are some arts-based careers that recent graduates have found?”

<a id="b-f08edeb5762ed64e"></a>
<a id="u-decisive-ec68f0a7f26f"></a>
<a id="u-decisive-84e5d387854f"></a>
<a id="u-decisive-5311286d94a2"></a>

<a id="v2-94-0"></a>

**Trial Runs: Generating Firsthand Evidence to Test Options**

A trial provides firsthand evidence under its actual conditions. The Heaths suggest, for example, that partners uncertain about living together could try two weeks before a longer commitment. The experience reveals routines and frictions that speculation may miss, but a short trial can omit rare conflicts, novelty effects and longer-term pressures. Define what it tests, discuss expectations, and use the evidence to update rather than claim an accurate long-term forecast is guaranteed.

<a id="b-dabe2bc603e1ef56"></a>
<a id="u-clear-thinking-976dbae62055"></a>
<a id="u-decisive-e131b11f76b5"></a>
<a id="u-decisive-afd4ba874837"></a>
<a id="u-decisive-b9216c6fe5c8"></a>

<a id="v2-95-0"></a>

**Shortform perspective: Gathering and Evaluating Information From Others**

Shane Parrish (author of *Clear Thinking*) acknowledges that people who pass along advice inevitably filter it through their biases. Roger Martin (*The Opposable Mind*) agrees with this basic point and goes further: even direct input from an expert can be skewed by their frame of reference—including their resources, personality, and other background factors—making it less applicable to your situation. He therefore recommends synthesizing information from multiple sources, especially for complex problems, so that individual biases counterbalance one another and you can reach a more objective understanding.

Words such as “better than,” “worse than” and “could mean” are prompts for questions, not a test separating subjective from objective statements. A measured numerical comparison can be objective; “could” can accurately express uncertainty. Ask what is being compared, by which criterion, with what evidence, and which part depends on a value preference. The goal is to inspect the claim, not classify it by a phrase.

If you are unfamiliar with statistics, you can avoid misinterpreting data by checking the sample size—the number of people a statistic represents. A claim that “100% of our 2020 art program graduates found jobs in the arts” means very different things if the graduating class had three students versus 300.

For career-related decisions, researchers advise seeking input from expert professionals, while for personal decisions you should turn to experienced friends or family. When reaching out to a career expert, informational interview techniques help: use an email template, research the person in advance, and keep the Q&A short.

<a id="b-8d34757c5fc9ba4f"></a>
<a id="u-decisive-bbd13f263085"></a>

<a id="v2-96-0"></a>

**Shortform perspective: Countering Confirmation Bias and Filter Bubbles**

The Internet is a primary information source for many people, but it amplifies confirmation bias through the “filter bubble effect”: our own confirmation bias limits what content we seek out in the first place, and algorithms reinforce that bias by showing us only ads and search results that match our existing preferences, surrounding us with information that continually strengthens our current beliefs. One entrepreneur recommends countering this effect by “Googling the opposite”—using search terms that lead to opposing viewpoints. A high school graduate considering a college, for example, could search the school’s name alongside words like “overrated” or “unhappy” to find perspectives on its downsides.

<a id="b-ae07d51697ade002"></a>
<a id="u-decisive-16e7c9bdc8a8"></a>

<a id="v2-97-0"></a>

**Shortform perspective: Equity Barriers to Trial Runs**

Shortform points out that trial runs have access costs. It cites a 2019 figure of 43% unpaid internships at US for-profit companies; without the original survey denominator, treat this as the summary’s reported figure rather than a verified national rate. Unpaid placements can exclude people unable to finance them and can reproduce wealth-related inequalities, including racial disparities. They are not literally accessible only to wealthy people, but their opportunity cost matters. The summary’s White House campaign example illustrates an institutional response: changing compensation can broaden who can afford to try a path.

*Sources and further reading:* [Combining information sources](https://www.shortform.com/app/book/the-opposable-mind#gather-important-information-from-all-sides) · [Assessing comparison claims](http://www.butte.edu/departments/cas/tipsheets/thinking/claims.html) · [Checking sample size](https://datajournalism.com/read/handbook/one/understanding-data/become-data-literate-in-3-simple-steps) · [Choosing advisers](https://www.cnbc.com/2019/07/09/stop-asking-can-i-pick-your-brain-harvard-says-this-is-how-successful-people-ask-for-advice.html) · [Informational-interview email](https://www.indeed.com/career-advice/interviewing/what-to-expect-in-informational-interview) · [Preparing for an interview](https://lifedesignlog.com/a-comprehensive-guide-to-virtual-informational-interviews/) · [Concise informational interviews](https://www.themuse.com/advice/5-tips-for-nonawkward-informational-interviews) · [Searching for opposing views](https://www.changemakeracademy.com/articles/google-the-opposite/) · [Unpaid-internship estimate](https://www.washingtonpost.com/lifestyle/magazine/his-quest-to-get-interns-paid-is-paying-off/2019/01/11/93df2b2a-ff2a-11e8-83c0-b06139e540e5_story.html) · [Internships and inequality](https://instituteforpr.org/unpaid-internships-and-inequality-in-public-relations/) · [White House internship-pay campaign](https://www.forbes.com/sites/alexandrasternlicht/2022/06/03/meet-the-28-year-old-behind-the-white-houses-decision-to-pay-its-interns/?sh=1db39e8260a6)

<!-- atlas:block-content b-aa5b59634d7d265e:end -->
<!-- atlas:block-content b-9f2843bcd67741f5:start -->
### Choose the tool that makes the trade-off visible

Start by listing consequences, then weight their importance. Add probabilities when outcomes are uncertain. A calculated expected monetary value remains a model-dependent input, not a universal command.

<a id="v2-91-1"></a>

**Decision tools matched to complexity**

Weinberg and McCann present a tiered set of decision-making tools suited to different levels of complexity.

For the least complex decisions, the pro-con list is the best-known tool: it sorts the pros and cons of a decision into parallel columns to make positive and negative consequences easier to see. For example, a lawyer deciding whether to accept a higher-paying position that requires moving away from family could draft a pro-con list of accepting the offer.

![superthinking-pro-con-list.png](https://short.zaynjarvis.com/media/source-notes/03eed585d3-superthinking-pro-con-list.png)

<a id="b-9f2843bcd67741f5"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-ad2332faf570"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-71afec896deb"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-588fb068456b"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-17c0b0e3b829"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-e9c5a57666b6"></a>

<a id="v2-90-0"></a>

**Shortform perspective: Flaws of the Pro-Con List and When It Works**

Despite its popularity, the pro-con list has several flaws, according to Gabriel Weinberg and Lauren McCann in *Super Thinking: The Big Book of Mental Models*. First, it creates a false dichotomy, since many consequences do not fit neatly as pros or cons. Second, it fails to weight different pros and cons, so you focus solely on the quantity of pros and cons. Third, the grass-is-greener mentality leads us to naturally overemphasize pros. For these reasons, a pro-con list may work well enough for simpler decisions but is not recommended for more complex decisions.

<a id="v2-81-1"></a>

Shortform adds a countervailing bias: the grass-is-greener mentality favors change, while status quo bias can exaggerate its costs. To improve the list, distinguish major from minor consequences before weighing them—for example, greater upward mobility may matter more than an extra week of paid leave. This addresses the weighting problem; it does not make every consequence fit neatly into one column.







<a id="b-e6f8542c0075de00"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-65a4286b9889"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-b9217c7fc7ca"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-4358be1ceac9"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-084c1ccf9693"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-94eeb9b020ff"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-713d48d948c6"></a>

<a id="v2-192-0"></a>

**Cost-benefit analysis as a structured decision tool**

For a more complex choice, Weinberg and McCann recommend comparing cumulative costs and benefits. Their simple scoring version assigns each consequence a value from −10 to 10 and adds the scores. A positive total favors the action only under the chosen scale, weights, baseline and constraints; it is not a universal test of whether an action is worthwhile. In the job-offer illustration, the negative score means the listed disadvantages outweigh the listed advantages on that scale. Monetary estimates can make some trade-offs more concrete: moving might cost $10,000, while an extra week of leave might be valued at $3,000. Check that the values are comparable, avoid double counting, and keep essential constraints visible.

![superthinking-cost-benefit-analysis.png](https://short.zaynjarvis.com/media/source-notes/144df93071-superthinking-cost-benefit-analysis.png)

<a id="b-b800f43e20dc4432"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-3f6ec5a54a9a"></a>

<a id="v2-193-0"></a>

**Shortform perspective: Cost-Benefit Analysis and Diminishing Marginal Utility**

While assigning monetary values to costs and benefits can be useful, it is important to remember that monetary value diminishes beyond a certain point, due to the diminishing marginal utility of income and wealth. For example, a $15,000 raise could tangibly improve quality of life for someone making $40,000 a year, but would likely have far less impact for someone already making $250,000. These diminishing returns should be considered alongside any cost-benefit analysis.

<a id="v2-91-tree"></a>

**A tree makes uncertainty explicit**

For the most complex decisions, a tree can make alternative outcomes and their probabilities explicit. Simple lists and weighted scores may leave uncertainty implicit; cost-benefit analysis can also incorporate probabilistic estimates. A tree is useful when you can define branches and conditional consequences meaningfully, as in a policy with several uncertain responses.

Crucially, decision trees allow us to calculate each decision’s expected value—the average monetary value resulting from a decision. To do so, multiply the likelihood of each outcome by its respective monetary value, then sum them up for each decision. For a decision with two possible outcomes, the formula is:

(Outcome 1 likelihood) × (Outcome 1 monetary value) + (Outcome 2 likelihood) × (Outcome 2 monetary value) = expected value

The lawyer illustration stipulates an 80% chance of becoming a partner after accepting the new position and a 20% chance of being fired under its pressure; staying is assigned a 100% chance of remaining an associate for the period considered. The outcomes are valued at $100,000, −$50,000 and $75,000. In the actual diagram, the decisions are on the left, outcomes in the middle and monetary values on the right. All three columns use rectangles, and the branch probabilities appear on the connecting arrows.

![superthinking-decision-tree.png](https://short.zaynjarvis.com/media/source-notes/2c4860b28c-superthinking-decision-tree.png)

Under these stipulated monetary values and probabilities, accepting has expected value 0.8×$100,000 + 0.2×(−$50,000) = $70,000, while staying gives $75,000. Maximizing expected money favors staying. Those are teaching assumptions, not measured career odds; preferences, nonfinancial consequences and utility can change the choice.

*Sources and further reading:* [Status quo bias](https://www.psychologytoday.com/us/blog/stretching-theory/202206/is-fear-change-holding-you-back#:~:text=Human%20decision%2Dmakers%20have%20an,making%20and%20stifle%20people's%20potential.) · [Limits of pro/con lists](https://hbr.org/2017/01/the-pros-and-cons-of-pros-and-cons-lists) · [Marginal utility of wealth](https://www.economicshelp.org/blog/12309/concepts/diminishing-marginal-utility-of-income-and-wealth/#:~:text=Diminishing%20marginal%20utility%20of%20income%20and%20wealth%20suggests%20that%20as,increase%20in%20satisfaction%20and%20happiness.&text=Utility%20means%20satisfaction%2C%20usefulness%2C%20happiness,to%20spend%20on%20a%20good.)

<!-- atlas:block-content b-9f2843bcd67741f5:end -->
<!-- atlas:block-content b-aab1e3ccabad0e67:start -->
### WRAP: gain distance and decide what matters

<a id="b-aab1e3ccabad0e67"></a>
<a id="u-decisive-caa90eedcd2e"></a>
<a id="u-decisive-b1b059c0db89"></a>
<a id="u-decisive-3761d541b4c1"></a>
<a id="u-decisive-d1beb0a3f426"></a>
<a id="u-decisive-e8ffa21e63ba"></a>
<a id="u-decisive-639f3cf4294c"></a>
<a id="u-decisive-99ac0bf23083"></a>
<a id="u-decisive-f1c4b3c34858"></a>
<a id="u-decisive-94a4950947d1"></a>

<a id="v2-100-0"></a>

**A — Attain distance before deciding**

A key challenge in decision-making is that strong present emotions can distort our judgment. The Heaths dedicate a step of their process to putting your current emotions into perspective, and offer several strategies for doing so.

One strategy is to imagine the point of view of someone else who is not experiencing the same strong emotions. This allows you to more rationally and objectively see the decision’s important facets, and you might then recognize the value of disrupting the status quo and seeking beneficial change. For any kind of decision, ask yourself: “If my friend or colleague was in this situation, what decision would they make, and why?”

A second strategy is to look far into the future. The authors state that we make better decisions when we give additional weight to our future feelings; considering how we will feel later makes us more open to changing the status quo. One technique they recommend is the “10/10/10 test,” created by business writer Suzy Welch. First, select one of your options. Then ask yourself: “If I chose this option, how would I feel about my decision 10 minutes, 10 months, and 10 years after I made it?” If you determine that you would feel positively about that option at all three points in the future, it is a strong option.

When a decision remains emotionally fraught even after trying perspective-taking or future-oriented reflection, it often signals a conflict between two or more of your core values. The best way to work through such a decision is to clarify which values matter most to you. One value-clarifying technique the authors recommend is the “stop doing list,” created by business consultant Jim Collins. The exercise has three steps: First, imagine receiving two phone calls—one telling you that you have inherited 20 million dollars, and another saying you will die in 10 years. Second, ask yourself what you want to spend more time doing to make those last 10 years meaningful, and what you need to stop doing to make that possible. Third, the Heath brothers add a final step: look at your schedule and remove everything you decided you want to stop doing. What you keep reflects what you value most, and what you cut reflects what you value less.

<a id="b-437cdfa37e593238"></a>
<a id="u-decisive-37fcc52b29d4"></a>

<a id="v2-102-0"></a>

**Shortform perspective: Predicting Future Feelings and Mental Time Travel**

Shortform places Gilbert’s account of affective-forecasting errors alongside the 10/10/10 exercise: present feelings and imagination can distort predictions of future happiness. Its suggestion that Gilbert “would likely” criticize the exercise is the editor’s extrapolation, not a documented verdict by Gilbert. The exercise can broaden perspective without supplying measured future feelings.

However, regardless of whether accurate prediction is possible, Annie Duke, author of *Thinking in Bets*, argues that imagining the future still positively impacts decision-making. She claims that by remembering past decisions and imagining future outcomes, we better manage our emotions and make more rational decisions—a strategy she calls “mental time travel.”

<a id="b-e182101b08ed042f"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-380913198a96"></a>
<a id="u-decisive-3eb8afa1766b"></a>

<a id="v2-103-0"></a>

**Shortform perspective: Deliberation, Satisfaction, and Perspective-Taking in Decisions**

The decision-making tools discussed by Weinberg and McCann rest on the assumption that conscious deliberation leads to better decisions. However, a set of recent studies found that even for complex decisions, people were generally more satisfied with choices made without deliberation. IKEA customers buying complex products, for example, were more satisfied when they purchased without deliberating than when they deliberated first. One possible explanation is that deliberation raises expectations, making people harder to satisfy. This does not necessarily mean deliberation produces worse *decisions*—only less satisfying ones—so the finding can still be reconciled with the idea that deliberation improves decision quality.

Similarly, imagining what someone else would do in your situation is a common decision-making exercise, but research suggests we are often wrong when we simply guess another person’s thoughts: we predict their thoughts more accurately when we actually talk to them. Even so, imagining another person’s perspective still has value as a mental exercise: recent research indicates it boosts our ability to generate creative and novel ideas.

<a id="b-a8e2bd5a98c96629"></a>
<a id="u-decisive-81f2831ccf4f"></a>

<a id="v2-101-0"></a>

**Shortform perspective: Cultural Values and Decision-Making Style**

When clarifying your values for an emotionally fraught decision, it is worth considering how cultural background shapes your approach. Research shows that culture influences whether you look inward or outward when making such decisions: in cultures that value individuality, people tend to look inward and consult their own wants and needs, whereas in cultures that value collectivity, people tend to look outward and decide based on others’ expectations. Both approaches have merit. Notice whether you default to looking inward or outward, and if that default is not helping you manage your emotions or reach a decision, try the opposite approach.

<a id="b-3016b9017658dd33"></a>
<a id="u-decisive-36c90ac694cf"></a>
<a id="u-clear-thinking-82ff25402fdb"></a>
<a id="u-decisive-50e58f05a921"></a>

<a id="v2-89-0"></a>

**Shortform perspective: Prioritizing Decisions and Reducing Time on Low-Stakes Choices**

Uncovering more options offers many benefits but may be time-intensive. As Oliver Burkeman argues in *Four Thousand Weeks*, life is finite, so we must thoughtfully manage our time to prioritize what matters most to us. Similarly, Barry Schwartz argues in *The Paradox of Choice* that having too many choices in all areas of our lives causes us to devote too much time to unimportant decisions.

To decide which decisions are worth spending time on, psychologist Dan Ariely, author of *Predictably Irrational*, recommends considering whether the decision you’re making has an expiration date and prioritizing decisions that do. For example, a person with a sick grandparent may prioritize the decision of how to spend time with them over a non-expiring decision such as whether to buy a grill.

Once we’ve determined which decisions are worth our time, we can reduce our decision-making time for less-important ones. A psychologist who warns that decision overload leads to procrastination offers several techniques:

- **Have someone else make the decision.** Find someone else who might be better at making that decision than you (and possibly even enjoy making it). For instance, if you’re unsure what to wear to an event, consider asking a fashionable friend to choose your outfit.
- **Choose a “go-to” option.** For recurring decisions, automatically go with the same, easily doable option—even if it’s not perfect. For instance, if you want to exercise more, choose a “go-to” exercise such as a 30-minute run, and only deviate from that plan if you have the time and energy to brainstorm something that requires more planning, such as weightlifting.

Schwartz’s satisficing approach sets a satisfactory threshold instead of searching indefinitely for the best imaginable choice. Prioritize criteria, but meet essential constraints as well: an option that serves one or two leading preferences may still be unacceptable if it fails a health, affordability or other binding requirement. The trade-off is a feasible good-enough choice versus the cost and regret of endless maximization.

<a id="v2-89-1"></a>

**Shortform perspective: Satisficing for Values-Driven Decisions**

Jim Collins’s stop-doing list and the Heath brothers’ related addition prompt us to ponder what we value most and identify ways that we’re wasting our time. But even after using a stop-doing list to clarify values, you may still be left with multiple competing priorities. For decisions involving several important values, there may be no single perfect choice. Psychologist Barry Schwartz, author of *The Paradox of Choice*, recommends satisficing—choosing an option that is satisfactory rather than perfect. His research shows that maximizers, who exhaustively evaluate every possibility in search of the best option, are often more unhappy with their decisions.

To illustrate, imagine someone deciding whether to stay in the city or move to the country. She feels drawn to rural life but also values being part of a large LGBTQA+ community, which the city offers. After working through a stop-doing list, she finds she values both time in the country and time with queer community equally. As a maximizer, she might spend years searching for a perfect small town with a vibrant queer scene, with no guarantee of success. As a satisficer, she might miss out on that ideal town, but by making a choice she can end the stress of deliberation and start figuring out how to make her chosen path work well.

<a id="b-f48e7569b72d6a3a"></a>
<a id="u-thinking-in-bets-87e17bbfd8e5"></a>
<a id="u-thinking-in-bets-1a20fa996157"></a>
<a id="u-thinking-in-bets-180468381e88"></a>
<a id="u-thinking-in-bets-c238fb419c87"></a>
<a id="u-thinking-in-bets-8be6dee0323a"></a>
<a id="u-same-as-ever-7fbd52f1bc8c"></a>
<a id="u-same-as-ever-ca1c2cae342a"></a>
<a id="u-same-as-ever-b72527962cd2"></a>
<a id="u-clear-thinking-e677dc21308e"></a>
<a id="u-clear-thinking-6b2d19175de6"></a>
<a id="u-same-as-ever-2d05ae16f0a9"></a>
<a id="u-same-as-ever-e40c8048a3fd"></a>

<a id="v2-331-0"></a>

**Mental time travel: learning from the past and imagining the future**

Duke’s “mental time travel” brings past experience and future consequences into a present choice. Looking forward, imagine concrete effects on your future self—for example, what regular exercise or neglected health might change. Looking backward, recall regret from a comparable decision so it can inform action now rather than remain only an unpleasant memory. The two directions can be combined, including for choices such as relocation or a career change whose consequences take time to unfold.

Scenario planning gives the forward exercise structure: develop relevant, meaningfully different scenarios and assign tentative probabilities where useful. This can reveal contingencies, suggest barriers against impulsive choices and soften emotional reactions to anticipated successes as well as setbacks. It cannot enumerate every future or prevent every surprise. Leave room for unmodeled events, update scenarios as evidence changes, and keep the original reasoning so a later result does not rewrite what seemed reasonable beforehand.

*Sources and further reading:* [Collins’s stop-doing list](https://www.shortform.com/app/book/good-to-great/chapter-6#hedgehog-thinking-always) · [Stumbling on Happiness](https://www.shortform.com/app/book/stumbling-on-happiness/1-page-summary) · [Affective forecasting](https://www.shortform.com/app/book/stumbling-on-happiness#poor-choice-1-you-think-youll-feel-in-the-future-as-you-do-now) · [Deliberation and satisfaction](https://pubmed.ncbi.nlm.nih.gov/16484496/) · [Talking versus guessing perspectives](https://pubmed.ncbi.nlm.nih.gov/29620401/) · [Perspective taking and creativity](https://knowledge.wharton.upenn.edu/article/perspective-taking-brain-hack-can-help-make-better-decisions) · [Finite time and priorities](https://www.shortform.com/app/book/four-thousand-weeks#fact-2-your-life-is-finite) · [Choice demands](https://www.shortform.com/app/book/the-paradox-of-choice#the-consequences-of-expanded-choice) · [Predictably Irrational](https://www.shortform.com/app/book/predictably-irrational/1-page-summary) · [Decision deadlines](https://www.shortform.com/app/book/predictably-irrational/1-page-summary#irrationality-trigger-7-options) · [Decision overload and delay](https://www.psychologytoday.com/us/blog/in-practice/201910/6-solutions-procrastination-caused-decision-fatigue) · [Maximizing and satisfaction](https://www.shortform.com/app/book/the-paradox-of-choice/chapter-4#how-maximizing-harms-decision-making) · [Satisficing](https://www.shortform.com/app/book/the-paradox-of-choice/chapter-4#maximizing-vs-satisficing) · [The Paradox of Choice: book](https://www.shortform.com/app/book/the-paradox-of-choice)

<!-- atlas:block-content b-aab1e3ccabad0e67:end -->
<!-- atlas:block-content b-33757fa1ea858768:start -->
### WRAP: prepare for error, then return to the decision

<a id="b-33757fa1ea858768"></a>
<a id="u-decisive-c61419ad1fdc"></a>
<a id="u-decisive-7c0f2f9cb2a4"></a>
<a id="u-decisive-ac12e2244dae"></a>
<a id="u-decisive-e04d0613aacd"></a>
<a id="u-decisive-b24f4125a511"></a>
<a id="u-decisive-b2c5e7a5079d"></a>
<a id="u-decisive-aa64909f2abb"></a>
<a id="u-decisive-0e705b7cd1a6"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-b1eaffe45abe"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-d14733c868ee"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-8821df73a516"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-6b3ea2b84917"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-f5a9d479a1f9"></a>

<a id="v2-105-0"></a>

**Hubris and the Need for Humble Post-Decision Preparation**

In *Decisive*, the Heath brothers identify hubris as the fourth bad decision-making habit, which takes effect after a decision has already been made. Our hubris—or extreme self-confidence—makes us feel certain about our predictions of a decision’s outcomes; when the actual outcome differs from what we predicted, we are caught by surprise and unprepared to deal with it. They illustrate this with the example of a woman who decides to sell her sailboat before leaving town. Confident that her luxury boat will sell quickly, she makes no plan for the possibility that it won’t. When the market turns out softer than she predicted, she is left scrambling.

The authors emphasize that making a choice does not end the decision-making process; after you reach a decision, you can take action to increase the odds that it will succeed. Hubris makes people overconfident in their predictions about the future, which in turn leads to poor preparation for a decision’s actual outcomes. You can overcome hubris with humility and raise the likelihood of success by accepting that you cannot predict the future, and instead preparing for a range of possible outcomes.

<a id="b-9c96836d8cdfad35"></a>
<a id="u-decisive-4a075cb6d7ef"></a>
<a id="u-decisive-595a1a19b08f"></a>
<a id="u-decisive-5c551d81d25a"></a>
<a id="u-decisive-6df41c05a88c"></a>
<a id="u-decisive-d07d950cdb7f"></a>
<a id="u-decisive-8347e43a7f51"></a>

<a id="v2-106-0"></a>

**Contingency Planning for Best-Case and Worst-Case Scenarios**

The Heaths recommend contingencies for both favorable and adverse outcomes. A new food-truck owner can insure against specified accident losses and arrange possible help from friends in food service if a critic’s favorable review doubles demand. Insurance depends on coverage, exclusions and payment limits; available staffing also needs confirmation. The mechanism is preparing resources for different outcomes, not ensuring that no worst case can be devastating.

<a id="b-147a57d7ff520c9c"></a>
<a id="u-decisive-ab9a5668b1cf"></a>
<a id="u-decisive-cd0a2a8012d4"></a>
<a id="u-decisive-2ed657320a79"></a>

<a id="v2-107-0"></a>

**Safety Factors: Buffering Against Underestimated Worst Cases**

A safety factor adds a cushion to an estimate. The food-truck example multiplies projected expenses by 1.5, creating a 50% buffer. Parrish elsewhere illustrates preparing for twice an imagined bad outcome. Neither multiplier guarantees solvency or covers unknown extreme losses. Choose margins against plausible stress scenarios, available resources and commitments, and revise them as information changes.

<a id="b-9581411eeadd167b"></a>
<a id="u-decisive-b6ebf64c0b83"></a>

<a id="v2-108-0"></a>

**Shortform perspective: Safety Margins and Reframing Risk Preparation**

The Heath brothers frame safety factor strategies—building a buffer against worst-case scenarios—as “rainy day funds.” For people who are more motivated by positive incentives, the same strategy can be reframed as a “sunny day fund.” A food truck owner who calculates a 1.5x budget cushion, for example, can motivate herself to save by telling herself that if the money is not needed for setbacks, she can spend it on something enjoyable like a vacation.

<a id="b-6c3a5b467e514864"></a>
<a id="u-decisive-ba8c857484f6"></a>
<a id="u-decisive-8dad5a0c8ef1"></a>
<a id="u-decisive-0bbb84aec19e"></a>
<a id="u-decisive-e322a4a28d51"></a>
<a id="u-decisive-0d4e3e6cbcf1"></a>

<a id="v2-104-0"></a>

**P — Prepare to be wrong**

The Heaths’ P step—Prepare to Be Wrong—turns a choice into a revisable plan. Prepare for favorable and adverse scenarios, leave a safety margin, and set tripwires for review. These arrangements help you notice when another decision is needed; they neither guarantee timely detection nor require assuming that every choice will fail.

The authors offer two types of alerts. The first is a deadline: after making a decision, choose a near-future date, and when that date arrives, evaluate the decision’s success so far and decide whether you need to change course.

The second is a pattern: after making a decision, list several possible warning signs that the decision is going badly. If you notice a pattern of more than one warning sign, it is a signal to re-evaluate the decision or follow it up with another one. For example, a teacher with several new students wants to make sure they are integrating well with their peers, and identifies sitting alone at lunch as a sign of poor integration. He observes the students each day at lunch; if he notices a student sitting alone more than two times, the pattern alerts him to decide how to better support that student.

<a id="b-ba92f273e55958d9"></a>
<a id="u-decisive-a241bc10b103"></a>
<a id="u-decisive-82ce9d904ec6"></a>
<a id="u-decisive-535e095946b5"></a>

<a id="v2-109-0"></a>

**Alerts: Countering Passive Living and Decision Drift**

Alerts help counter passive living, a pattern the Heath brothers identify in which people move through life without actively reflecting on their daily behavior. When we live passively, we often fail to recognize opportunities to follow up on our decisions; an alert prompts us to notice those moments. For example, someone who has decided to buy a car may later slip into passive living and fail to notice the car’s gradual deterioration, never stopping to decide whether to get a tune-up or replace it with a better vehicle. Alerts also serve as a counterweight to hubris, another issue the Heaths highlight.

<a id="v2-81-2"></a>

**Shortform perspective: Alerts, Reflection, and the Cyclical Nature of Decision-Making**

Alerts—predetermined triggers that prompt a new decision—reveal that decision-making is cyclical: the outcomes of one decision create the conditions for the next. A deadline alert, for instance, gives you a chance to reflect on your decision at a set time. Annie Duke recommends that when reflecting on outcomes, you distinguish between results caused by luck and those caused by skill. This lets you reinforce behaviors that contributed to success, stop behaviors that led to failure, and acknowledge good or bad fortune that was outside your control.

Because confirmation bias makes us see what we want to see, we may miss warning signs that a decision is heading toward a negative outcome. When setting pattern-style alerts—those that watch for a recurring sign of trouble—it can help to enlist a friend, colleague, or family member who can spot warning signs and pull you out of your bias.

<a id="b-599d872380165c85"></a>
<a id="u-decisive-dd758b00c019"></a>

<a id="v2-110-0"></a>

**Shortform perspective: Scenario Planning and Contingency Planning**

Duke’s scenario planning imagines a useful range of possible outcomes and assigns provisional probabilities before choosing. The Heaths’ contingency planning prepares actions for outcomes after a choice. The two can form a cycle: explore scenarios, choose, prepare responses, then revise with new evidence. Neither requires—or can promise—an exhaustive list of every possible future.

<a id="b-d17e5259ccb6f5d1"></a>
<a id="u-decisive-3556f1a15ac1"></a>

<a id="v2-111-0"></a>

**Heuristics, Passive Living, and Poor Decisions**

Why do we tend to live passively? Daniel Kahneman’s research in *Thinking, Fast and Slow* shows that the brain sometimes uses heuristics—automatic shortcuts for processing information. This intuitive, efficient mode of judgment is useful in many situations, such as reading bathroom door signs to know which one to enter. Kahneman argues, however, that we make poor decisions when we act on these instincts without questioning them—in other words, when we passively let them rule our lives.

<a id="b-b6c276b140495e99"></a>
<a id="u-decisive-c75ba3c1fb64"></a>

<a id="v2-112-0"></a>

**Shortform perspective: Hubris: Universal Mindset or Power-Linked Condition?**

Shortform contrasts the Heaths’ everyday use of hubris with research discussing an acquired, power-linked hubris syndrome. These use different scopes: overconfidence can affect ordinary choices without implying a clinical or enduring syndrome. The comparison adds a possible institutional risk of power, not a direct refutation of the Heaths’ general advice to prepare for error.

<a id="b-57d70025a37001db"></a>
<a id="u-decisive-437ca43a65f1"></a>

<a id="v2-113-0"></a>

**Shortform perspective: Optimism, Caution, and Mindset in Decision-Making**

A cautiously optimistic mindset offers benefits that extend beyond decision-making into broader life satisfaction. Psychological research suggests that people who balance optimism with caution experience higher long-term contentment. The optimal balance, however, shifts with the task: some psychologists recommend pairing caution and optimism when setting goals, then adopting a more optimistic mindset while carrying out those plans.

*Sources and further reading:* [Deep Work](https://short.zaynjarvis.com/books/deep-work/) · [Positive framing of reserves](https://www.verywellmind.com/the-incentive-theory-of-motivation-2795382) · [Hubris and power](https://academic.oup.com/brain/article/132/5/1396/354862) · [Optimism and caution](https://www.psychnewsdaily.com/realists-are-happier-than-unrealistic-optimists) · [Task-dependent optimism](https://www.psychologytoday.com/us/blog/dont-forget-the-basil/201809/beyond-optimism-versus-pessimism) · [Mindsets in planning and action](https://psycnet.apa.org/record/1995-43739-001)

<!-- atlas:block-content b-33757fa1ea858768:end -->
<!-- atlas:block-content b-b04fb47af3b5be83:start -->
### Spend effort in proportion to the decision

<a id="v2-105-1"></a>

**Mental Models for Efficient Decision-Making**

Weinberg and McCann outline three mental models for making decisions efficiently. First, before approaching specific decisions, identify your north star—your overarching goal or vision. When you orient decisions toward your north star, your abilities compound and you become more efficient at pursuing your goals. Second, after identifying your north star, perform deep work, a concept from Cal Newport’s book *Deep Work*: this involves spending extended periods of time dedicated to one important task, free of distractions and interruptions. In the long run, deep work yields greater returns than juggling various tasks at once. Third, perform high-leverage activities to get more out of your decisions: just as a lever in physics allows you to lift heavier objects than you could on your own, high-leverage decisions create oversized results from less effort.

<a id="b-b04fb47af3b5be83"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-37cfc3e7444e"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-13e4e478cccf"></a>

<a id="v2-115-0"></a>

**Efficiency, Focus, and Leverage in Decision Execution**

Weinberg and McCann discuss deep work as an alternative to multitasking, which diffuses concentration across various activities. Cal Newport adds a further framing: deep work is also an antidote to task-switching, the frequent shifting of attention from one task to another. Because mental residue from a previous task takes several minutes to clear, task-switching likewise erodes efficiency.

The concept of leverage also shapes how effectively you act on decisions. In some contexts “leverage” carries a negative connotation of exploiting an advantage, as when a corrupt politician uses influence to pass unpopular bills. However, Shane Parrish and Rhiannon Beaubien argue in *The Great Mental Models Volume 2* that leverage can be an extension of reciprocity: it lets you offer others the best possible exchange for your services, benefiting everyone involved.

<a id="b-b2ab4130f3a013b8"></a>
<a id="u-noise-b1eb26ea6917"></a>

<a id="v2-152-0"></a>

**Shortform perspective: Second-Order Decisions as Cognitive Burden Reducers**

Rules and standards are examples of what Cass Sunstein and Edna Ullmann-Margalit call second-order decisions: strategies for reducing cognitive burden when decisions are too numerous, too repetitive, too difficult, or too ambiguous to make one by one. Other second-order decisions include:
- **Presumptions**: rule-like guidelines that allow for exceptions in some cases.
- **Routines**: recurring habits such as always brushing your teeth right before bed.
- **Small, reversible steps**: low-stakes trials such as pet-sitting for a neighbor’s dog before committing to adopting one of your own.
- **Random picking**: choosing without deliberation, such as throwing a dart at a map to pick a vacation destination.
- **Delegation**: letting someone else decide, such as allowing your partner to choose dinner.
- **Heuristics**: decision shortcuts such as the matching operation described elsewhere in the guide.

*Sources and further reading:* [Leverage and reciprocity](https://www.shortform.com/app/book/the-great-mental-models-volume-2/1-page-summary) · [Second-order decisions](https://papers.ssrn.com/sol3/papers.cfm?abstract_id=193848)

<!-- atlas:block-content b-b04fb47af3b5be83:end -->
<a id="theme-noise-judgment-and-groups"></a>

## Noise in organizations and groups

<!-- atlas:block-content b-26ecdff0fa3108ea:start -->
### Noise is unwanted variation in judgment

A group may agree and still be wrong. A system may average out errors across cases while treating particular people unfairly. Diagnosis must specify the judgment, the target and the cost.

<a id="b-26ecdff0fa3108ea"></a>
<a id="u-noise-77269065814d"></a>

<a id="v2-120-0"></a>

**Authorship and Scope of *Noise***

*Noise* is co-authored by Daniel Kahneman, Olivier Sibony, and Cass R. Sunstein, drawing on their combined expertise across psychology, strategy consulting, and legal scholarship. Kahneman is a Nobel Prize-winning psychologist and author of the award-winning *Thinking, Fast and Slow*; Sibony is a strategy professor, business consultant, and author of business strategy books; Sunstein is a legal scholar and co-author of the award-winning *Nudge*. The book also builds on decades of research and the authors' own work as noise consultants in business. It is written for a general audience, but is especially relevant to anyone leading an organization that relies on human judgment.

<a id="b-5a25a5f320f71dc2"></a>
<a id="u-noise-d75d6381aa2c"></a>
<a id="u-noise-477a98d01a61"></a>
<a id="u-noise-c011634b292d"></a>
<a id="u-noise-e2ac8aa7f692"></a>
<a id="u-noise-421b1aa8a251"></a>
<a id="u-noise-54dc1e3a495b"></a>
<a id="u-noise-c6bee37a5922"></a>
<a id="u-noise-0dc5cb184549"></a>
<a id="b-ca710ff48fabfe0d"></a>
<a id="u-noise-4993aa1e0464"></a>
<a id="u-noise-2788dae24afb"></a>
<a id="u-noise-1f19046002e3"></a>
<a id="u-noise-5bacf75f3319"></a>
<a id="u-noise-cb4212136033"></a>

<a id="v2-121-0"></a>

**Definition and Types of Noise in Judgment**

In *Noise*, by Daniel Kahneman, Olivier Sibony, and Cass R. Sunstein, **noise** is defined as unexpected and unwanted variance in human judgments. It is one of the two main errors in human judgment, the other being bias; reducing both improves judgment quality and thereby reduces unfairness, financial loss, and even loss of life.

A **judgment** is a mental attempt to assign a value to something in order to choose a course of action. The authors divide judgments into two broad categories:
- **Predictions** aim to come as close as possible to some correct value or answer. For example, insurance underwriters calculate a premium that is neither too low (which would lose the company money) nor too high (which would drive away customers), and doctors diagnose patients by identifying the correct cause of an ailment. In principle, predictive accuracy can be measured by comparing the prediction to the correct answer once it is known.
- **Evaluations** have no single correct answer; they require the decision-maker to balance pros and cons as best they can. Examples include a judge sentencing a criminal or deciding whether to grant asylum, or a teacher grading an essay. Because there is no correct answer, evaluative accuracy cannot be measured the same way predictive accuracy can.

Judgments are generally neither purely factual nor purely opinion-based: a doctor reading routine blood-panel results is not making a judgment (though she may make one about the cause of an anomaly), and a personal preference for one band over another is not a judgment either.

Shortform notes that a similar predictive balancing of risk and profit applies to any field resembling insurance, and that in the worst case, noise in such calculations can lead to a full-blown financial collapse. It also observes that measuring predictive accuracy is complex in its own right, because many predictions are too vague or too qualified to be meaningfully judged.

<a id="b-f22743f7adb5849f"></a>
<a id="u-noise-a8c8f40f3c4b"></a>
<a id="u-noise-e06b1915a0af"></a>

<a id="v2-122-0"></a>

**Consistency is one part of judgment quality**

We expect professional judgments to be accurate and fair, while recognizing that reasonable, qualified people can disagree. Noise concerns unwanted variation in comparable judgments—not every difference in preference, values, or materially different cases.

For numerical judgments against a fixed target under squared-error loss, mean squared error decomposes into squared bias plus variance. The two terms enter with equal weight; their actual magnitudes need not be equal. Reducing variance lowers this error measure if bias stays fixed. An intervention that also changes bias requires checking both effects. The authors emphasize noise because it is often neglected: its spread becomes visible across repeated judgments, whereas a persuasive story about one case can hide it.

<a id="b-5389d1f9fce54ad1"></a>
<a id="u-noise-f33fa8ece514"></a>
<a id="u-noise-9b351c61d950"></a>
<a id="u-noise-7817c949c0ec"></a>
<a id="u-noise-8f149fe1695b"></a>
<a id="u-noise-17fb1ca0203c"></a>
<a id="u-noise-c71dc647f3b1"></a>

<a id="v2-123-0"></a>

**The Target Metaphor: Noise, Bias, and Error**

To illustrate why both noise and bias matter, the authors use a shooting-range target metaphor: a judgment is the target, inconsistency equals noise, and inaccuracy equals bias. The four possible combinations are:
- Shots clustered tightly around the bullseye: both accurate and consistent (no bias, no noise).
- Shots clustered tightly but some distance from the bullseye: consistent but not accurate (no noise, but biased).
- Shots generally centered on the bullseye but not tightly grouped: generally accurate but inconsistent (not biased, but noisy).
- Shots spread widely and not centered on the bullseye: both noisy and biased.

![noise_noisytarget.png](https://short.zaynjarvis.com/media/source-notes/4c3a1ee118-noise_noisytarget.png)

If the target is flipped so you can see the shots but not the bullseye, you can no longer detect bias or accuracy, but you can still easily see noise, because noise refers to the spread of the shots. This means you can detect and correct for noise without knowing the correct answer to a prediction. It also means you can detect noise in evaluative judgments, where there is no correct answer by which to measure quality. In both cases, you can address noise without needing to know whether the judgments were biased, because noise is the degree of inconsistency between one judgment and the next.

We might be tempted to think that noise averages out over time. The authors argue that it does not: if the target is a just sentence, an accurate diagnosis, or a prudent business decision, every miss is costly, and these costs do not cancel each other out—they compound.

<a id="b-3575a4a88995b820"></a>
<a id="u-noise-2671551f4f5f"></a>
<a id="u-noise-1e7b4ab56ae0"></a>

<a id="v2-125-0"></a>

**Evaluative Judgment, Noise, and the Term “Judger”**

Shortform uses the generic term “judger” to describe anyone making a professional judgment, to avoid confusion with courtroom judges—who also qualify as judgers under this definition. This framing is especially relevant for evaluative judgments, where individual subjectivity naturally plays a role: we accept that some teachers and some judges are more or less lenient than others. Even so, we expect schools, courts, and other public institutions to be fair and consistent. A gap between a B+ from one teacher and an A- from another falls within normal variation, but an A from one grader and an F from another signals that something has gone wrong.

<a id="b-e3c361b862806998"></a>
<a id="u-noise-92b2b62cee1b"></a>
<a id="u-noise-88082113f5d8"></a>
<a id="u-noise-62f3313fddb1"></a>

<a id="v2-126-0"></a>

**Three Sources of Noise: Start With Level Differences**

Noise is variability in judgments, not all variability in outcomes. It can reflect stable differences between people as well as fluctuations within a person. The book distinguishes level noise and pattern noise; occasion noise is the transient component of pattern noise.

**Level noise** occurs when one person’s average judgments differ consistently from the average of all judgers. For example, some teachers grade more or less harshly than others over time, and some economic forecasters are more or less optimistic than others over time. Level noise describes each judger’s overall pattern compared with the overall average across all judgers.

Shortform notes that level noise may not be consistent over time and may itself be noisy. One study shows that graders inflate scores over time because they mistake their growing comfort with grading for an improvement in the quality of the work being graded. This effect does not invalidate *Noise*’s point so much as it illustrates how complicated the problem is.

<a id="b-e5bc167e95428fc0"></a>
<a id="u-noise-961b48cf7d30"></a>
<a id="u-noise-78a2125c2b8d"></a>
<a id="u-noise-9b147e725748"></a>
<a id="u-noise-762265517be3"></a>
<a id="u-noise-822011b86048"></a>

<a id="v2-127-0"></a>

**Pattern and Occasion Noise**

#### Pattern Noise
Pattern noise is the deviation that arises when a judger is unusually affected by a specific situation for one reason or another. For example, a forecaster who is generally more optimistic than most might become more pessimistic than her peers when evaluating a startup company. Pattern noise stems from people’s personalities and unique experiences. Some of it is stable over time, and some is transient—driven by current or recent circumstances. A forecaster who lost money on startups earlier in her career and is always cautious about them exhibits stable pattern noise. A forecaster who read about a spectacular startup failure that morning and feels cautious about startups right now exhibits transient pattern noise.

#### Occasion Noise
Occasion noise describes variability within a single person caused by numerous seemingly random factors. Studies cited by the authors show that judgments can be affected by: weather (college admissions officers weigh academic credentials more heavily on cloudy days and non-academic factors more heavily on sunny days); mood and sports (judges sentence more harshly after their local football team loses and more leniently after a win); time of day (doctors are more likely to prescribe opioids toward the end of the day, possibly because tired, stressed, rushed clinicians make diagnostic mistakes and reach for quick fixes); order in a series of judgments (judges become increasingly likely to deny asylum after granting it several times in a row, and vice versa, likely due to an unconscious drive to maintain balance); and order of information presentation (hearing that a politician is smart, driven, charismatic, and ruthless creates a different impression than hearing ruthless, charismatic, driven, and smart).

<a id="b-b5fa38379a640334"></a>
<a id="u-noise-cd0c78a48ab3"></a>
<a id="u-noise-6235532f14f1"></a>

<a id="v2-128-0"></a>

**How Level, Pattern, and Occasion Noise Combine**

For analytical purposes, noise in human judgment is divided into three types, though in practice any combination may be at work in a single situation. A shoplifting case sentenced by Judge Thompson illustrates all three:
- **Level noise**: Judge Thompson, on average, delivers much lighter sentences than his colleagues — a reason for the defendant to be optimistic.
- **Pattern noise**: Because his parents own a small retail business that has long struggled with shoplifting, he sentences shoplifters much more harshly than his peers — bad news for the defendant.
- **Occasion noise**: He has just returned from a relaxing vacation and is in an excellent mood that morning, so the defendant might still get a break.

Sentences would differ with a different judge, a different crime, or even a different day in court. This variability is one source of noise in the justice system, and the same dynamic applies to any system that relies on human judgment.

Shortform suggests a distinction between personally specific transient effects and circumstances that can affect many people. This should not become a separate definition: occasion noise need not affect everyone in the same direction, and shared situational influences need not cancel when judgments are averaged.

Occasion noise is particularly difficult to predict or notice in real time, and factors like weather cannot be standardized to ensure uniform judgments. Several techniques may help reduce its influence:
- Some forms of occasion noise, such as that caused by the order in which information is presented, can be controlled through structured decision-making strategies.
- Other forms, such as overprescription of opioids by fatigued doctors, may be prevented by watching out for information overload and excessive stress, though this is often hard to implement in practice.
- Factors like weather, mood, and time of day can be mitigated by having multiple judges assess a situation independently before comparing notes. Because different judges are affected by different situational factors, their judgments tend to balance each other out. Ideally, judges also follow a procedure that specifies which information to attend to — the sample hiring procedure discussed later in this guide offers one concrete illustration of how such a procedure can minimize occasion noise.

*Sources and further reading:* [Noise](https://www.littlebrown.com/titles/daniel-kahneman/noise/9780316451383/) · [Noise and financial risk](https://www.shortform.com/app/book/the-big-short) · [Testable predictions](https://www.shortform.com/app/book/superforecasting/chapter-3) · [Grading drift](https://www.sciencedaily.com/releases/2018/03/180312084826.htm) · [Overload and stress](https://www.shortform.com/app/book/blink/part-5)

<!-- atlas:block-content b-26ecdff0fa3108ea:end -->
<!-- atlas:block-content b-9f646c7789a04a75:start -->
### How interpretation and conversation introduce variation

<a id="b-9f646c7789a04a75"></a>
<a id="u-noise-0671fe09a9e4"></a>
<a id="u-noise-646e9feccefc"></a>
<a id="u-noise-2bf0de56351f"></a>
<a id="u-noise-414868f4e054"></a>
<a id="u-noise-5b5056934457"></a>
<a id="u-noise-bcfb3b8b31d9"></a>
<a id="u-noise-2fad7b8d51fa"></a>
<a id="u-noise-2edff7fb69c9"></a>
<a id="u-thinking-in-bets-5b1e415da7c4"></a>
<a id="u-thinking-in-bets-103001ca2a33"></a>

<a id="v2-129-0"></a>

**Psychological Sources of Noise: The Matching Operation**

Noise arises from several sources: personal biases such as preferences, backgrounds, affiliations, and beliefs, which produce level noise and pattern noise; more random factors like mood, weather, and the order or timing of decisions and information, which contribute to occasion noise; and fundamental features of how the mind perceives the world and how people behave in group settings.

One key psychological source is an intuitive **matching operation**: we attempt to predict or evaluate something by comparing it to similar things we know more about. This operation introduces noise because of the oversimplifications inherent in the procedure, as well as limits on our ability to discern fine quantitative differences. For example, we can easily tell whether it is sunny or cloudy, or whether it is hot or cold, but if we are asked to rank a series of temperatures from coldest to warmest, we quickly make mistakes. According to the authors, studies have found that we can rank things into about seven levels of quality or intensity before ranking errors set in. We do reasonably well when directly comparing one item to another, but given a set of items to rank or categorize, we make errors more easily than we might expect.

Any judgment that requires assessing things on a scale becomes noisier as the scale becomes less defined. Without proper context and a shared frame of reference for what values mean and how they should be assigned, judgers are forced to guess, making the judgment arbitrary; since each person guesses differently, the scale becomes noisy.

Shortform notes that this matching operation is a type of heuristic—a mental shortcut for solving a difficult problem quickly, in which the mind substitutes something easier or more familiar to generate an answer, a concept explored in *Thinking, Fast and Slow*.

<a id="b-9af182f3ab20a91d"></a>
<a id="u-noise-a600301d1418"></a>
<a id="u-noise-c91d1c9a5bd5"></a>
<a id="u-noise-f6ed384ae4f1"></a>
<a id="u-noise-38c20c8800a8"></a>

<a id="v2-130-0"></a>

**Psychological Source of Noise: Cause-and-Effect Thinking**

Kahneman, Sibony, and Sunstein identify cause-and-effect thinking as a major psychological source of noise in judgment. Because the human mind naturally approaches the world through causal narratives and tends to consider one case at a time, it struggles to recognize noise, which is a statistical property that emerges across many cases. This causal preference is further distorted by hindsight: once an outcome is known, people retroactively pick out factors from the situation and label them as the cause. Most events, the authors note, are neither completely surprising nor completely expected; because ordinary, "normal" events receive little conscious attention, they feel as though they would have been fully predictable, even though reliable prediction ahead of time would not have been possible. The result is a widespread sense that the causes of events are obvious and inevitable, and that future outcomes can be predicted—obscuring how arbitrary and contingent most events actually are from a predictive standpoint, since the relevant causes often only become visible at the same moment the outcome itself becomes known.

<a id="v2-129-1"></a>

**Social Sources of Noise: Information Cascades and Group Qualities**

Individual judgment already contains noise arising from how each person interprets the world. When people work in groups to reach judgments, social factors introduce additional sources of noise.

One major social source is the **information cascade**. Popularity—whether real or perceived—shapes how people evaluate information: an idea that gains early public support or popularity is more likely to prevail, regardless of its inherent merit. When one person voices an opinion, the next speaker is more likely to agree unless they have strong reasons to disagree, and the effect grows as more people align with the initial view. In group judgment settings where many members start undecided or hold mixed feelings, the final decision is usually driven by whichever opinion launched the cascade.

In *Thinking in Bets*, Annie Duke identifies three core qualities of a good group for learning from each other and improving decision-making. First, the group cares about accuracy: members call out each other's biases and engage in civil disagreement. Second, members hold each other accountable: they discourage each other from giving in to irrational or self-destructive impulses. Third, the group welcomes diversity of thought: different perspectives are important for generating new ideas and helping each other see what would otherwise be missed. These qualities directly counter the social dynamics that produce information cascades and other group noise.

<a id="b-bcb8c013f40d0647"></a>
<a id="u-noise-9f85d544611b"></a>
<a id="u-noise-f90c2bfd3268"></a>

<a id="v2-131-0"></a>

**Social Sources of Noise in Groups**

Because most people end up in agreement after a cascade, observers tend to assume the outcome was inevitable. It is not: a different starting point could have produced a different result. This path dependence is not obvious in everyday experience because each real-world situation only unfolds once.

Groups are also susceptible to polarization, meaning members shift toward more extreme versions of their initial opinions. For example, if every member of a hiring committee feels only mildly enthusiastic about candidate A at the outset, by the end of the meeting they may feel passionately excited about that candidate. Conversely, if some members are mildly enthusiastic about candidate A while others are mildly enthusiastic about candidate B, polarization can produce a stalemate in which half the committee strongly supports A and strongly opposes B, and the other half does the opposite.

<a id="b-558daf5e122bd42c"></a>
<a id="u-noise-ba3fc76a1dda"></a>
<a id="u-noise-c42425932176"></a>
<a id="u-noise-33a008d03968"></a>

<a id="v2-132-0"></a>

**Cognitive and Social Sources of Noise**

Many of the cognitive roots of noise echo ideas from Kahneman’s *Thinking, Fast and Slow*. One key mechanism is the human tendency to default to causal, narrative explanations even for random events — a pattern closely tied to the narrative fallacy. For example, if a company lays off employees by drawing names from a hat, observers who do not know the selection method will often invent coherent stories: if a friend is fired, they may recall a recent argument with the boss and assume a grudge; if someone they dislike is fired, they may attribute it to skill or character flaws. In both cases, observers perceive the outcome as logical and foreseeable, when it was in fact entirely random.

Information overload also amplifies noise. Both Malcolm Gladwell and Nate Silver point out that too much information leads to bad decisions, either because people fail to focus on what is most important, or because they become overwhelmed and fall back on familiar patterns and preconceived notions.

At the group level, social dynamics further amplify noise. Information cascades and polarization can reinforce each other, and their interaction helps explain groupthink: by random chance, the first speaker nudges the group in a particular direction and starts an information cascade, then polarization pushes the group decisively further in that direction — even if no member entered the meeting with strong convictions about any position. In *Originals*, Adam Grant argues that in corporate settings, groupthink also stems directly from company attitudes toward dissenting voices. Practices such as being transparent, inviting dissent, and choosing leaders who genuinely welcome criticism can reduce the risk of groupthink in decision-making. These principles are also relevant to the wisdom of crowds: crowd wisdom can only be tapped when a group includes people with diverse viewpoints and when those people feel free to voice their ideas.

*Sources and further reading:* [Blink: selecting information](https://www.shortform.com/app/book/blink/part-5#2-be-selective-with-the-information-you-consider) · [The Signal and the Noise](https://www.penguinrandomhouse.com/books/305826/the-signal-and-the-noise-by-nate-silver/) · [Originals](https://www.shortform.com/app/book/originals) · [Dissent in organizations](https://www.shortform.com/app/book/originals/chapter-7)

<!-- atlas:block-content b-9f646c7789a04a75:end -->
<!-- atlas:block-content b-cd33519b1c0423e6:start -->
### Audit, structure and check the judgment

<a id="b-cd33519b1c0423e6"></a>
<a id="u-noise-111175840adf"></a>
<a id="u-noise-a2c64f8b2ed9"></a>
<a id="u-noise-783267b99838"></a>

<a id="v2-134-0"></a>

**Detecting and Measuring Noise Through Audits**

Detecting and measuring noise is the first step in reducing it. Administrators generally assume their organizations make judgments consistently, and they may resist change until they see the problem directly. To quantify noise in a company, organization, or system, the authors of *Noise* describe a noise audit process used in their consulting work, with detailed guidelines provided in the book's appendix. In general terms, an organization gives a common set of sample cases to all members whose role requires making judgments about such cases—for example, an insurance company would distribute a set of sample claims to all its adjusters. The people being audited complete their judgments independently, and the results are then compared to measure the amount of variability across the organization.

<a id="b-d557ee036e4fda1f"></a>
<a id="u-noise-4a7386ed6ed7"></a>
<a id="u-noise-c73e570f3e8f"></a>
<a id="u-noise-cb60379627b4"></a>

<a id="v2-135-0"></a>

**Approaches to Reducing or Eliminating Noise**

The authors outline two broad categories of solutions for reducing or eliminating noise from judgments once it is understood: mechanical judgment tools—statistical models and computer algorithms—that can replace or augment human judgment, and a set of practices for reducing noise within human decision making itself.

<a id="b-ff46c6f1fe5fe9a6"></a>
<a id="u-noise-fef4f7b5ec0d"></a>
<a id="u-noise-956894bbb627"></a>

<a id="v2-136-0"></a>

**Decision Hygiene as a Framework for Improving Human Judgment**

The authors’ primary approach to improving human judgments is “decision hygiene”: consistent, preventative measures designed to minimize the chance of noise. Decision hygiene is a loose collection of suggestions, practices, and principles. With one exception—a sample hiring procedure—the authors do not prescribe a specific, systematic course of action; Shortform infers that organizations can implement whichever suggestions are relevant and practicable in their context.

<a id="b-98268d99d500a836"></a>
<a id="u-noise-f0d8dc4863f6"></a>
<a id="u-noise-e086733e2974"></a>
<a id="u-noise-7685ea23bcef"></a>
<a id="u-noise-0e88ef9f2817"></a>
<a id="u-noise-f166b302802d"></a>
<a id="u-noise-f11c0940e969"></a>
<a id="u-noise-9c7ef7df08cd"></a>
<a id="u-noise-c094fa79dd57"></a>
<a id="u-noise-c13ce1e7c0ee"></a>

<a id="v2-137-0"></a>

**Decision Hygiene: Better Scales, Judgers, and Statistical Thinking**

The authors of *Noise* outline a set of decision hygiene strategies for reducing noise, several of which address the psychological sources of noise directly.

#### Use interpretable scales
A numerical job rating is noisy when reviewers attach different meanings to “6.” Comparative judgments can help when raters share a relevant comparison set: for example, locating communication performance in the top 20% or next 20%. Rankings are not universally better; they can force distinctions in a nearly uniform group, change with the reference population and fail to show whether everyone improved. Specify the objective, reference group and concrete level descriptions before scoring.

A well-designed comparative scale also provides concrete **anchor points**—clear descriptions or markers for each level on the scale, where a specific value is tied to a relevant example of the thing being evaluated (such as a "C" grade representing average work on a paper). To minimize noise, anchor points must be provided in advance so every judge starts from the same frame of reference.

#### Choose and Train Better Judgers
The quality of human judgment can be meaningfully improved by selecting better judgers in the first place and by training existing judgers to improve their techniques and processes.

#### Think Statistically
Our normal, causal mode of thinking is prone to errors and biases that show up as noise. To improve accuracy, we must adopt a statistical perspective: rather than treating each case as entirely unique, we should classify it as a member of a broader category of similar cases. When predicting the likelihood of an outcome, we should first consider how common that outcome is across the entire category. For example, when predicting whether a particular student will graduate from college, we should start by knowing the overall graduation rate for all incoming college students.

<a id="b-12ce59bc11b5a76a"></a>
<a id="u-noise-1ae1905e6a77"></a>
<a id="u-noise-9e40d49a4a9a"></a>
<a id="u-noise-9a81a8fc14c3"></a>

<a id="v2-142-0"></a>

**Shortform perspective: Structured Judgment Tools: Scales and Decomposition**

Vague rating scales are a common source of noise. The problem can be understood through the lens of matching operations and question substitution from *Thinking, Fast and Slow*: without clear guidance, people faced with an ambiguous rating question such as “How would you rate X’s communication out of 10?” may unconsciously replace it with an easier, more subjective question like “How impressed am I by X’s communication?” or “How clear do I find X?”

One remedy is a set of predefined, behaviorally meaningful scale anchors. These give raters a common interpretation of scores. Shortform connects them to psychological anchoring, but the practical justification is clearer: specify relevant reference examples rather than leave each person to invent a scale. Arbitrary anchors can introduce shared bias even when ratings become more consistent.

Another debiasing technique is to break judgments into smaller components, which reduces the room for idiosyncratic interpretation. However, some fields may be inherently more noisy and more resistant to these techniques than others. Mental health, for example, is often cited as a field with high diagnostic inconsistency. While the authors attribute this inconsistency to the varying training and theoretical orientations of professionals, additional factors include overlapping and mutually influencing conditions (such as depression and anxiety), the difficulty of separating mental and physical health, and fundamental disagreements among professionals about whether a given set of symptoms constitutes a disorder or simply a difference. Reducing noise in such fields may require analysis and reform that go beyond the scope of standard noise-hygiene techniques.

<a id="b-5e2f4f72585958cc"></a>
<a id="u-noise-bf692bccaad3"></a>
<a id="u-noise-c0d21b7a9fbf"></a>
<a id="u-noise-86a0fc3f26ff"></a>
<a id="u-noise-aa56a81bfa89"></a>
<a id="u-noise-3d486c4e613b"></a>
<a id="u-noise-4728eaa2acbf"></a>

<a id="v2-138-0"></a>

**Decision Hygiene: Structuring Judgment and Information**

Two further decision hygiene strategies reduce noise by structuring how judgments are made and how information flows.

#### Break Judgments Into Smaller Components
It is easier to avoid noise when you break an overall decision into a set of smaller, more concrete subjudgments. Standardized procedures, checklists, guidelines, and assessments support this approach. For example, educators can reduce noise in essay grading by using rubrics: asking graders to assign separate scores for originality, logical clarity, organization, and grammar before calculating a final grade makes the judgment more straightforward. This decomposition also helps ensure that every judge follows the same procedures and attends to the same factors.

The authors acknowledge that this strategy is not perfect. In mental health, for instance, the DSM—a reference work designed to support and standardize psychiatric diagnoses—has done little to reduce diagnostic noise. One reason is that psychiatrists and psychologists tend to interpret signs and symptoms through the lens of their own training and background: different theoretical understandings of the mind and of mental disorders shape how different professionals interpret the facts presented to them.

#### Sequence Information Carefully
Because judgments are subject to influence from information, contextual clues, confirmation bias, and other factors, it is important to carefully control and sequence the information that judges receive. The authors provide several implementation guidelines:
- As a basic rule, judges should only be given what they need when they need it.
- Independent judgments must be truly independent: if the person verifying a result knows the first person’s conclusion, they are more likely to confirm it.
- Judges should document their conclusions at each step of the process; if new information leads them to change their decisions, they should explain and justify why.

<a id="b-f8b3f17bcd843105"></a>
<a id="u-noise-77860d102349"></a>
<a id="u-noise-4210aff182c7"></a>
<a id="u-noise-d5d42e6c8b7d"></a>

<a id="v2-139-0"></a>

**Selecting and Training Better Judgers**

When identifying good judgers, two considerations apply. In fields with objectively right or wrong outcomes, judgers can be evaluated by their measurable, objective results. In fields where expertise cannot be reduced to a simple metric, judgers in any domain can still be assessed by three traits correlated with better judgment: overall intelligence, cognitive style, and open-mindedness. The authors emphasize that intelligence alone is not sufficient; the other two traits are at least as important.

The authors further note that some members of the general population are superforecasters, whose predictions are consistently more accurate than those of the average trained expert; ideally, these are the people who should be hired or appointed as judgers. Superforecasters exhibit several identifiable traits that can be used both to select better judgers and to train current ones: they are open-minded; they are willing to update their opinions and predictions when new evidence arises; they naturally think statistically and spontaneously consider factors such as base rates; and they break problems down and evaluate components using probability rather than relying on a holistic "gut feeling" about the answer.

<a id="b-a1965bfa3baee418"></a>
<a id="u-noise-2507349a5aa6"></a>

<a id="v2-140-0"></a>

**Shortform perspective: Selecting and Training Better Judgers**

The quality of judgments depends heavily on who is doing the judging. *Noise*’s discussion of superforecasters draws on the work of Philip E. Tetlock and Dan Gardner in *Superforecasting*, where they frame the difference between strong and weak forecasters using the classic hedgehog-and-fox metaphor. Hedgehog personalities tend to view the world through the lens of one big idea, make snap judgments, and are extremely confident in their predictions. Fox personalities, by contrast, gather small bits of information across many topics, approach problems slowly and from multiple angles, and are cautious and qualified in their predictions. Tetlock and Gardner argue that foxes make better predictors than hedgehogs.

These fox-like skills can also be practiced: people can learn to recognize and avoid their own cognitive biases, generate multiple perspectives on a problem, and break problems down into smaller questions. These techniques are essentially the same as many of the recommendations in *Noise*.

<a id="b-6f981eb29e297e05"></a>
<a id="u-noise-b7525aa02d98"></a>

<a id="v2-133-0"></a>

**Statistical Thinking as a Noise-Reduction Tool**

Failure to think statistically is a major source of judgment error and noise, a theme central to *Thinking, Fast and Slow*. The core recommendation from that book — to explicitly account for base probabilities — aligns with the advice in *Noise*.

Shortform connects this with Nate Silver’s Bayesian approach: start from a prior, assess new evidence under competing hypotheses, and update without counting the same evidence twice. The calculation appears in the probability section. Silver’s suggestion that better Bayesian reasoning could have helped avert replication problems is a methodological argument, not proof that Bayes alone would prevent misconduct, weak designs or selective publication.

<a id="b-e1e030cb0d5a8df3"></a>
<a id="u-noise-e4a02c0b1d46"></a>
<a id="u-noise-529100141eee"></a>
<a id="u-noise-a6e8818a0601"></a>
<a id="u-noise-9cb94f157c05"></a>
<a id="u-noise-ca182f894ddc"></a>

<a id="v2-141-0"></a>

**Using Rules and Standards to Reduce Noise**

One strategy for reducing noise by breaking judgments into smaller parts is to implement rules and/or standards. The two approaches differ in structure and precision: rules offer explicit guidance typically tied to objective measures—for example, a maximum allowable blood alcohol content above which a driver can be charged with drunk driving. Standards are suggestive guidelines that require subjective interpretation and implementation—for example, law enforcement officers are trained to recognize potential signs of impairment such as erratic driving and to administer field sobriety tests.

When choosing between rules and standards, the authors advise first determining which approach will produce fewer errors. They also note that rules cannot always be implemented: rule-makers may be unable to agree, for instance due to political or moral differences, or they may lack the information needed to craft an appropriate rule. In some cases, the best approach is to combine rules and standards. Mandatory sentencing guidelines follow this model: they set a minimum and maximum sentence for a given crime (a rule) while leaving judges to determine a just sentence for each individual case (a standard).

*Sources and further reading:* [Overlapping mental-health conditions](https://www.psychologytoday.com/us/blog/the-power-rest/201807/psychiatric-diagnosis-is-difficult-and-so-is-treatment) · [Mental and physical health](https://www.mentalhealth.org.uk/a-to-z/p/physical-health-and-mental-health) · [Diagnostic disagreement](https://www.wired.com/story/mental-health-psychiatry-diagnosis-dsm/) · [Identifying superforecasters](https://www.shortform.com/app/book/superforecasting/1-page-summary) · [Hedgehog thinking](https://www.shortform.com/app/book/superforecasting/chapter-4#hedgehogs) · [Fox thinking](https://www.shortform.com/app/book/superforecasting/chapter-4#foxes) · [Recognizing cognitive bias](https://www.shortform.com/app/book/superforecasting/part-2#they-avoid-cognitive-biases) · [Generating perspectives](https://www.shortform.com/app/book/superforecasting/part-2#they-generate-multiple-perspectives) · [Decomposing questions](https://www.shortform.com/app/book/superforecasting/part-2#they-use-fermi-estimation)

<!-- atlas:block-content b-cd33519b1c0423e6:end -->
<!-- atlas:block-content b-9bc2c39dceb4dec4:start -->
### Aggregate without manufacturing agreement

<a id="b-9bc2c39dceb4dec4"></a>
<a id="u-noise-f70c4b506f57"></a>
<a id="u-noise-74138d769c1a"></a>
<a id="u-noise-d396ebee8831"></a>
<a id="u-noise-681abb505870"></a>
<a id="u-noise-b7a2dcbc28e3"></a>
<a id="u-noise-e12c317b0e0f"></a>

<a id="v2-143-0"></a>

**Aggregate independent judgments, then examine disagreement**

Averaging can reduce variation when errors are not perfectly correlated; it does not guarantee truth lies in the middle. Shared bias can survive aggregation. Seek relevant expertise and genuinely different information, use a common scale, and choose an aggregation rule suited to the loss being minimized.

The estimate–talk–estimate procedure has three stages:
1. Each person privately records an estimate before hearing others.
2. Participants explain evidence and reasoning and examine disagreements.
3. Each privately estimates again; aggregate these second estimates under the chosen rule.

Initial independence reduces cascades and polarization. Discussion is for finding evidence and mistakes, not forcing outliers toward consensus. Shortform notes that social pressure can still create false agreement; policy Delphi and argument Delphi explicitly emphasize a range of arguments or policy perspectives rather than making consensus the goal.

<a id="b-c8d6452608285916"></a>
<a id="u-noise-43db59dac2d9"></a>
<a id="u-noise-52ec5fd4ac95"></a>

<a id="v2-144-0"></a>

**Shortform perspective: Aggregating Judgments and the Wisdom of Crowds**

Averaging judgments before one decision differs from letting noisy decisions accumulate across people. The first can reduce variation under suitable conditions; the second does not repair the harm to someone already given an unusually severe sentence or a wrong diagnosis. Neither operation removes a shared bias automatically.

An individual can borrow part of the logic by seeking different sources and considering evidence against a favored answer—negative empiricism. Reading broadly and deliberately constructing alternatives can expose omissions. These perspectives still come through one mind and may share assumptions; they are not independent judges or independent samples merely because there are several of them.

<a id="b-d0d026ffddf3b720"></a>
<a id="u-thinking-in-bets-37deb5264d9b"></a>
<a id="u-thinking-in-bets-1853acbca70e"></a>
<a id="u-thinking-in-bets-cdd6c5699362"></a>
<a id="u-thinking-in-bets-6c41aa8688c7"></a>
<a id="u-thinking-in-bets-7ff18f090357"></a>
<a id="u-thinking-in-bets-071cb88e8fab"></a>
<a id="u-thinking-in-bets-f03218118143"></a>
<a id="u-thinking-in-bets-22dbe4a8f697"></a>
<a id="u-thinking-in-bets-ae37ea06ed5f"></a>
<a id="u-thinking-in-bets-801314c17fec"></a>

<a id="v2-150-0"></a>

**Group Norms for Better Judgment: The CUDOS Framework**

Once a group is formed, clear rules of engagement are needed to keep it focused on accuracy and avoid the social noise of cascades and polarization. Duke draws on sociologist Robert K. Merton's guidelines for how the scientific community should function, captured in the acronym CUDOS:
- **Organized skepticism**: All ideas are subject to scrutiny, criticism, and dissent. It is not argumentative or confrontational when everyone engages with their own uncertainty; the group should reward civil dissent and debate with approval, making healthy discussion part of its norms.
- **Communism (in Merton's sense)**: Data is commonly owned, belonging to everyone. In a decision group, sharing data means being honest about the factors that went into your decision, providing as much detail as possible, and omitting nothing—especially the things you are tempted to leave out.
- **Disinterestedness**: Don't let conflicts of interest or other biases influence the work. Such conflicts can include what Duke calls “resulting”: allowing knowledge of an outcome to influence how you evaluate the decision. One way to keep the group from being swayed by outcomes is to withhold the outcome until after the discussion happens.
- **Universalism**: All evidence must be treated by the same standard set of criteria, no matter what the source is. This does not mean you cannot evaluate whether a source is credible, but you should not disregard information without a fair evaluation just because you have negative feelings about the source—such as a news outlet with political leanings you disagree with or a colleague you find annoying.

<a id="b-82d789a06e9b00d3"></a>
<a id="u-thinking-in-bets-fa5010130986"></a>
<a id="u-thinking-in-bets-50bc257877f5"></a>

<a id="v2-151-0"></a>

**Group Accountability for Better Decisions**

Self-critique is an important skill, but other people can help you see your blind spots. They bring their own unique life experiences and give you the chance to view ideas from angles you had not considered before. However, you first need to find someone willing to have those discussions: not every person has the bandwidth to give serious feedback on your life choices, or wants their own decisions or beliefs picked apart in return. That is why forming a group around good decision-making practices matters.

*Sources and further reading:* [Consensus pressure](https://mpra.ub.uni-muenchen.de/4999/1/MPRA_paper_4999.pdf) · [Policy Delphi](https://journals.sagepub.com/doi/pdf/10.2304/pfie.2013.11.6.755#:~:text=A%20policy%20Delphi%20seeks%20to,problem%20%28Turoff%2C%202002%29.) · [Argument Delphi](https://ieeexplore.ieee.org/document/7089162) · [Negative empiricism](https://www.shortform.com/app/book/the-black-swan/chapter-3#flaw-1-the-error-of-confirmation)

<!-- atlas:block-content b-9bc2c39dceb4dec4:end -->
<!-- atlas:block-content b-6f431bd5b481f867:start -->
### Models, interviews and the limits of standardization

<a id="b-6f431bd5b481f867"></a>
<a id="u-noise-87f0e3818925"></a>
<a id="u-noise-facf24b6ef97"></a>
<a id="u-noise-e107d23d8201"></a>
<a id="u-noise-d94a8d5a5670"></a>
<a id="u-noise-ac59e0aef1b3"></a>
<a id="u-noise-6277f59bf058"></a>
<a id="u-noise-9e7d3c61fb91"></a>
<a id="u-noise-fdf25243df88"></a>

<a id="v2-145-0"></a>

**Mechanical Judgment: Statistical Models and Algorithms**

#### Statistical Models
A statistical prediction rule consistently combines selected inputs, often through a weighted formula. A college-graduation model might use high-school grades, test scores, extracurricular activity, and parents’ education. In the settings reviewed in Noise, simple models often outperform unaided judgment. Even rough or randomly chosen weights can work surprisingly well when applied to relevant predictors whose directions are correctly specified; this is not a claim that arbitrary inputs or arbitrary models always beat people. Consistency removes one source of error, while predictor selection, measurement, training data and validation still determine accuracy. Subjective adjustments can add noise, so their claimed benefit should be tested on new cases.

#### Computer Algorithms
Here the authors discuss **learning-based prediction algorithms**, one part of the much broader category of computer algorithms. Some learn patterns from large training datasets and use them to make predictions; neither large datasets nor learning define algorithms in general. Learning can reveal useful relationships, but accuracy must be tested on relevant new cases. The authors warn that biased training data can reproduce bias—for example, a recidivism model can inherit distortions in criminal-justice records. Data, objectives, evaluation and use all matter.

Shortform cites real-world examples: after years of development, Amazon discovered its recruitment algorithm systematically favored men over women, and Facebook's advertising algorithms have been criticized for helping spread everything from fake news to hate speech.

<a id="b-9ad98e8f761f7d93"></a>
<a id="u-noise-4ec80a57f2fe"></a>

<a id="v2-146-0"></a>

**The Role of Mechanical Judgment Relative to Human Judgment**

Although the authors describe several mechanical judgment methods, they are more interested in reducing noise in human judgment than in replacing human judgment entirely with machines. This is partly because, as they note, the comparison considered here concerns prediction tasks people already attempt, where consistent rules can improve accuracy. The authors frame the effectiveness of mechanical judgments less as a blanket solution to noise and more as a demonstration of how much benefit noise reduction itself can deliver.

<a id="b-628320439c8d4b09"></a>
<a id="u-noise-ac00ef2cb2fa"></a>
<a id="u-noise-6afc25b6c87b"></a>

<a id="v2-147-0"></a>

**Shortform perspective: Mechanical Judgment and Hybrid Approaches**

Although the authors of *Noise* ultimately favor improving human judgments over replacing them, they may not state this priority as clearly as they could. They devote considerable space to explaining models and algorithms and defending them against criticism, which can create a misleading impression of how central algorithms are to the book’s proposed agenda — especially since some reviewers have focused on the dangers of algorithms as a major critique of *Noise*.

Shortform, drawing on Nate Silver’s baseball discussion, raises a possible role for combining models with scouting. Its account describes teams emphasizing measurable offensive performance after *Moneyball*, while inadequate measures of defense left important performance poorly represented; conceding runs could offset gains elsewhere. It also describes teams combining statistical models with scouts’ observations, including judgments about psychological demands. These observations suggest complementary information worth testing, not a controlled comparison proving that hybrids cause the greatest success or that scouts reliably observe what models cannot. Ask whether the added judgment contributes information absent from the model, how accurately it does so, and whether the combined rule improves predictions on new cases. This leaves the warning about subjective noise intact.

<a id="b-2ad91c83a1e8ffe0"></a>
<a id="u-noise-c582246e7421"></a>
<a id="u-noise-cae33ef145ad"></a>
<a id="u-noise-479c82cde46c"></a>
<a id="u-noise-19193e369d4a"></a>

<a id="v2-148-0"></a>

**Example: Google’s Structured Hiring Procedure**

To illustrate how decision hygiene practices apply in a real-world setting, the authors provide an overview of Google’s hiring process, which proceeds as follows:
1. Determine what skills are most important to the position you are hiring for.
2. Develop scales to measure each candidate on each skill identified in step 1.
3. Interview each candidate multiple times with different interviewers—Google uses four interviews. The purpose of the interviews is to rate candidates on their skills. Interviews must be conducted independently; interviewers cannot compare notes yet.
4. The hiring team meets to discuss their results, review the data they have amassed, and finally share their impressions and make an overall decision.

While this procedure specifically describes a hiring process, the authors note that it can easily be adapted to other types of business decisions, such as whether to make an investment or whether to acquire or merge with a rival company.

<a id="b-40059953edbc524d"></a>
<a id="u-noise-33cb19318229"></a>

<a id="v2-149-0"></a>

**Shortform perspective: Decision Hygiene Techniques in Google’s Hiring Procedure**

In a “Putting It All Together” analysis, Shortform maps Google’s hiring procedure onto several core decision hygiene techniques—connections the book’s authors note but do not lay out in a clear, methodical way:
- **Decompose the judgment**: Step 1 breaks the larger question “Whom should we hire?” into smaller, discrete components.
- **Use better scales**: Step 2 applies insights about scale design to ensure high-quality, consistent data collection.
- **Sequence information and enforce rules and standards**: Step 3 keeps interviews truly independent of one another, supported by clear rules and standards that keep interviews consistent and focused.
- **Aggregate independent judgments**: Step 4 brings multiple interviewers together to reach a joint decision.
- **Delay subjective impressions**: At a broader level, the whole process sequences information by asking interviewers to hold off on subjective reactions until the final step. This preserves room for personal responses to candidates, but mediates those reactions through the hard data collected earlier and the collective wisdom of multiple interviewers—each of whom has already formed an independent view of the candidate.

<a id="b-8cf084e7729d066c"></a>
<a id="u-noise-ecd197d6d935"></a>
<a id="u-noise-25a8a6de21a8"></a>
<a id="u-noise-eb5056a2072f"></a>
<a id="u-noise-b74f00dbaa37"></a>

<a id="v2-153-0"></a>

**Shortform perspective: Conceptual Debates Around Noise**

Critics of *Noise* have raised several conceptual and statistical objections to the book’s framework. Some argue its core claims are a repackaged version of earlier work or common-sense ideas. While the specific definition of noise as variance in human judgment appears novel, the broader statistical concept of noise has a long history: Fischer Black framed noise (in contrast to information) as a fundamental component of economic models, and Nate Silver defined it as junk data opposed to meaningful signal in his work on prediction. The book also builds extensively on Daniel Kahneman’s earlier *Thinking, Fast and Slow*; many sources of noise trace back to the cognitive errors and biases described there, and *Noise* can be read as an exploration of how those errors scale up within systems and organizations. Shortform notes that *Noise* acknowledges these connections but mostly glosses over them, and that it spells them out more fully because doing so helps clarify the relevant concepts.

A conceptual dispute concerns which disagreements count as unwanted error when a judgment evaluates values rather than predicts an observable event. The mean-squared-error decomposition is a mathematical identity relative to a specified target; the target need not be known to derive it, though estimating accuracy requires an appropriate target or evidence. If the target, loss function or legitimate values differ, reducing variance alone cannot settle the disagreement. Keep that normative question separate from the arithmetic.

A further caution is that low noise should not be mistaken for accuracy. Noise reduction only tightens the spread of judgments, bringing them closer to each other; it does not guarantee they are closer to the true target. A target-shooting metaphor illustrates this: shots could cluster tightly around a point far from the bullseye. Silver makes a parallel point in *The Signal and the Noise*, using a nearly identical graphic to represent precision rather than accuracy. He warns that mistaking precision for accuracy can have devastating consequences, as seen in the 2008 financial crisis, where high precision (low noise) created an illusion of correctness that masked uncertainty and bias.

<a id="b-2c084e21610fdfc3"></a>
<a id="u-noise-fbea07e75b45"></a>

<a id="v2-154-0"></a>

**Shortform perspective: Treating Evaluations Like Predictions**

Shortform highlights what it regards as a genuinely original insight from the book: we should treat evaluations the same way we treat predictions.

*Noise* distinguishes between two types of judgments: predictive judgments (such as forecasting a stock’s future value) and evaluative judgments (such as grading an essay). It is commonly accepted that evaluative judgments are inherently subjective, that there is no correct answer against which to measure their quality, and therefore that there is no way to improve the quality and accuracy of evaluations. There is also far less literature on improving evaluations than on improving predictions.

Yet if we accept *Noise*’s arguments that evaluations and predictions are the same kind of thing — both are judgments, both suffer from noise to the same degree and for the same reasons, and reducing noise is beneficial — then it follows that we can improve our evaluative judgments. We can do so by applying to evaluations the same advice many authors have already offered for making better predictions. The book makes this point early on, but it deserves special emphasis because the insight appears to be truly original.

*Sources and further reading:* [Amazon’s hiring-model bias](https://www.theguardian.com/technology/2018/oct/10/amazon-hiring-ai-gender-bias-recruiting-engine) · [Facebook’s recommendation effects](https://fivethirtyeight.com/features/facebooks-algorithm-is-broken-we-collected-some-spicy-suggestions-on-how-to-fix-it/) · [Critiques of algorithms in Noise](https://www.theguardian.com/books/2021/jun/03/noise-by-daniel-kahneman-olivier-sibony-and-cass-sunstein-review-the-price-of-poor-judgment) · [Moneyball](https://www.shortform.com/app/book/moneyball) · [Noise: originality critique](https://a-ortmann.medium.com/on-noise-the-book-87b419585fd9]) · [Earlier work on noise](https://statmodeling.stat.columbia.edu/2021/05/23/thinking-fast-slow-and-not-at-all-system-3-jumps-the-shark/) · [Noise: common-sense critique](https://www.prospectmagazine.co.uk/arts-and-books/signal-failure-daniel-kahnemans-fascinating-and-flawed-new-book-noise) · [Fischer Black on noise](https://onlinelibrary.wiley.com/doi/full/10.1111/j.1540-6261.1986.tb04513.x)

<!-- atlas:block-content b-6f431bd5b481f867:end -->
<a id="theme-systems-feedback-and-leverage"></a>

## Systems, feedback and intervention

<!-- atlas:block-content b-ca026d2ccecaf3e2:start -->
### Start with the stock, the flows and the purpose

A system diagram is useful when it identifies a mechanism you can trace: what accumulates, what changes it, what feeds back, and how long the response takes.

<a id="b-ca026d2ccecaf3e2"></a>
<a id="u-thinking-in-systems-c0b0912e5d1c"></a>
<a id="u-thinking-in-systems-6bb6ae0f7e72"></a>
<a id="u-thinking-in-systems-08e8f634f496"></a>
<a id="u-thinking-in-systems-86110fd003dd"></a>
<a id="u-thinking-in-systems-1cc0b897d2a7"></a>
<a id="u-thinking-in-systems-94306fc5d7fe"></a>
<a id="u-thinking-in-systems-8c30c8a27970"></a>
<a id="u-seeking-wisdom-bevelin-c1d5c918b828"></a>
<a id="u-seeking-wisdom-bevelin-4ba1ce9e811e"></a>
<a id="u-seeking-wisdom-bevelin-52efadce10b5"></a>
<a id="u-seeking-wisdom-bevelin-85673a11195e"></a>
<a id="u-seeking-wisdom-bevelin-a89c157f8846"></a>
<a id="u-thinking-in-systems-63bcc3718960"></a>
<a id="u-thinking-in-systems-8279a273d36f"></a>
<a id="u-thinking-in-systems-53fef6fbe7ca"></a>
<a id="u-thinking-in-systems-eec4e9139842"></a>
<a id="u-thinking-in-systems-a6cf8eb4f586"></a>
<a id="u-thinking-in-systems-9559ad8fdc6d"></a>
<a id="u-thinking-in-systems-19de267b115d"></a>
<a id="u-thinking-in-systems-ddc7df9da879"></a>
<a id="u-thinking-in-systems-6443eaab1833"></a>
<a id="u-seeking-wisdom-bevelin-cb198275ce92"></a>
<a id="u-seeking-wisdom-bevelin-5848a34167fd"></a>
<a id="u-seeking-wisdom-bevelin-b6031e8f18c0"></a>
<a id="u-seeking-wisdom-bevelin-3f35ac408b33"></a>
<a id="u-seeking-wisdom-bevelin-43f888869078"></a>

<a id="v2-155-0"></a>

**Definition and Core Components of Systems**

A system is a set of elements interconnected in a way that achieves its function. It is composed of three core components: elements (the individual things in the system), interconnections (the relationships between the elements), and purpose or function (what the system achieves). Many things in the world operate as systems. For example, a football team consists of players on the field with specific interacting roles, plus coaches, support staff, and fans; within a corporation, people, machines, and information work together to achieve the corporation's goals, and the corporation itself sits within the larger system of the economy.

According to Peter Bevelin in *Seeking Wisdom*, a system is a complex web of interdependent components—for example, a business with its employees, processes, and clients. Systems thinking means analyzing how those interdependent parts affect the whole, and its core principle is that every action takes place within a larger system, so you must always consider the far-reaching consequences of what you do. A failure to practice systems thinking is a major cause of flawed decision-making and unintended consequences.

*Thinking in Systems* is an introduction to systems analysis. Many aspects of the world operate as complicated systems rather than simple cause–effect relationships, and many real-world problems arise from defects in how those systems function. Understanding how systems work and how to intervene in them is key to producing the changes you seek.

<a id="b-7da68df63a655d71"></a>
<a id="u-thinking-in-systems-03f053296110"></a>
<a id="u-thinking-in-systems-3bc74dbdc5a4"></a>
<a id="u-thinking-in-systems-30b89c9b19d0"></a>
<a id="u-thinking-in-systems-c69e14e82b5c"></a>
<a id="u-thinking-in-systems-d58a2654769f"></a>
<a id="u-thinking-in-systems-ce9c2df25d2d"></a>
<a id="u-thinking-in-systems-7382c101446e"></a>
<a id="u-thinking-in-systems-eee242b18acf"></a>
<a id="u-thinking-in-systems-58c83b1658c6"></a>
<a id="u-thinking-in-systems-3eecc74f1861"></a>

<a id="v2-156-0"></a>

**Stocks and Flows**

Stocks and flows are the foundation of every system. A stock represents the elements in a system that you can see, count, or measure—commonly thought of as an inventory, a store, or a backlog. Flows are the means by which stocks change over time: inflows increase the level of stock, while outflows decrease it.

A simple illustration is a bathtub: the stock is the amount of water in the tub, the inflow is water coming from the faucet (which raises the stock), and the outflow is water leaving through the drain (which lowers the stock). This relationship can be drawn as a stock-and-flow diagram.

Many real systems are analogous to the bathtub. For fossil fuels, the stock is the underground reservoir; mining reduces the stock, while natural geological processes slowly increase it. The world population is a stock of people, growing through births and shrinking through deaths.

![thinking-in-systems-stock-flow.png](https://short.zaynjarvis.com/media/source-notes/25147338d3-thinking-in-systems-stock-flow.png)

<a id="b-c485276435fa0729"></a>
<a id="u-thinking-in-systems-7087907e7829"></a>
<a id="u-thinking-in-systems-23d9f8fa0136"></a>
<a id="u-thinking-in-systems-9115a7d29125"></a>
<a id="u-thinking-in-systems-25ffe617c77e"></a>
<a id="u-thinking-in-systems-a37799c81497"></a>
<a id="u-thinking-in-systems-6fa62d88e90d"></a>
<a id="u-thinking-in-systems-cb2f46bf4bab"></a>

<a id="v2-157-0"></a>

**Stocks as Buffers, Delays, and Shock Absorbers**

Stocks take time to change. While you can adjust a flow almost instantly — turning on a faucet takes only a second — filling a bathtub takes minutes. Stocks change gradually because it takes time for the flows to actually flow; as a result, they act as buffers, delays, lags, and shock absorbers in a system.

From a human perspective, this slowness has both benefits and drawbacks. On the positive side, stocks represent stability: they allow inflows and outflows to be out of balance for a period of time. A bank account, for example, stores money and gives your life stability — if you lose your job and your income inflow stops, you can draw on that stock to keep living while you figure out a solution.

On the negative side, a slowly changing stock means things cannot change overnight. If a population’s skills are made obsolete by technology, for instance, you cannot re-educate the workforce instantaneously; it takes time for information to work its way through the system and for new skills to flow through to the population.

<a id="b-653e54f00e5de0e0"></a>
<a id="u-thinking-in-systems-904eb0edfeac"></a>
<a id="u-thinking-in-systems-a38ebaaebe99"></a>
<a id="u-thinking-in-systems-2f55c9f0cb61"></a>

<a id="v2-158-0"></a>

**Feedback Loops as Drivers of Persistent System Behavior**

Systems often produce behavior patterns that persist over time. In some cases a system appears self-correcting, with stocks hovering around a steady level; in others it spirals out of control, either rising exponentially or shrinking rapidly. When a behavior pattern is this persistent, it is typically governed by a feedback loop. Loops form when changes in a stock alter the flows into or out of that stock. The broader category of feedback loops includes both balancing loops, which stabilize systems, and reinforcing loops, which drive runaway change.

<a id="b-8be0d31526a62e40"></a>
<a id="u-thinking-in-systems-ff04cdfa47e0"></a>
<a id="u-thinking-in-systems-cd355be7aca0"></a>
<a id="u-thinking-in-systems-0f7165fa3cf7"></a>
<a id="u-thinking-in-systems-d9315de65f30"></a>
<a id="u-thinking-in-systems-47e20f287d89"></a>
<a id="u-thinking-in-systems-32a0b94835a7"></a>
<a id="u-thinking-in-systems-db3f01d2f101"></a>
<a id="u-thinking-in-systems-3ecfd57e9db5"></a>

<a id="v2-159-0"></a>

**Reinforcing Feedback Loops and Runaway Dynamics**

Reinforcing feedback loops — also called positive feedback loops, vicious cycles, virtuous cycles, flywheel effects, snowballing, compound growth, or exponential growth — amplify rather than dampen change in a stock, causing it to grow faster or shrink faster. The mechanism works in both directions: as a stock level rises, its inflow increases (or its outflow decreases), pushing the stock even higher; conversely, as a stock level falls, its inflow decreases (or its outflow increases), pushing the stock even lower.

Reinforcing loops appear in both positive (growth) and negative (collapse) forms. Positive examples include world population growth, where more people produce more offspring, which further increases the population stock, and a growing national economy, where more factories and workers enable greater production, and greater production funds more factories and better education, which in turn expand productive capacity. Negative examples include soil erosion in agriculture, where eroded soil reduces root growth, which causes still more erosion, and panic buying during emergencies, where a shrinking stock of a good such as toilet paper makes people buy it more fervently, depleting the stock even faster.

<a id="b-3d7c8a7be601ff38"></a>
<a id="u-thinking-in-systems-86d4a5a742f5"></a>
<a id="u-thinking-in-systems-6a176fb87ead"></a>
<a id="u-thinking-in-systems-be92e9e3cca5"></a>
<a id="u-thinking-in-systems-210d91c64910"></a>
<a id="u-thinking-in-systems-241252f1f158"></a>
<a id="u-thinking-in-systems-5019afbd4787"></a>

<a id="v2-162-0"></a>

**Balancing Feedback Loops**

Balancing feedback loops—also called negative feedback loops or self-regulation—operate around an acceptable setpoint for a stock. If the stock moves relative to this acceptable level, the flows adjust to push it back toward the setpoint. If the stock dips below the level, inflows increase and outflows decrease to raise the stock. If the stock rises above the acceptable level, inflows decrease and outflows increase to lower the stock.

An intuitive example is keeping a bathtub's water level steady: if the level is too low, you plug the drain and turn on the faucet; if the level is too high and water spills over, you open the drain and turn off the faucet.

<a id="b-1e342f845e638947"></a>
<a id="u-thinking-in-systems-540a6a94b3ed"></a>
<a id="u-thinking-in-systems-aa1e7592d18a"></a>
<a id="u-thinking-in-systems-faf318baf5b4"></a>
<a id="u-thinking-in-systems-afe33eae164b"></a>
<a id="u-thinking-in-systems-5da2b8e71460"></a>
<a id="u-thinking-in-systems-574b8d9f8686"></a>
<a id="u-thinking-in-systems-7a3efe05fc64"></a>
<a id="u-thinking-in-systems-618b156dab12"></a>
<a id="u-thinking-in-systems-37a40487408b"></a>

<a id="v2-163-0"></a>

**Example: World Population as a System**

World population illustrates a stock with births as an inflow and deaths as an outflow, measured in people per unit time. More people can generate more births and deaths if the corresponding per-person rates stay fixed. The diagram simplifies those dependencies into reinforcing and balancing loops. Shortform calls zero a setpoint for deaths; here that is a modeling interpretation, not a claim that populations deliberately seek zero.

The net flow determines the immediate direction: more births than deaths means growth; the reverse means decline; equality means no net change. Exponential growth specifically requires a constant positive net per-capita rate in this simplified model. Merely observing more births than deaths does not establish that growth law. A fall in births, rise in mortality, or both can reverse the direction.

Various circumstances can shift the relative strength of the birth or death loop:
- Data suggest that as countries get wealthier, birth rates fall. Poorer countries with high current birth rates may therefore not retain those high rates as their economies develop.
- A lethal, contagious disease can drastically increase the death rate. During the HIV/AIDS epidemic, for example, population projections for areas with high HIV prevalence had to account for higher mortality.
- Birth rates can also fall due to social factors, such as reduced interest in raising children or widespread fertility issues.

![thinking-in-systems-population.png](https://short.zaynjarvis.com/media/source-notes/c7bc17344a-thinking-in-systems-population.png)

<a id="b-39a63c8266981329"></a>
<a id="u-thinking-in-systems-bc0daee1a4e4"></a>
<a id="u-thinking-in-systems-3d41d2c22c2f"></a>
<a id="u-thinking-in-systems-72e3c21de40b"></a>
<a id="u-thinking-in-systems-5ad481246c3d"></a>
<a id="u-thinking-in-systems-684e0de63afa"></a>

<a id="v2-164-0"></a>

**Building More Complex System Models**

From basic building blocks—stocks, flows, and feedback loops—you can build up to more complicated systems that model the real world. A short overview covers only one simple system in detail to show how systems analysis reveals its behavior, but the full treatment explores a range of more complex models:
- One stock plus two balancing loops represents a thermostat keeping a room's temperature stable.
- Delays introduce oscillations into system behavior, as when a car sales manager tries to keep inventory consistent.
- Models of extracting a non-renewable resource, such as fossil fuels.
- Models of extracting a renewable resource, such as fish in the sea.

<!-- atlas:block-content b-ca026d2ccecaf3e2:end -->
<!-- atlas:block-content b-b78307de6d7a78c4:start -->
### Intervene in the loop that sustains the problem

<a id="b-b78307de6d7a78c4"></a>
<a id="u-thinking-in-systems-35615b824542"></a>
<a id="u-thinking-in-systems-50e935d91c7b"></a>
<a id="u-thinking-in-systems-a35d766c9bf6"></a>
<a id="u-thinking-in-systems-4b8995368876"></a>
<a id="u-thinking-in-systems-6064d69c57c7"></a>

<a id="v2-160-0"></a>

**Escalation as a Multi-Actor Reinforcing Loop**

Escalation — also known as “keeping up with the Joneses” or an arms race — is a reinforcing feedback loop in which two or more competitors each hold an individual stock and aim to possess the largest stock of the group. When one competitor falls behind, it pushes hard to catch up and become the new winner; the higher one stock rises, the higher all the other stocks are pushed to reach in turn. The dynamic can continue at enormous cost to every participant until one or more parties drops out or collapses. Historical examples include the Cold War, where the Soviet Union and the United States monitored each other’s munitions and strove to amass the larger arsenal at a cost of trillions of dollars; a more mundane example is competitive advertising, where rivals’ campaigns grow increasingly pervasive and aggressive in an effort to capture more attention.

<a id="b-d78cbf8f6e812f07"></a>
<a id="u-thinking-in-systems-b8a59bc3dcda"></a>
<a id="u-thinking-in-systems-a496a4660b0b"></a>
<a id="u-thinking-in-systems-83773e28cfe4"></a>
<a id="u-thinking-in-systems-f7807a1271a1"></a>

<a id="v2-161-0"></a>

**Interventions to Break Escalation**

Dampen the loop through which competitors respond to one another. A credible mutual agreement can cap escalation even when neither side gets everything it wants. Unilateral restraint is another possibility when rivals mainly seek a relative advantage and you can tolerate the resulting gap. It does not ensure they will be satisfied: hostile aims, opportunism or an inability to withstand the gap can make withdrawal dangerous. The intervention depends on what actually drives the feedback.

<a id="v2-155-2"></a>

**Addiction: The Burden-Shifting Feedback Loop**

“Shifting the burden to the intervenor” is a reinforcing dependence loop: outside help relieves a symptom while local capacity to address its cause weakens; the recurring problem then calls for more help. The mechanism requires that loss of capacity, not merely the presence of assistance. The summary uses elder care as an illustration: smaller homes and reduced family care skills can increase reliance on institutions. Professional care and social support can also build capability or meet needs a family cannot safely meet. Their existence alone does not establish a harmful loop or justify withdrawing them.

Meadows’s three-part intervention logic is to diagnose why the system cannot address the root problem, design help that strengthens rather than replaces that capacity, and plan a withdrawal or transition when the system can sustain the function. The timing follows restored capacity and the consequences for people affected; prompt withdrawal regardless of those conditions can worsen the problem.

<a id="b-cecc9bed4979ad86"></a>
<a id="u-thinking-in-systems-139709272321"></a>
<a id="u-thinking-in-systems-23d26eab994f"></a>
<a id="u-thinking-in-systems-a159d3e3985a"></a>
<a id="u-thinking-in-systems-306905f99140"></a>

<a id="v2-172-0"></a>

**Common System Problem Archetypes**

Systems with similar structures tend to fall into recurring problem archetypes. Three well-known patterns are policy resistance, the rich-get-richer dynamic, and drift to low performance.
- **Policy resistance** describes situations where a policy appears to have little effect because the actors in the system push back against its influence; the war on drugs is a classic example.
- **The rich-get-richer** pattern occurs when winners capture an ever-greater share of limited resources and progressively outcompete losers, as seen with market monopolies.
- **Drift to low performance** arises when performance standards are tied to past performance rather than absolute benchmarks, creating a vicious cycle of declining standards—for example, a business losing market share consoles itself each time by thinking, "well, it's not that much worse than last year."

<!-- atlas:block-content b-b78307de6d7a78c4:end -->
<!-- atlas:block-content b-bd10f37fc4ecd525:start -->
### Capacity, hierarchy and moving bottlenecks

<a id="b-bd10f37fc4ecd525"></a>
<a id="u-thinking-in-systems-23be567534ab"></a>
<a id="u-thinking-in-systems-1aa7de3151b5"></a>
<a id="u-thinking-in-systems-2f7f82423231"></a>
<a id="u-thinking-in-systems-1a42a652fe73"></a>
<a id="u-thinking-in-systems-ce18daf1cb66"></a>
<a id="u-thinking-in-systems-1ea3e0e2e3cb"></a>
<a id="u-thinking-in-systems-f7855a819fe4"></a>

<a id="v2-165-0"></a>

**Hierarchy in Systems**

In a hierarchy, subsystems are grouped under a larger system. For example, individual cells in the body are subsystems of an organ; organs are subsystems of the body as a whole; and a person is a subsystem of larger systems such as family, company, and community. In an efficient hierarchy, subsystems work well more or less independently while serving the needs of the larger system, and the larger system's role is to coordinate between subsystems and help them perform better. This layered arrangement improves efficiency because each subsystem can manage its own internal affairs without heavy coordination with other subsystems or the larger system.

Problems can arise at both levels: if a subsystem optimizes only for itself and neglects the larger system, the whole system can fail—for instance, a single cancerous cell optimizes for its own growth at the expense of the larger human system. Conversely, if the larger system exerts too much control, it can suppress self-organization and efficiency.

<a id="b-de23bf606eeab3dd"></a>
<a id="u-thinking-in-systems-ce87e09172d7"></a>
<a id="u-thinking-in-systems-446c584b3672"></a>
<a id="u-thinking-in-systems-95595042fa95"></a>
<a id="u-thinking-in-systems-84987b42b7ed"></a>
<a id="u-thinking-in-systems-73291f3f63b9"></a>

<a id="v2-166-0"></a>

**Three Properties of Strong Systems**

Strong systems have three properties that allow them to accomplish their purposes remarkably well, persist for long periods without particular oversight, and survive environmental changes: resilience (the ability to bounce back after being stressed), self-organization (the ability to make itself more complex), and hierarchy (the arrangement of a system into layers of systems and subsystems). Creating systems that ignore these three properties leads to brittleness, causing systems to fail under changing circumstances.

<a id="b-4808ed3378f007e7"></a>
<a id="u-thinking-in-systems-32059a878895"></a>
<a id="u-thinking-in-systems-d94e65d357cc"></a>
<a id="u-thinking-in-systems-1c48ab45344e"></a>
<a id="u-thinking-in-systems-52246f21da3b"></a>
<a id="u-thinking-in-systems-497b3dd38ded"></a>

<a id="v2-167-0"></a>

**Resilience and Feedback Loops**

Resilience can be understood as the range of conditions under which a system can continue functioning normally; the wider that range, the more resilient the system. The human body, for instance, resists disease from foreign agents, repairs itself after injury, and survives across a wide range of temperatures and dietary conditions.

The stability that underpins resilience comes from feedback loops operating at multiple layers of abstraction: baseline loops that restore the system, often with redundant loops that can substitute for one another through different mechanisms and on different time scales; meta-feedback loops that restore the baseline feedback loops themselves; and even meta-meta feedback loops that generate improved meta-loops and feedback structures.

Systems are sometimes designed with goals other than resilience in mind. A common tradeoff is optimizing for productivity or efficiency by eliminating feedback loops that appear unnecessary or costly. This can make the system highly brittle: it narrows the range of conditions under which normal operation is possible, so that even minor disturbances can push the system out of balance.

<a id="b-3613691dd33d00c6"></a>
<a id="u-seeking-wisdom-bevelin-5ac7644f84f4"></a>
<a id="u-seeking-wisdom-bevelin-fecf8c885b22"></a>
<a id="u-thinking-in-systems-1d1877633d32"></a>
<a id="u-thinking-in-systems-6e994acca7da"></a>
<a id="u-thinking-in-systems-3e1de4ef90cd"></a>
<a id="u-thinking-in-systems-0bb27f25ca08"></a>
<a id="u-thinking-in-systems-7d51e8f5de6c"></a>

<a id="v2-168-1"></a>

**Self-Organization, Adaptation, and the Costs of Suppression**

Self-organization means that a system is able to make itself more complex, which allows it to diversify, adapt, and improve itself. Earth’s biology is a self-organizing system: billions of years ago, a soup of chemicals in water formed cellular organisms, which then evolved into multicellular organisms and eventually into thinking, talking humans. Some organizations quash self-organization, possibly because they optimize for performance and seek homogeneity, or because they fear threats to stability; this can explain why some companies reduce their workforces to machines that follow basic instructions and suppress disagreement. Suppressing self-organization can weaken a system’s resilience and prevent it from adapting to new situations.

<a id="v2-155-1"></a>

**Limiting Factors and System Performance**

A limiting factor constrains performance at the current level of other inputs. Improving a nonbinding factor may yield little, while improving the constraint can help until a different constraint takes over. This does not mean every bottleneck failure destroys the whole system or every improvement elsewhere is worthless.

Bevelin illustrates the concept of limiting factors with a household example: imagine you recently had a child and hope to continue growing your family. The city you live in is a limiting factor in your home system, because if you keep paying such high rent, you may never be able to afford a larger home to accommodate a larger family. To change this limiting factor, you could move to a new city with lower rent so you can afford a larger living space.

<a id="b-38cdee2b7d9238ff"></a>
<a id="u-seeking-wisdom-bevelin-246a3b886fd2"></a>

<a id="v2-170-0"></a>

**Limiting Factors and Common Errors in System Analysis**

While Bevelin claims that we can avoid errors by identifying and improving a system’s limiting factors, he does not explore in depth what types of errors we make when we fail to consider limiting factors. In *Thinking in Systems*, Donella Meadows identifies two major such errors. First, people sometimes misidentify the limiting factor, which prevents them from supporting the system’s growth or improvement. For instance, you might initially misidentify the small size of your apartment as the factor limiting your family’s growth, but moving to a larger apartment would only increase your rent and potentially make a second child unaffordable — the real limiting factor is the city’s sky-high rent. Second, limiting factors change over time, especially as a system grows; a system may fail to reach its goals if you assume the limiting factor stays the same. For example, after you move your family to a new city with lower rent and can afford a larger home, new factors may limit your family’s potential growth, such as the cost of education in your new neighborhood.

*Sources and further reading:* [Thinking in Systems](https://short.zaynjarvis.com/books/thinking-in-systems/)

<!-- atlas:block-content b-bd10f37fc4ecd525:end -->
<!-- atlas:block-content b-bbeb6017c1caccc0:start -->
### Follow the consequences beyond the first effect

<a id="v2-155-3"></a>

**Far-Reaching and Unintended Consequences**

Because people operate inside complex systems, their actions can ripple outward in ways they did not foresee; failing to anticipate those ripples can produce negative outcomes elsewhere in the system. Bevelin illustrates this with the example of a family system. Suppose you accept a promotion to earn more money and better provide for your family. The new role keeps you so busy that you spend less time at home; your partner is left with more parenting responsibilities, creating tension in the relationship; your children, upset by that tension, begin acting out at school.

To raise the chance that your actions produce intended outcomes, Bevelin advises trying to predict their far-reaching consequences. Before taking the promotion, for instance, you could discuss the impact with your partner and brainstorm ways to lighten their parenting load—such as asking other family members for childcare help or enrolling the children in an engaging after-school program.

<a id="v2-168-0"></a>

**System Complexity, Unintended Consequences, and Safety Buffers**

Because systems are extremely complex, it is impossible to examine the entire system surrounding an action or to predict every effect it will have. Peter Bevelin argues that instead of trying to do the impossible, you should expect that your actions will have some unintended negative consequences and plan for those outcomes. He shares a strategy from Warren Buffett for planning for unexpected negative results: build safety factors into your predictions. A safety factor is a buffer added to a prediction; if the prediction turns out to be too optimistic, the buffer helps prevent disaster. Buffett says his company will not buy a stock if they estimate its value is only slightly higher than its price; instead, they build a safety factor into purchasing decisions by buying only when they estimate the value will be significantly higher than the price.

<a id="b-bbeb6017c1caccc0"></a>
<a id="u-seeking-wisdom-bevelin-dfe343a8949f"></a>
<a id="u-seeking-wisdom-bevelin-052a04ea3e05"></a>
<a id="u-seeking-wisdom-bevelin-246021feb209"></a>
<a id="u-seeking-wisdom-bevelin-747ef0166e01"></a>

<a id="v2-169-0"></a>

**Unintended Consequences, Resilience, and System Visualization**

Actions within systems can produce far-reaching, unintended effects. The unpredictability of such consequences is sometimes called “the butterfly effect,” a term coined by a meteorologist who found that tiny changes in initial conditions could produce large, unexpected outcomes elsewhere in a system; he reasoned that the flap of an extra butterfly’s wings in one part of the world could contribute enough air current to trigger a tornado elsewhere.

To improve your ability to predict far-reaching consequences, Donella Meadows offers a visual technique in *Thinking in Systems*: draw a diagram of the system by mapping its parts and their connections. This forces you to notice how elements interrelate, making it easier to see how one change might trigger another. For example, a manager restructuring a department could create an organizational chart to visualize hierarchies and the relationships among employees.

Meadows’s ideas on resilience also offer another way to plan for unintended negative consequences, especially useful for situations that do not involve numeric predictions — such as planning to improve your mental health — where Buffett’s safety factor approach, which works well for adjustable calculated predictions, is less applicable. Meadows argues that the best way to protect a system against unexpected outcomes is to make it resilient in advance. A resilient system performs well across a wide range of situations, both positive and negative, and has built-in backup mechanisms that act as a safety net. For example, you can build resilience in your mental health by investing in multiple layers of support: therapy, medication, and a personal support network. If one layer fails unexpectedly — for instance, if you forget to refill your medication — your other backup systems can help prevent a crisis.

<a id="v2-169-1"></a>

**Small-Scale Systems in Daily Life**

Most of Bevelin’s examples involve large, complex systems such as businesses, factories, organizations, and economies, but smaller systems also shape daily life significantly, and the same logic of actions and consequences applies to them. The human body is a system of interdependent organs, fluids, and hormones that serve the whole; changing one part can affect another, as when a medication prescribed for one problem causes side effects such as insomnia. A family is a system of relationships in which harsh words to one member can send a signal that harshness is acceptable, leading others to be unkind and creating a negative ripple effect. Everyday habits also form a system: in *Atomic Habits*, James Clear describes a system as a set of behaviors that together produce an outcome. For someone aiming to lose weight, a system might include healthier eating, more sleep, and more frequent exercise; the positive feelings from exercise can then reinforce the other healthy habits.

<a id="b-8fc1828f7db5060e"></a>
<a id="u-thinking-in-systems-f398becdb229"></a>
<a id="u-thinking-in-systems-74a85b1350f6"></a>
<a id="u-thinking-in-systems-b65b935b10e4"></a>
<a id="u-thinking-in-systems-3f8a4ab7d831"></a>
<a id="u-thinking-in-systems-bde15c30c346"></a>
<a id="u-thinking-in-systems-8768d4d2003e"></a>
<a id="u-thinking-in-systems-3be539fac1ec"></a>
<a id="u-thinking-in-systems-04fa80c3079f"></a>

<a id="v2-171-0"></a>

**Cognitive Biases in Perceiving Systems**

People study systems in order to predict their behavior and identify the most effective ways to change them, yet systems frequently behave in ways that defy expectations. At the root of this gap is a fundamental cognitive limitation: human brains favor simplicity and can only process a finite amount of complexity. We also tend to reason in straightforward cause-and-effect terms and over short time horizons, which keeps us from seeing the full consequences of our interventions. These blind spots prevent us from perceiving systems as they actually are, from designing robust systems, and from intervening productively.

One recurring cognitive bias is a tendency to focus on stocks more than flows, and within flows, to pay more attention to inflows than outflows. For example, when thinking about global population growth, people naturally fixate on rising birth rates as the driver and give less thought to how declining mortality from improved medical care also expands the population. Similarly, a company trying to grow its headcount will instinctively hire more people, rather than devoting equal thought to reducing the outflow of employees who quit or are fired. This asymmetry is one illustration of how our drive for simplicity leads us to overlook system complexity and develop incomplete ideas about how to intervene.

<a id="b-a2a927ca5fcc1d77"></a>
<a id="u-thinking-in-systems-80b4165a7dfe"></a>
<a id="u-thinking-in-systems-3af74a354275"></a>
<a id="u-thinking-in-systems-d3af9f44228d"></a>
<a id="u-thinking-in-systems-bfcfe6b69ac3"></a>
<a id="u-thinking-in-systems-e88fd08fe8dc"></a>

<a id="v2-173-0"></a>

**Practicing Systems Thinking**

Learning to think in systems is a lifelong pursuit. The world is endlessly complex, so there is always something new to learn; even when you believe you have a solid grasp of a system, it can behave in surprising ways that force you to revise your model. And even if you understand a system well and are confident about what should change, implementing that change is an entirely separate challenge.

Guidance for becoming a better systems thinker includes the following practices:
- **Observe behavior first to understand the system.** Research its history to learn how it reached its current state; gather data, chart key metrics over time, and tease out the relationships between them.
- **Expand your boundaries.** Think across both short and long timescales—consider how the system might behave ten generations from now. Draw on multiple disciplines: understanding complex systems requires familiarity with fields as broad as psychology, economics, religion, and biology.
- **Articulate your model.** As you build understanding, put your ideas on paper by drawing a system diagram that lays out the system's elements and how they connect. Drawing the diagram makes your assumptions about the system and its workings explicit.
- **Expose your model to other credible people and invite their feedback.** They will challenge your assumptions and push you to deepen your understanding. Admitting your mistakes and redrawing your model trains your mental flexibility.
- **Decide where to intervene.** Most interventions focus on tweaking mere numbers within the system structure, such as department budgets or national interest rates. Far higher-leverage intervention points exist, such as weakening the effect of reinforcing feedback loops, improving the system's capacity for self-organization, or resetting the system's goals.
- **Probe your intervention to its deepest human layers.** When investigating why interventions fail, you may encounter deep questions about human existence. You might lament that people in the system are blind to obvious data and assume the problem would vanish instantly if only they saw things as you do. But this raises a more fundamental question: How does *anyone* process the data they receive, and how do people view the same data through very different cognitive filters?

*Sources and further reading:* [Butterfly effect](https://www.technologyreview.com/2011/02/22/196987/when-the-butterfly-effect-took-flight/) · [System diagrams](https://short.zaynjarvis.com/books/thinking-in-systems/#improving-as-a-systems-thinker) · [Organizational charts](https://www.lucidchart.com/blog/types-of-organizational-structures) · [Medication side effects](https://www.goodrx.com/conditions/generalized-anxiety-disorder/insomnia-cant-sleep-due-to-anxiety-treatments)

<!-- atlas:block-content b-bbeb6017c1caccc0:end -->
<a id="theme-mental-models-and-second-order-effects"></a>

## Mental models across domains

<!-- atlas:block-content b-2c7aca579cc017a1:start -->
### Use a model by stating its premises

A metaphor can suggest a question. It cannot transfer a mathematical result or a causal claim without the assumptions that made the original result work.

<a id="b-2c7aca579cc017a1"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-626d1eb9307d"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-60f72ff430b3"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-007229c593d3"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-9a288d8c99f0"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-eb916b8319db"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-82d4ede8c248"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-486f2901c241"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-842d0e6a4d03"></a>
<a id="b-964ba847ce6db64d"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-1d5e2795bffb"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-cda13a323a55"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-e921ee890d00"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-eff2892baa9f"></a>

<a id="v2-174-0"></a>

**Mental Models as a Framework for Decision-Making Under Uncertainty**

Every day, people face a steady stream of decisions ranging from trivial choices like what to eat for breakfast to high-stakes ones like which career path to pursue. Because the information available is often messy and ambiguous, the best choice is rarely self-evident. According to Gabriel Weinberg and Lauren McCann in *Super Thinking: The Big Book of Mental Models*, the remedy for this uncertainty is mental models.

Weinberg and McCann define mental models as concepts and patterns that help us understand various situations across distinct topics. For instance, the mental model of supply and demand dictates that the relationship between supply and demand determines prices: if supply exceeds demand, the price decreases, and if demand exceeds supply, the price increases. They argue that mental models provide a framework for optimizing decisions amid uncertainty, and that using them well can lead to *super thinking*—the ability to accurately understand the world and its underlying mechanisms. Super models eliminate misconceptions and inefficiencies in reasoning, instead offering reliable patterns and heuristics for analyzing situations. By improving the ability to evaluate information, they strengthen decision-making; when used in concert, disparate super models form a toolkit that can inform decisions in virtually any situation.

A model developed in one field can suggest a mechanism elsewhere, which Weinberg and McCann call a “super model.” Critical mass in nuclear physics is the quantity of fissile material sufficient under specified conditions to sustain a chain reaction. The business analogy concerns participation feedback: a dating app may become more attractive once a useful pool of compatible users exists. Geography, matching quality, retention and competition matter; reaching a single customer count does not guarantee explosive growth.

<a id="b-1f33659174aa5354"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-804e2177c4c9"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-554ed0b7fe68"></a>

<a id="v2-175-0"></a>

**Shortform perspective: Where Definitions of Mental Models Differ**

In *The Great Mental Models Volume 1*, Shane Parrish and Rhiannon Beaubien describe mental models as representations of how things work. By contrast, Weinberg and McCann use the term more broadly, encompassing a wide array of concepts beyond mere representations of how things work; they even classify generic phenomena such as multitasking and organizational culture as mental models. Parrish and Beaubien further argue that mental models serve an even broader purpose: cultivating wisdom. Although the exact nature of wisdom is unclear, they hold that it involves the ability to see creative solutions to various problems and is therefore practical—it helps us know what to do in the face of unexpected challenges.

<a id="b-ad63843c52e887a5"></a>
<a id="u-clear-thinking-d45f46a0b527"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-e6f65858937a"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-fdbd6527f738"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-262967cdb57e"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-6ad886620c92"></a>

<a id="v2-176-0"></a>

**Shortform perspective: Who Built These Toolkits—and How the Guide Maps Them**

Two major bodies of work on mental models frame the field through different authorial backgrounds and structural approaches.

Shortform’s guide opens its coverage of *Super Thinking* by explaining what mental models are, where they come from, and why they are valuable in decision-making. Weinberg is the founder and CEO of DuckDuckGo, a search engine valued at over $100 million, and contributes business expertise to the book’s discussions of mental models for successful companies. McCann is a widely published statistician who conducted research for pharmaceutical giant GlaxoSmithKline, and brings expertise to formal decision-making models. The book covers more than 300 mental models. Shortform's guide organizes its coverage in four parts: a broad introduction to mental models; models for avoiding pitfalls in reasoning; constructive models to apply instead; and models for building successful businesses. Shortform's supplementary perspectives expand on—and occasionally push back against—the models the authors recommend.

Shane Parrish, a Canadian entrepreneur and writer, is the founder of [Farnam Street](https://fs.blog/), a popular blog and media company dedicated to [helping people improve their decision-making and practice lifelong learning](https://fs.blog/start/). He developed expertise in decision-making and teamwork while working as a cybersecurity expert for a Canadian intelligence agency. With fellow intelligence agent Rhiannon Beaubien, Parrish has written *The Great Mental Models*, a four-volume collection covering [general thinking](https://www.shortform.com/app/book/the-great-mental-models-volume-1); [physics, chemistry, and biology](https://www.shortform.com/app/book/the-great-mental-models-volume-2); [systems and mathematics](https://www.shortform.com/app/book/the-great-mental-models-volume-3/1-page-summary); and economics and art. Parrish also hosts *[The Knowledge Project](https://fs.blog/episodes/)* podcast, where he interviews experts from business, science, and other fields.

<a id="b-6b2eed7926d8b6fe"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-592dc93bb4b3"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-91e090db578e"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-f8306dba9879"></a>

<a id="v2-178-0"></a>

**The rigidity of intuitive and conventional reasoning**

Weinberg and McCann warn that conventional thinking and intuition can carry assumptions into settings where they no longer work. Bloodletting is their historical example. The practice has a long, changing history; humoral theory—organized around blood, phlegm, black bile and yellow bile—provided an influential rationale for many uses. The summary compresses that history into roughly 3,000 years and a late-nineteenth-century rejection. Those dates should not imply one unchanged practice sustained solely by one belief. Challenges to its effectiveness and changes in medical evidence helped displace its broad use. The decision lesson is to examine the support for a familiar practice, rather than treat familiarity as validation.

<a id="b-da2ac89790024721"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-842287a52e6b"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-d26850c7e4b3"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-b8886b720841"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-6029dd2f71b9"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-5aa9c0fda155"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-639d33125b3e"></a>

<a id="v2-179-0"></a>

**Reasoning From First Principles and De-Risking Assumptions**

Reasoning from first principles makes the starting premises explicit and builds an argument from them. The summary’s career premise—“upward mobility is valuable”—is a person’s value judgment, not a self-evident truth everyone must share. Clear premises make reasoning easier to inspect; they do not guarantee truth or a good choice. Copernicus provides a historical case of questioning a conventional Earth-centered model, not proof that one obvious principle settled astronomy. His model retained assumptions such as uniform circular motion and epicycles, and its assessment belongs to a longer history of mathematical modeling and evidence. See [Stanford Encyclopedia of Philosophy: Copernicus](https://plato.stanford.edu/entries/copernicus/).

Even seemingly self-evident first principles can be wrong, however, so Weinberg and McCann offer a second mental model to guard against false assumptions infiltrating your reasoning: *de-risk* your assumptions by testing their validity with objective measures. The concrete form of de-risking varies by assumption. For example, a conservative political candidate who assumes they will comfortably win historically Republican districts might de-risk that assumption by commissioning an opinion poll.

<a id="b-b545ab69a03db2f5"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-edc9a355684f"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-ad13434985a0"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-03191eb0ebfb"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-59d75d77398f"></a>

<a id="v2-180-0"></a>

**Shortform perspective: First-Principles Reasoning: Philosophical Roots, Limits, and Social Backlash**

Shortform connects first-principles reasoning with Descartes’s search for secure foundations in Meditations on First Philosophy, while questioning whether all starting assumptions can be self-evident. Euclidean geometry offers a useful distinction: in an equivalent formulation of the parallel postulate, exactly one line through a point outside a given line is parallel to it. Coherent non-Euclidean geometries alter that assumption. The lesson is to state and test a model’s premises where possible, rather than call an assumption true merely because it feels obvious.

Because conventional beliefs are deeply ingrained, unconventional conclusions derived from first principles often provoke social backlash. Galileo, for instance, was branded a heretic by the Roman Catholic Church and sentenced to lifelong house arrest for endorsing the Copernican view that Earth revolves around the Sun. People defer to conventional thinking in the first place because it is convenient and lets them avoid grappling with the complexity of the real world, as Steven Levitt and Stephen Dubner observe in *Freakonomics*. Because conventional thinking typically rests on anecdotes rather than quantitative data, it breeds misconceptions, and appealing to conventional wisdom often pulls us away from the truth.

Shortform also points out an important limitation of de-risking assumptions: it works well for empirical assumptions but is far less practical for value judgments. For instance, Weinberg’s company DuckDuckGo operates on the assumption that promoting internet privacy is valuable, and PETA assumes that speciesism—the view that certain species are inherently superior—is wrong. It is unclear how to de-risk either assumption, since we cannot run experiments to determine the truth of speciesism or the inherent value of privacy. When your assumptions cannot be de-risked, it is advisable to proceed with caution.

<a id="b-cd5887e738d42120"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-0c89082c7f7d"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-63d4d8fa2930"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-0801c47d553d"></a>

<a id="v2-181-0"></a>

**Fundamental attribution error and charitable interpretation**

Another reasoning problem arising from our frame of reference is the fundamental attribution error: we misattribute others' motivations to character or personality rather than external factors. For example, if our server seems cold, we might conclude they are rude and impolite, rather than considering that they might have just had a bad day. To counter this error, Weinberg and McCann encourage seeking the most respectful interpretation of others' behaviors—essentially, viewing others' actions charitably rather than cynically. This advice aligns with Hanlon's Razor, a mental model stating that we should not attribute someone's behavior to malice if it can instead be attributed to carelessness.

<a id="b-26f53ec1475bb0ec"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-9c50dbdc65ad"></a>

<a id="v2-184-0"></a>

**Shortform perspective: Hanlon’s Razor and Its Limits**

Hanlon’s Razor is only intended as a general heuristic, and it can lead us astray in situations where malice is a better explanation than carelessness, even if both are possible. In *Pride and Prejudice*, Jane Bennett exemplifies the danger of misusing Hanlon’s Razor: when Caroline Bingley repeatedly tries to sabotage Jane’s budding relationship with her brother, Jane repeatedly—and implausibly—gives Caroline the benefit of the doubt.

*Sources and further reading:* [Super Thinking](https://www.penguinrandomhouse.com/books/562923/super-thinking-by-gabriel-weinberg-and-lauren-mccann/) · [NRC: critical mass](https://www.nrc.gov/reading-rm/basic-ref/glossary/critical-mass) · [The Great Mental Models, Vol. 1](https://www.shortform.com/app/book/the-great-mental-models-volume-1) · [The Great Mental Models, Vol. 1: summary](https://www.shortform.com/app/book/the-great-mental-models-volume-1/1-page-summary) · [Farnam Street](https://fs.blog/) · [Farnam Street’s purpose](https://fs.blog/start/) · [Mental Models, Vol. 2](https://www.shortform.com/app/book/the-great-mental-models-volume-2) · [Mental Models, Vol. 3](https://www.shortform.com/app/book/the-great-mental-models-volume-3/1-page-summary) · [The Knowledge Project](https://fs.blog/episodes/) · [Bloodletting and humoral theory](https://www.medicalnewstoday.com/articles/bloodletting-why-doctors-used-to-bleed-their-patients-for-health) · [Bloodletting’s decline](https://bcmj.org/premise/history-bloodletting) · [Earth-centered astronomy](https://www.britannica.com/science/Ptolemaic-system) · [Copernican model discussion](https://www.uu.edu/dept/physics/scienceguys/2004Jan.cfm#:~:text=Copernicus%20adopted%20a%20heliocentric%20view,circles%20around%20the%20fixed%20Sun.) · [Cartesian foundationalism](https://plato.stanford.edu/entries/descartes-epistemology/#Foun) · [Euclidean parallel postulate](https://www.britannica.com/science/parallel-postulate) · [Non-Euclidean geometry](https://www.famousscientists.org/gauss-and-non-euclidean-geometry/#:~:text=Gauss%20realized%20that%20self%2Dconsistent,publish%20any%20non%2DEuclidean%20geometry.) · [Galileo’s trial](https://www.history.com/this-day-in-history/galileo-is-accused-of-heresy) · [Freakonomics](https://short.zaynjarvis.com/books/freakonomics/) · [DuckDuckGo’s privacy values](https://duckduckgo.com/) · [Speciesism](https://www.peta.org/about-peta/#:~:text=Our%20Mission%20Statement&text=PETA%20opposes%20speciesism%2C%20a%20human,and%20in%20the%20entertainment%20business.)

<!-- atlas:block-content b-2c7aca579cc017a1:end -->
<!-- atlas:block-content b-d9fe917eb66113a5:start -->
### Long-term effects, shared resources and incentives

<a id="b-d9fe917eb66113a5"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-3d5149845361"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-96df15481333"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-1357964e6a97"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-c82c22523ec1"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-bca5f9dffb79"></a>

<a id="v2-185-0"></a>

**Short-Termism, the Boiling Frog, and the Precautionary Principle**

The boiling-frog story is a metaphor for overlooking gradual deterioration, not a reliable account of frog behavior. Weinberg and McCann use it to explain short-termism: actions that improve current results can undermine future capacity. A company that markets only existing products while neglecting development may look efficient now but lack a viable next generation of products. Track the slow-moving condition itself, rather than waiting for each small change to feel dramatic.

To combat short-termism, Weinberg and McCann recommend heeding the *precautionary principle*: act with extreme caution when an action’s potentially harmful consequences are unknown. More concretely, this means analyzing potential risk by considering the harmful consequences an action could lead to, and asking yourself whether there is reason to think the action has dangerous consequences. If the answer is yes, pause before acting. By following this principle, they argue, we are less likely to encounter harmful consequences in the future.

<a id="b-1c39b72af8414add"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-80093f6492bd"></a>

<a id="v2-186-0"></a>

**Shortform perspective: The Precautionary Principle Versus Expected-Utility Theory**

Although the precautionary principle seems intuitive, legal scholars have argued that it has untenable consequences. In particular, when all possible decisions have potentially harmful consequences, the precautionary principle leads to stalemate. For example, critics of nuclear power object to its potentially catastrophic consequences, while the alternative—fossil fuels—might lead to devastating climate change. The precautionary principle offers little guidance in such situations.

More broadly, these cases illustrate a tension between the precautionary principle and expected-utility theory, the view that the rational decision in a given context maximizes utility—that is, it yields the highest net benefit on average. Because high-risk decisions sometimes have sufficiently high rewards, expected-utility theory might endorse decisions that the precautionary principle would caution against. For example, the potentially vast rewards of transitioning to nuclear energy could justify doing so despite the risks.

To reconcile these approaches, some have argued that the precautionary principle is simply a helpful practical heuristic, not a foolproof law of decision theory—in other words, a useful guideline even if theoretically questionable.

<a id="b-a7f5123514664e3c"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-62ad0dc2f581"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-cf5791ac8f3f"></a>

<a id="v2-188-0"></a>

**Collective harm from individually reasonable choices**

The tyranny of small decisions describes individually attractive choices accumulating into an undesirable collective outcome. Voting illustrates a collective-action problem: the small perceived influence of one vote can discourage participation, although a democracy needs participation in aggregate. The tragedy of the commons specifically concerns a shared, rival resource, such as a fish stock, where individual extraction can deplete the whole. Groceries bought during early Covid-19 panic are private goods, not public goods; simultaneous stockpiling can still create a coordination failure and temporary shortages. Public goods, such as a broadly available warning signal, instead raise problems including free riding. Keep these mechanisms distinct when choosing remedies.

<a id="b-a34facfebb629aaa"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-f8dcfdb1d760"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-0d7ca459f231"></a>

<a id="v2-187-0"></a>

**Shortform perspective: Short-Termism, Cultural Variation, and Collective Harm**

Shortform notes research traditions comparing short- and long-term orientations across cultures, often contrasting the United States with East Asian societies. Such group-level tendencies depend on context and measurement; they do not determine an individual’s horizon or make every institution in a region behave alike.

Kant’s universalizability test adds an ethical question: could one rationally will the maxim behind an action to apply universally? A practice of lying can undermine the trust it depends on. That is an argument within his moral framework, not a proof that every collectively costly action is irrational under every account of decision theory. Asking what happens if everyone follows the same rule can reveal collective effects in voting, truth-telling and resource use, while those cases retain different mechanisms.

<a id="b-f5edccd9a8af8af3"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-6bb6111f38df"></a>

<a id="v2-183-0"></a>

**Shortform perspective: Framing Effects and the Ethics of Nudging**

Thaler and Sunstein’s nudge changes choice architecture without forbidding options or materially changing economic incentives. Defaults and presentation can alter attention and behavior while leaving alternatives accessible. A small menu of suggested tips illustrates framing and anchoring, provided custom amounts remain possible; an automatic paid subscription after a trial may exploit inertia if terms or exit are obscure. Germany’s bottle-return system includes a refundable deposit, not simply a payment for recycling, and its financial incentive should be distinguished from a pure nudge.

Choice architecture can serve the provider instead of the chooser. High default tip percentages may anchor a payment and create social pressure; this is not best described as availability bias alone. Assess transparency, ease of choosing otherwise, whose goal is served and whether the user can meaningfully decline.

<a id="b-a0fb528d5b478326"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-e7918dc9a314"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-5b54683c8d6b"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-ba7850396f83"></a>

<a id="v2-194-0"></a>

**Deterrence and the carrot-and-stick model**

To prevent conflict from arising in the first place, deterrence is a useful strategy: it involves using a threat to dissuade the opposing side from starting a conflict. For example, parents of rebellious children might deter them from sneaking out by threatening to ground them if they are caught.

The carrot-and-stick model combines rewards for desired conduct with costs for undesired conduct. In behavioral terminology, negative punishment specifically removes a valued stimulus; “negative” does not mean every unpleasant penalty. A professor’s attendance-related grade changes illustrate incentives, but their effects depend on credibility, fairness, what grades are meant to measure, and possible gaming. Combining levers does not guarantee cooperation or make threats harmless.

<a id="b-b0cedf8a5ccd13ee"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-943e4fa9b1c1"></a>

<a id="v2-195-0"></a>

**Shortform perspective: Positive Reinforcement Versus Punishment**

How well positive reinforcement works compared to punishment depends on the goal you are trying to achieve. In general, positive reinforcement is more effective at encouraging action, while punishment is more effective at encouraging inaction. That said, some argue this asymmetry does not hold for children, for whom rewards consistently work better than punishment.

<a id="b-88eeda180a1b1062"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-903e196c43b3"></a>

<a id="v2-189-0"></a>

**Shortform perspective: Containment: Benefits and Risks**

While containment can prevent domino effects, it can also carry adverse consequences. As a historical example, consider Harry Truman’s Cold War containment strategy aimed at stopping Soviet expansion, under which the US provided military and economic aid to countries at risk of Soviet invasion to curb communist influence. Some scholars argue, however, that Truman’s containment strategy inadvertently inhibited diplomacy and reduced the possibility of dialogue with the Soviet Union. In short, containment is not always a risk-free approach.

<a id="b-9c57eb05f21471f3"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-6e9a8abea72d"></a>

<a id="v2-190-0"></a>

**Shortform perspective: Appeasement and Hindsight Bias**

The most famous historical example of appeasement is widely regarded as a failure: in 1938, British Prime Minister Neville Chamberlain signed the Munich Agreement, allowing Germany to annex the Sudetenland in the hope of appeasing Germany and avoiding further conflict. That decision famously proved futile, as conflict broke out anyway. Although the Munich Agreement is often cited as a cautionary tale against appeasement, this reading may reflect hindsight bias—the tendency to see an outcome as more predictable in hindsight than it actually was at the time.

*Sources and further reading:* [Precautionary-principle critique](https://dash.harvard.edu/handle/1/29998410) · [Nuclear-energy risks](https://www.greenamerica.org/fight-dirty-energy/amazon-build-cleaner-cloud/10-reasons-oppose-nuclear-energy) · [Fossil fuels and climate](https://www.hsph.harvard.edu/c-change/subtopics/fossil-fuels-health/#:~:text=Producing%20and%20burning%20fossil%20fuels,coal%2C%20oil%20and%20natural%20gas.) · [Expected utility](https://people.hec.edu/mongin/wp-content/uploads/sites/36/2018/08/O12.MonginExpectedHbk97.pdf) · [Precaution as a heuristic](https://academic.oup.com/lpr/article/5/1/19/990788) · [Pandemic panic buying](https://sustainability.utdallas.edu/news/20200406a/) · [Cultural time horizons](https://fosburit.com/culture/long-term-orientation-vs-short-term-orientation/) · [Kant’s universalizability test](https://plato.stanford.edu/entries/kant-moral/#ForUniLawNat) · [Choice-menu design](https://www.shortform.com/app/book/nudge/techniques-3-and-4) · [Defaults](https://www.shortform.com/app/book/nudge/technique-1) · [Bottle-deposit systems](https://allaboutberlin.com/guides/pfand-bottles) · [Tip prompts and pressure](https://www.today.com/money/guilt-tipping-are-square-mobile-payments-making-us-tip-everyone-t126151) · [Reinforcement and punishment](https://hbr.org/2017/09/what-motivates-employees-more-rewards-or-punishments#:~:text=This%20may%20seem%20odd%2C%20but,be%20more%20effective%20than%20punishments.) · [Rewards and children](https://www.psychologytoday.com/us/blog/family-affair/200809/rewards-are-better-punishment-here-s-why) · [Cold War containment](https://history.state.gov/milestones/1945-1952/kennan) · [Containment and diplomacy](https://www.jstor.org/stable/24914870) · [Munich Agreement](https://www.britannica.com/event/Munich-Agreement)

<!-- atlas:block-content b-d9fe917eb66113a5:end -->
<!-- atlas:block-content b-2a3ff95ed8efa94f:start -->
### Build an advantage, then keep adapting

<a id="b-2a3ff95ed8efa94f"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-c5320beb1c97"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-9626f82865b3"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-9470d1a64f0d"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-41dda2218dca"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-39ae0fc87fed"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-5b70501f881b"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-eb5af062db1a"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-2854822f5e68"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-128bb29a057d"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-1babe0b9b114"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-1f6ade00e2e5"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-1b83ba6199ba"></a>
<a id="b-ef8130004e333ae6"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-1a16df710673"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-b1461ea67302"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-c83ac92910ff"></a>

<a id="v2-197-0"></a>

**Sustainable Competitive Advantage: Secrets and Product/Market Fit**

A sustainable competitive advantage is an edge that competitors cannot quickly erase. The summary uses Tesla’s battery capabilities to illustrate the potential value of controlling an important input. Producing a component internally is not automatically cheaper or better; capital cost, scale, expertise, supplier alternatives and execution determine whether integration creates an advantage.

Weinberg and McCann distinguish two kinds of secrets. One is little-known information that creates an opportunity: the summary’s Bob Voulgaris case describes finding betting edges through basketball analysis before the relevant pricing caught up. The other is a known possibility others are reluctant to pursue: electric cars existed around 1900, and Tesla later challenged assumptions about their commercial viability. Early existence and Tesla’s later valuation do not prove that electric cars led all sales in 1900 or that one insight alone caused business success. The authors recommend curiosity and contact across fields because one field’s familiar knowledge may be another’s overlooked opportunity.

Secrets can reveal unmet needs and support product/market fit: customers want the product enough to sustain a business. Apple’s place among many smartphone producers illustrates differentiated value, not proof that the first entrant to achieve a supposedly “perfect” fit always wins. Distribution, cost, timing and competitors also shape outcomes; fit can change as the market changes.

Customer development tests assumptions by learning from prospective and current customers and iterating the product. Interviews, experiments and focus groups can reveal wants and obstacles. Focus groups are not necessarily random samples and do not establish demand merely because participants like an idea. Connect what people say with actual behavior and the market segment the business intends to serve.

<a id="b-9fd560881165b155"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-e236b027bf95"></a>

<a id="v2-198-0"></a>

**Shortform perspective: Finding Secrets: Thiel’s Framework and the Wright Brothers Example**

Weinberg and McCann credit Peter Thiel’s book *Zero to One* as the origin of the “find secrets” framework, and Thiel proposes additional steps for uncovering secrets: believe that secrets are still out there and resist the assumption that all relevant secrets have already been found; examine questions that mainstream thinkers ignore or dismiss as worthless, as these often point to secrets; and be willing to look foolish by questioning conventional assumptions, since rigid adherence to those assumptions blocks discovery.

Shortform uses the Wright brothers to illustrate questioning prevailing assumptions about practical powered flight. Wing warping altered lift across the wings to help control roll; it was not simply a trick for increasing total lift. Their December 1903 flight depended on a combination of control, propulsion, aerodynamic work and testing. The transferable mechanism is investigating neglected constraints, not proving that contrarian belief alone produces a breakthrough.

<a id="b-6bf11463a3d4db21"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-1cac9b20d1b3"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-7f1394bbb2ea"></a>

<a id="v2-199-0"></a>

**Shortform perspective: Product/Market Fit, the Product Life Cycle, and Customer Development**

Products that achieve product/market fit still go through the product life cycle, which often ends in market decline as they fall out of favor with consumers. For this reason, it is important to keep adapting even after reaching product/market fit, or success may be short-lived. This echoes Simon Sinek’s advice in *The Infinite Game*: treat business as an infinite game where the goal is simply to keep playing.

Additionally, while Weinberg and McCann discuss customer development in the context of specific products, the concept also applies to entire business models; start-ups in particular can use customer development to identify the strengths and weaknesses of potential business models, for example by running surveys to assess consumer needs and then building models to meet those needs.

<a id="b-b1d0485151dac89f"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-f1de2e404a96"></a>

<a id="v2-200-0"></a>

**Shortform perspective: Sustainable Versus Transient Competitive Advantage**

Although sustainable competitive advantage is widely hailed as the holy grail of business, some strategists argue instead for pursuing transient advantages: short-term edges that collectively yield long-term success. Proponents of this view contend that sustainable competitive advantage has become unrealistic, because globalization and increased digitization have made long-term edges harder to sustain. They therefore recommend seeking out and exploiting a sequence of short-term advantages.

Amazon illustrates this approach: it started as a platform for selling books, then added CDs, expanded into categories such as toys and electronics, and eventually launched a streaming service. None of these individual moves was sustainable on its own—other companies also sold CDs and general goods online, and competing streaming services emerged—but each provided a short-term advantage that Amazon exploited for a period. That said, these short-term edges may collectively have produced a long-term edge, because the sheer breadth of products available on Amazon is now itself a sustainable competitive advantage.

<a id="b-d4e75fb73ff6804d"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-d6cab9da4bd8"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-f142b143ecc8"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-1c94839d3b34"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-588f646b8201"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-2584b7188f52"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-869dc702900b"></a>

<a id="v2-203-0"></a>

**Adaptability, Natural Selection, and Overcoming Inertia**

The framework of natural selection, drawn from evolutionary theory, provides a mental model for business adaptability. In biology, natural selection dictates that genes contributing to genetic fitness—the ability to survive and reproduce—are most likely to be passed down; for example, longer-necked giraffes can reach food more easily, improving their survival and reproductive odds, so that trait is more likely to appear in future generations. Weinberg and McCann argue that an analogous selection process operates in societal contexts: ideas that adapt to society’s evolving preferences are more likely to thrive. Businesses that adapt to these shifting preferences are better positioned for success, while those unwilling to adapt are quickly left behind. The authors present a set of mental models designed to maximize adaptability and put a business on the path to success.

Inertia in mechanics concerns persistence of motion unless a net force changes it; the business use is an analogy for resistance to altering routines and commitments. Question assumptions and try feasible changes when the environment shifts. Netflix added streaming in 2007 while its DVD business continued; it did not abandon physical rentals at that moment. The example illustrates building a new capability alongside an existing operation, with the difficulties and timing of transition still important.

<a id="b-5d78e1fa1b0e7e6f"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-f90a230cc662"></a>

<a id="v2-204-0"></a>

**Shortform perspective: Adaptation as an Infinite Game**

Framing adaptation as an infinite game offers another perspective on sustained business success. Simon Sinek argues that by constantly adapting to a dynamic society, you treat business as an infinite game—like Tetris—where the objective is to continue playing indefinitely, and you will be more successful than if you view business as a finite game—like chess—that has a fixed endpoint. To operate with this infinite-game mindset, Sinek encourages three practices: proactively prepare to pivot so you can position yourself to thrive when societal preferences shift; study worthy competitors to learn additional ways to adapt and survive; and build trusting teams, because employees in a trusting environment are more likely to share ideas for improving the company.

<a id="b-c1c636bbb16bf1f1"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-3dae4d3f4be3"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-0f988ffb99ad"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-7e316738b2b9"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-459bb6099e21"></a>

<a id="v2-201-0"></a>

**Momentum, tipping points, and adoption curves**

Closely related to inertia is the concept of momentum. To succeed in business, Weinberg and McCann propose pursuing ideas and products whose momentum is increasing. A key way to identify such opportunities is to seek out ideas and products that are about to reach a tipping point—the moment when a system begins to shift rapidly. As an illustration, in the 1990s Baltimore reached a tipping point that sparked a syphilis epidemic, as a large number of people in high-density areas were exposed. In a business context, tipping points occur when products are adopted not only by early adopters—people willing to try new ideas without social pressure—but also by the early majority—people willing to try ideas popularized by early adopters. By staying attuned to trends among early adopters, you can anticipate which products are likely to gain momentum in the future.

<a id="b-ee5ae464e0b072ce"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-1f93cf6bc7d2"></a>

<a id="v2-202-0"></a>

**Shortform perspective: Tipping Points and Crossing the Chasm**

To maximize momentum for your own products, you need to actively reach tipping points by crossing the “chasm” between the early market and the mainstream market. In *Crossing the Chasm*, Geoffrey Moore outlines a four-step guideline for doing so: first, use your intuition to choose a niche market to target; second, develop a “simplified whole product” for that niche—roughly, your core product plus its necessary external components, such as a smartphone paired with its operating system and charging devices; third, position your product by crafting a position statement that situates it among competitors in the market; and fourth, secure a distribution channel to sell your product to potential customers.

As a concrete example, for a dating app you might intuitively target divorcees as your niche, build a beta website for a smaller member pool, position the app as a viable alternative to existing options by highlighting its advantages, and finally buy advertisements in venues most viewed by divorcees.

<a id="b-c32cb8d1d3a96138"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-d08d70939e3a"></a>

<a id="v2-210-0"></a>

**Shortform perspective: Inertia, Friction, and Organizational Stagnation**

In *The Great Mental Models Volume 2*, Parrish and Beaubien discuss a physics model closely related to inertia: friction, a force that hinders motion. In a business context, external factors can create friction and hinder adaptation. For example, computer companies might encounter friction via hardware manufacturing issues, and start-ups might encounter friction in the form of funding problems. When acting together, friction and inertia can lead to stagnation: inertia causes entities that are currently stationary to tend to stay stationary, while friction makes it difficult for external factors to drive a change of path.

*Sources and further reading:* [Tesla battery integration](https://cleantechnica.com/2020/07/16/teslas-5-biggest-competitive-advantages/) · [Voulgaris’s betting analysis](https://www.espn.com/blog/playbook/dollars/post/_/id/2935/meet-the-worlds-top-nba-gambler) · [Early electric-car history](https://www.businessinsider.com/electric-car-history-2017-5#the-electric-car-burst-onto-the-scene-in-the-late-1800s-and-early-1900s-1) · [Zero to One](https://short.zaynjarvis.com/books/zero-to-one/) · [Wright brothers’ experimentation](https://www.nytimes.com/2003/12/09/news/earliest-days-takeoff-how-the-wright-brothers-did-what-no-one-else-could.html) · [Product life cycle](https://hbr.org/1965/11/exploit-the-product-life-cycle) · [The Infinite Game](https://www.shortform.com/app/book/the-infinite-game/shortform-introduction) · [Customer development](https://hbr.org/2013/05/why-the-lean-start-up-changes-everything) · [Transient advantage](https://hbr.org/2013/06/transient-advantage) · [Competition and durability](https://store.hbr.org/product/the-end-of-competitive-advantage-how-to-keep-your-strategy-moving-as-fast-as-your-business/10605) · [Natural-selection analogy](https://www.nhm.ac.uk/discover/what-is-natural-selection.html) · [Netflix’s streaming transition](https://www.uschamber.com/co/good-company/growth-studio/successful-companies-that-reinvented-their-business) · [Infinite-game perspective](https://www.shortform.com/app/book/the-infinite-game/1-page-summary) · [Preparing to pivot](https://www.shortform.com/app/book/the-infinite-game/guideline-4) · [Learning from competitors](https://www.shortform.com/app/book/the-infinite-game/guideline-3) · [Trusting teams](https://www.shortform.com/app/book/the-infinite-game/guideline-2) · [Baltimore epidemic example](https://www.shortform.com/blog/tipping-point-examples/) · [Crossing the Chasm](https://short.zaynjarvis.com/books/crossing-the-chasm/)

<!-- atlas:block-content b-2a3ff95ed8efa94f:end -->
<!-- atlas:block-content b-7e1701109cf5e029:start -->
### Teams need culture, fit and useful practice

<a id="b-7e1701109cf5e029"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-550966ac452b"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-a323880672a6"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-ff3b14982860"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-dcde637098a7"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-d285c66c707b"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-ae5701ee0f9b"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-e777a90b87bd"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-00bece0601dd"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-db997ab24a25"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-f0a2c33831a5"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-f1e291dbb221"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-2e148a5d1237"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-31d0d02969ae"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-a58104d41d7d"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-cef5ecc2ac6e"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-206776e2be7e"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-89ca70af66d2"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-dc19621e3d18"></a>

<a id="v2-205-0"></a>

**Building 10x Teams: Culture, Individual Management, and Deliberate Practice**

Weinberg and McCann argue that adaptability alone is not enough for success—you must also tap into employees’ potential to build thriving teams. Their specific goal is to build *10x teams*: groups of competent employees positioned to thrive by leveraging their unique skills. The seemingly obvious strategy of recruiting only all-star employees will not work, they contend, because *Joy’s Law* holds that most of the smartest, most talented people in the world always work for someone else. Since you cannot count on every hire being a natural all-star, it is critical to instead develop and harness the potential of the employees you already have.

The first step toward building 10x teams is creating a strong *organizational culture*—the shared beliefs, practices, and values of a given organization. Weinberg and McCann emphasize that a strong culture is not one-size-fits-all, but several essential steps can help leaders shape a positive culture:

- Designate a concrete north star.
- Make your core values explicit.
- Reward employees for outstanding performances.
- Embody your desired norms and values.

By taking these steps, leaders can actively steer company culture in their intended direction.

A key mental model for garnering support within an organization is *winning hearts and minds*. Originating in military contexts, it means appealing directly to people’s emotions and reason to gain their backing. During Ukraine’s war against Russia in 2022, for example, President Volodymyr Zelenskyy employed this model by framing the war as a fight against tyranny and evil while simultaneously countering disinformation from Russian media. In business, winning hearts and minds fosters intrinsic motivation among employees: by clearly communicating values and showing the tangible needs the company meets, leaders can win employees over to the organization’s cause and help them persevere through challenges. By contrast, employees motivated only extrinsically will readily leave for a better offer, such as a higher-paying job.

Beyond team-building, maximizing employee potential requires managing each person in the manner best suited to them. A core principle is placing employees in roles tailored to their skills: introverts may be better suited to isolated positions such as strategic roles, while extroverts may perform better in communicative environments like sales. Similarly, understanding an organization’s specific needs clarifies which types of employees to recruit. Generalists—who prefer learning about a broad range of topics—can thrive in start-ups, where employees must handle a wide array of tasks. Specialists—who prefer deep, focused expertise in individual topics—can thrive in larger organizations that can afford to invest in dedicated experts.

Placing employees in the optimal role is not enough; they also need ongoing guidance and nurturing to reach their full potential. Specifically, managers should encourage *deliberate practice* among their teams. Developed in earlier research by Anders Ericsson and colleagues and later discussed in *Peak*, deliberate practice involves working at the edge of your ability while receiving consistent feedback on performance. For example, a novice chess player might analyze positions just beyond their current understanding, then review that analysis with an expert coach.

<a id="b-baab083ceb21d336"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-dd153b697aac"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-f67a82f24f28"></a>

<a id="v2-206-0"></a>

**Shortform perspective: Team Composition: Introverts, Extroverts, and Star Employees**

Studies indicate that in the workplace, extroverts are both higher-paid and promoted more quickly than introverts, which creates incentives for introverts to act more extroverted. However, introverts who behave like extroverts in the short term can experience fatigue and diminished performance over the long term, so it is important not to excessively pressure introverts to adopt more extroverted styles at work.

Even when star employees are available to hire, they can create team-level challenges. For example, hiring a star can lead to more interpersonal conflict—possibly driven by jealousy and resentment—which in turn hurts productivity. Similarly, giving too much positive feedback to a star employee can inflate their ego and harm overall team performance. For these reasons, star employees are not always worth the associated costs and complications.

<a id="b-9364767fd2b46cc1"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-2dcd07ae8e8d"></a>

<a id="v2-207-0"></a>

**Shortform perspective: Deliberate practice requires feedback, not necessarily competition**

Ericsson’s account, developed before and discussed in Peak, concerns structured work on specific weaknesses, tasks beyond current comfortable performance, informative feedback, correction and repeated effort. Reliable measures or expert criteria make improvement visible. Competition and incentives can motivate some learners, but are not a necessary condition for every instance of deliberate practice. A contest without targeted feedback is not enough. See [Ericsson’s retrospective account of the 1993 research](https://pmc.ncbi.nlm.nih.gov/articles/PMC6824411/).

<a id="b-829449567b5c20e7"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-c7860821e28f"></a>

<a id="v2-208-0"></a>

**Shortform perspective: Generalists and Specialists: From Individuals to Companies**

The generalist-versus-specialist distinction applies not only to individuals but also to entire companies, as Shane Parrish and Rhiannon Beaubien note in *The Great Mental Models Volume 2*. Generalist companies like Walmart offer wide ranges of products, while specialist companies like Untuckit—which sells casual men’s shirts—focus on narrower niches.

<a id="b-d8ad4c6a24a75669"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-8135994cd246"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-4db9adabc7d1"></a>

<a id="v2-209-0"></a>

**Shortform perspective: Organizational Culture: Competing Views and Remote Work Realities**

Whereas Weinberg and McCann treat organizational culture as just one component of a thriving business, Patrick Lencioni argues in *The Advantage* that culture is the single greatest advantage a company can achieve. He offers several complementary practices for strengthening culture: encourage healthy debate among the leadership team so the best idea wins on its merits; foster peer-to-peer accountability to reduce the burden on leaders of enforcing accountability alone; and set achievable short-term goals to ensure steady progress and prevent stagnation. Though these practices do not draw on mental models, they add useful perspective on the foundations of a healthy culture and pair well with Weinberg and McCann’s recommendations for nurturing effective, high-performing teams.

Weinberg and McCann’s guidance on winning hearts and minds is oriented toward in-person companies, but the Covid-19 pandemic shifted many organizations to remote work, where cohesive culture can be harder to build because there is less in-person interaction to reinforce norms and values. Several strategies can help mitigate this challenge: hold monthly all-hands meetings to enable clear communication and strengthen a sense of belonging; run fun team-building activities to reinforce personal connections between team members; create channels for casual conversation to foster a warmer environment; and recognize individual accomplishments to remind employees they are valued.

*Sources and further reading:* [Zelenskyy’s communication](https://news.yahoo.com/ukraines-information-war-winning-hearts-121457925.html?guccounter=1&guce_referrer=aHR0cHM6Ly93d3cuZ29vZ2xlLmNvbS8&guce_referrer_sig=AQAAAKnCHWA8kvb1Z2wXTRyytDXW_ghowX76IhM6AsUbudU3Eeb2K1nv8gAre-YrO6sJ2_7qukXAI9i-KJwulwSMAFEHBdUXSQyA4oL2Ol8K6vnUYwLAsxC2hY6B8zeXtF4MROhRKvgTH7I7sXMnnE9eg45RvQrE1VzNayBrdclUqcaB) · [Peak](https://www.shortform.com/app/book/peak/1-page-summary#deliberate-practice) · [Extraversion and careers](https://psycnet.apa.org/doiLanding?doi=10.1037%2Fapl0000415) · [Acting extraverted and fatigue](https://hbr.org/2022/10/stop-telling-introverts-to-act-like-extroverts) · [Star hires and conflict](https://hbr.org/2004/05/the-risky-business-of-hiring-stars) · [Praise and ego](https://hbr.org/2017/06/how-to-manage-your-star-employee) · [The Advantage](https://www.shortform.com/app/book/the-advantage) · [Remote-work transition](https://www.pewresearch.org/social-trends/2020/12/09/how-the-coronavirus-outbreak-has-and-hasnt-changed-the-way-americans-work/) · [Remote-team culture](https://www.forbes.com/sites/forbesbusinessdevelopmentcouncil/2021/02/02/12-actionable-ways-to-build-and-sustain-culture-in-a-remote-environment/?sh=634e1e187ba5)

<!-- atlas:block-content b-7e1701109cf5e029:end -->
<!-- atlas:block-content b-8d984e91e245855e:start -->
### Contain damage without inventing a single cause

<a id="b-8d984e91e245855e"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-0360a87c65bf"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-9e1f078e9c9a"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-bcbc607e15cf"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-800171ba6ad2"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-8011803c077c"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-1e4a27fba150"></a>
<a id="u-super-thinking-the-big-book-of-mental-models-5882ae126d3b"></a>

<a id="v2-211-0"></a>

**Minimizing Damage From Competitor Conflict: Containment and Appeasement**

Even companies with sustainable competitive advantages must remain wary of conflict with competitors. To that end, Weinberg and McCann offer mental models designed either to avoid conflict entirely or to reduce its fallout once it begins.

Containment limits further damage after a problem emerges. A food recall can stop additional exposure to a harmful product. Competitive threats need a different analysis: the summary contrasts Blockbuster’s refusal of a proposed $50 million Netflix acquisition with its 2010 bankruptcy, but that sequence does not establish that one rejected deal caused the collapse. Buying a rival is also not mechanically equivalent to recalling a product. Identify the threat, feasible responses and costs rather than infer a single missed move made failure inevitable.

If all else fails, Weinberg and McCann recommend *appeasement*, even if it requires making significant concessions to the other side. For example, if you are locked in a legal battle with competitors that threatens your company’s reputation, it may be worth making a settlement offer to appease them. You accept a short-term concession in order to prevent a possible long-term catastrophe.

*Sources and further reading:* [Netflix acquisition proposal](https://www.independent.co.uk/news/world/americas/blockbuster-ceo-netflix-meeting-laugh-b2009607.html) · [Blockbuster’s bankruptcy](https://www.businessinsider.com/the-rise-and-fall-of-blockbuster-video-streaming-2020-1#:~:text=But%2C%20in%202010%20Blockbuster%20filed,DVD%2Dby%2Dmail%20service.)

<!-- atlas:block-content b-8d984e91e245855e:end -->
<a id="theme-algorithms-and-bounded-optimization"></a>

## Algorithms under real constraints

<!-- atlas:block-content b-39315a6d53dc5606:start -->
### Stopping and exploration solve different problems

Before borrowing an algorithm, state the objective, available information, constraints and cost of a trial. Change those and the recommended action may change.

<a id="b-39315a6d53dc5606"></a>
<a id="u-algorithms-to-live-by-bcffda6a7b30"></a>
<a id="u-algorithms-to-live-by-f0e47be066b7"></a>
<a id="u-algorithms-to-live-by-5da70cbccd0e"></a>
<a id="u-algorithms-to-live-by-8b0109f56a84"></a>
<a id="u-algorithms-to-live-by-2aba577e123c"></a>
<a id="u-algorithms-to-live-by-731142103636"></a>
<a id="u-algorithms-to-live-by-775e8d6bcbaa"></a>

<a id="v2-212-0"></a>

**Core premise: computer algorithms as life strategies**

In *Algorithms to Live By*, Brian Christian and Tom Griffiths argue that computer science, often seen as a highly specialized field, contains a wealth of practical knowledge for improving everyday life. The core insight is that humans and computers face fundamentally similar problems: both must deploy limited resources—including memory, attention, and time—as optimally as possible. Because computers can process tasks with remarkable efficiency and quickly generate creative solutions to complex problems, the authors contend that people can achieve similar results by adopting the same algorithmic strategies computers use. The authors present eleven such “algorithms to live by.” Shortform groups them into four categories: decision-making algorithms, organizational algorithms, problem-solving algorithms, and a small set of miscellaneous algorithms that do not fit neatly into the other groups.

<a id="b-84dd472484ca5f09"></a>
<a id="u-algorithms-to-live-by-d0dc0ebef95c"></a>
<a id="u-algorithms-to-live-by-00cd47fdb3fd"></a>
<a id="u-algorithms-to-live-by-4928a8edb249"></a>
<a id="u-algorithms-to-live-by-a0c23396cc56"></a>

<a id="v2-213-0"></a>

**Optimal stopping: what the 37% rule optimizes**

The classic secretary problem assumes a known number N of candidates arriving in random order, observable relative ranks, irrevocable rejection, and the goal of maximizing the probability of selecting the single best candidate. For large N, an approximately optimal rule is to inspect the first N/e—about 37%—without selecting, then accept the first later candidate better than all previously seen. If none appears, the search can end without selecting the best; a practical final-candidate fallback does not restore that lost opportunity. The maximum success probability also approaches 37%.

Christian and Griffiths use job hunting to motivate the question: an early offer is the best so far but lacks some desired features—accept it or keep searching? Real searches may have unknown totals, nonrandom arrivals, negotiable deadlines and recall. In those cases, 37% is a model-derived benchmark, not a universal percentage of your job search or life.

<a id="b-1e9d07be62062fcb"></a>
<a id="u-algorithms-to-live-by-cb0c6ad538ec"></a>

<a id="v2-224-0"></a>

**Shortform perspective: Changing the goal changes the stopping rule**

Shortform cites Hannah Fry’s discussion of accepting a good candidate rather than insisting on the absolute best. The 37% success rate does not refute the classic result; it is the result under its demanding objective. If the best candidate arrives during the observation phase, the rule rejects it and no later candidate can beat the benchmark.

Fry’s illustrative figures in the source are: accepting a top-5% option, an exploration fraction of 22% and a 57% success rate; accepting a top-15% option, 19% exploration and 78% success. These figures belong to her specified example, not a general theorem for arbitrary candidate counts or search conditions. The transferable point is to state what “good enough” means, then use a stopping policy suited to that objective and the opportunities actually available.

<a id="b-7cc324c0aada28f0"></a>
<a id="u-algorithms-to-live-by-d8e11be55159"></a>
<a id="u-algorithms-to-live-by-f82f61430a43"></a>
<a id="u-algorithms-to-live-by-f37c05b6f9b6"></a>
<a id="u-algorithms-to-live-by-dcb43e521ccc"></a>

<a id="v2-214-0"></a>

**Multi-armed bandits: learn where uncertainty is worth testing**

In a multi-armed bandit problem, repeated choices among slot machines with initially unknown returns must balance earning rewards now with learning which machine pays better. An Upper Confidence Bound policy ranks an option by an estimated mean reward plus an uncertainty bonus. A poorly sampled option can receive more exploration because its mean remains uncertain; as evidence accumulates, that bonus typically falls. This is optimism about a plausible bound on the mean, not permission to choose the largest imaginable jackpot regardless of its probability.

The life analogy is to try promising, underexplored opportunities, learn from outcomes, and redirect effort when the evidence weakens. Its usefulness depends on repeatable feedback, a meaningful comparison of rewards, the time left to benefit from learning, and affordable experiments. Check downside and trial costs before exploring: a one-off decision with ruinous losses is not the bounded repeated-reward problem that justifies a standard UCB rule.

<a id="b-7eed715d41368413"></a>
<a id="u-algorithms-to-live-by-2b04fcaaddcf"></a>

<a id="v2-227-0"></a>

**Shortform perspective: Put a boundary on exploration losses**

Shortform brings in Taleb’s The Black Swan as a caution about tail exposure. A bank can earn millions on loans for forty years while still carrying a large, poorly represented default risk. A favorable record does not by itself measure the chance or severity of a rare systemic loss. That does not make all historical evidence meaningless; it means its limitations matter when the outcome distribution is asymmetric or poorly known.

Before pursuing upside, ask how much an experiment can lose, whether losses can cascade, and whether you can survive being wrong. This adds a risk-design question to the bandit analogy. It should not be described as a refutation of UCB or as evidence that Christian and Griffiths’s original book ignores all downside risk.

*Sources and further reading:* [Fry on optimal stopping](https://ideas.ted.com/when-should-you-settle-down/) · [Taleb on tail exposure](https://www.shortform.com/app/book/the-black-swan/chapter-7)

<!-- atlas:block-content b-39315a6d53dc5606:end -->
<!-- atlas:block-content b-08c5a9acb8e167f8:start -->
### Simplify a prediction without deleting constraints

Bayesian updating has its full explanation in the uncertainty section. The next model concerns generalization: what should remain in a model when additional flexibility fits noise?

<a id="b-08c5a9acb8e167f8"></a>
<a id="u-algorithms-to-live-by-1aa3fd615065"></a>
<a id="u-algorithms-to-live-by-7d5e256cedd0"></a>
<a id="u-algorithms-to-live-by-c9d8ebf336e5"></a>
<a id="u-algorithms-to-live-by-1056c79b3291"></a>

<a id="v2-216-0"></a>

**Overfitting: simplify without deleting what makes a choice viable**

A model overfits when its flexibility captures peculiarities or noise in the observed data at the expense of performance on new data. Too many adjustable variables relative to the available evidence can contribute, but variable count alone does not define the problem. Simpler models, regularization and testing on held-out cases can help.

Christian and Griffiths’s everyday analogy is to stop letting minor details overwhelm a few important factors. In choosing a job, enjoyment can deserve substantial weight. It cannot automatically replace salary needed to live, health, location or other binding constraints. Identify essentials first, then remove detail that adds complexity without improving the decision. “Use one or two factors and ignore everything else” is too blunt to serve as a universal rule.

<a id="b-7d5844440323891e"></a>
<a id="u-algorithms-to-live-by-3111446fe0c2"></a>

<a id="v2-226-0"></a>

**Shortform perspective: Overfitting, minimalism, and focusing on what actually matters**

The idea of overcoming overfitting by focusing on what matters and ignoring the rest extends beyond individual decisions to life as a whole. In *Minimalism*, Joshua Millburn and Ryan Nicodemus apply this same logic to lifestyle design. Modern humans, they argue, have a tendency to overfit their lives: they try to make themselves happier by adding more things—luxury cars, fancy homes, picturesque vacations—instead of focusing on the few factors that actually drive well-being. These possessions and experiences merely distract from the things that offer the most lasting value: personal health, loving relationships, and a sense of contribution to others. On the whole, removing things from your life that do not add value is a more sustainable path to happiness than constantly chasing bigger and better new pleasures.

*Sources and further reading:* [Minimalism](https://www.shortform.com/app/book/minimalism)

<!-- atlas:block-content b-08c5a9acb8e167f8:end -->
<!-- atlas:block-content b-784d2e75ce691966:start -->
### Order work and store what you need

<a id="b-784d2e75ce691966"></a>
<a id="u-algorithms-to-live-by-7723548da5ec"></a>
<a id="u-algorithms-to-live-by-1267f56a1cb6"></a>
<a id="u-algorithms-to-live-by-13cd9b8fb206"></a>
<a id="u-algorithms-to-live-by-a45c51cae853"></a>

<a id="v2-217-0"></a>

**Scheduling: choose the objective before choosing the order**

There is no universally best task order. For a fixed set of available, independent tasks on one processor, with known durations and no preemption, shortest processing time first minimizes average completion time. It can also make a long to-do list feel more manageable, which is the motivational use described by Christian and Griffiths.

If tasks have weights, ordering by descending weight divided by processing time minimizes the weighted sum of completion times under the corresponding assumptions. This ratio gives useful priority information; it does not mean that selecting the highest value per hour at every instant solves every scheduling problem. Deadlines, dependencies, changing arrivals, setup costs and the aim of completing a major project can change the appropriate rule. Estimate duration and value, state the objective, then choose the ordering that serves it.

<a id="b-8a84828616730c4e"></a>
<a id="u-algorithms-to-live-by-3312af305921"></a>

<a id="v2-229-0"></a>

**Shortform perspective: Scheduling trade-offs: shortest task first versus hardest task first**

Shortform places Brian Tracy’s Eat That Frog! beside shortest-processing-time scheduling. Tracy is addressing procrastination on important work: an endless supply of easy tasks can displace difficult tasks with greater value. His advice is to identify important work, break it into manageable steps and protect time for it. This differs from minimizing mean completion time for a fixed task set; it does not refute that result. Difficulty and value are not identical, so choose priorities using the actual goal rather than assuming the hardest task is always most valuable.

<a id="b-fc7cc1b968588d1f"></a>
<a id="u-algorithms-to-live-by-d7c64a62cf09"></a>
<a id="u-algorithms-to-live-by-600622e49d30"></a>
<a id="u-algorithms-to-live-by-f66d82fc40fd"></a>

<a id="v2-218-0"></a>

**Caching: organizing belongings by frequency of use**

The first organizational algorithm focuses on efficient access to belongings: to efficiently access any collection, segment it based on frequency of use. The principle is drawn from how computers use caches—they group the data that needs to be accessed most often and search those high-priority stores first. Christian and Griffiths recommend applying the same caching logic to physical possessions by keeping small piles of the most frequently used clothes, books, and files within arm's reach of where they are typically needed.

<a id="b-fe5cadcfcd8cb28f"></a>
<a id="u-algorithms-to-live-by-3a578f703751"></a>

<a id="v2-231-0"></a>

**Shortform perspective: Frequency-based caching versus Marie Kondo’s tidying philosophy**

A frequency-based caching approach to organizing belongings is directly challenged by Marie Kondo in *The Life-Changing Magic of Tidying Up*. Kondo argues that sorting belongings by frequency of use is a common organizational mistake. While storing everything in easily reachable caches may save a few seconds, she contends it carries a greater cost: the clutter of countless piles scattered throughout the home. In her view, this kind of “organization” is really disorganization in disguise. More often than not, people simply drop their belongings wherever they happen to be, then build their routines around the locations of these ad-hoc caches. Furthermore, the system provides no easy way to memorize where everything is, so items stored in unusual places become hard to find.

<a id="b-b4af42815019e592"></a>
<a id="u-algorithms-to-live-by-12df39eccd6e"></a>
<a id="u-algorithms-to-live-by-bae7a35c68c0"></a>
<a id="u-algorithms-to-live-by-93f7f37784df"></a>
<a id="u-algorithms-to-live-by-a4acbffc8472"></a>

<a id="v2-219-0"></a>

**Bucket sort: partition when categories are cheap and useful**

Bucket sort first assigns items to ordered categories, then sorts within each category. Its efficiency depends on the cost of classification and how items are distributed among buckets; it is not always the best sorting method. For twenty years of meeting VHS tapes, year labels provide a useful first partition, followed by ordering tapes within each year. If almost all tapes belong to one year or dates are hard to identify, the expected advantage changes.

<a id="b-91df83f5c447a24f"></a>
<a id="u-algorithms-to-live-by-e6297464ac17"></a>

<a id="v2-230-0"></a>

**Shortform perspective: Divide-and-conquer sorting: quicksort and the limits of Bucket Sort**

Shortform supplements the summary with quicksort: choose a pivot, partition items according to whether they precede or follow it, then recursively sort the partitions. Absence from a one-page summary does not show the original book omitted it. Bucket sort works best when useful categories can be assigned cheaply and distribute work well; heavily uneven buckets can reduce its advantage. Quicksort has its own pivot and worst-case issues. Compare the actual distribution and partitioning cost, rather than treating every division into smaller groups as automatically efficient.

*Sources and further reading:* [Eat That Frog!](https://www.shortform.com/app/book/eat-that-frog) · [Quicksort](https://betterexplained.com/articles/sorting-algorithms/) · [Bucket-sort limitations](https://infohubblog.com/bucket-sort-advantages-and-disadvantages.html)

<!-- atlas:block-content b-784d2e75ce691966:end -->
<!-- atlas:block-content b-9bb22a84ccdc8a82:start -->
### Approximate and explore with a feasibility check

<a id="b-9bb22a84ccdc8a82"></a>
<a id="u-algorithms-to-live-by-e1a5f5bbc1c1"></a>
<a id="u-algorithms-to-live-by-3b41049aaac1"></a>
<a id="u-algorithms-to-live-by-58d64e057055"></a>
<a id="u-algorithms-to-live-by-17111f64aa9f"></a>

<a id="v2-220-0"></a>

**Hard to solve is not the same as impossible to compute**

Some problems have exact solutions but demand impractical time or memory at the scale we care about. Christian and Griffiths recommend trading some precision for a workable answer when the decision allows it. This differs from an undecidable problem, for which no general algorithm solves every instance.

Constraint relaxation removes or weakens some restrictions to solve an easier problem. The relaxed solution can provide a bound or suggest a useful approach, but it may violate the original requirements. Restore those requirements and check feasibility before acting. Sometimes a near-optimal answer is adequate; in other settings a small error changes safety or acceptability, so the allowable approximation must be chosen deliberately.

<a id="b-98dc1feccf092856"></a>
<a id="u-algorithms-to-live-by-62b70c379b20"></a>

<a id="v2-234-0"></a>

**Shortform perspective: Quantum computing does not remove all hard limits**

Shortform raises quantum computing as a possible source of future improvements. Qubits have quantum states described by amplitudes, and algorithms exploit interference and, in some cases, entanglement; they are not merely ordinary bits whose value is unknown. Quantum algorithms offer advantages for particular problems, not a general shortcut for every difficult optimization task.

Building useful computations at scale also requires controlling noise and errors. The fact that quantum hardware exists does not establish that it is universally efficient, that error correction is solved, or that approximation will soon be unnecessary. Treat this as a bounded technological possibility, not a reason to disregard resource limits when making decisions now.

<a id="b-33c1432b3c715e7b"></a>
<a id="u-algorithms-to-live-by-96137f3c6be3"></a>
<a id="u-algorithms-to-live-by-ba6c78c4220e"></a>
<a id="u-algorithms-to-live-by-45daed7ab74d"></a>
<a id="u-algorithms-to-live-by-9c385e0005d5"></a>

<a id="v2-221-0"></a>

**Escape a local optimum with bounded exploration**

Hill climbing repeatedly tries a nearby change and keeps improvements. It can stop at a local optimum: no allowed small move improves the objective, although a better region exists elsewhere. Random restarts or occasional worse moves can help explore another region. Such randomness is a deliberate search device, not a demand to behave irrationally.

The everyday analogy is to try unfamiliar hobbies or investigate a new location when small changes no longer help. Moving to a randomly chosen town has much larger consequences than sampling a hobby. Generate candidates, test them cheaply where possible, and check housing, work and other constraints before an irreversible commitment.

<a id="b-da259fd36cf57ec6"></a>
<a id="u-algorithms-to-live-by-aca98479be82"></a>

<a id="v2-228-0"></a>

**Shortform perspective: Randomness as a tool for discovery and ego dissolution**

Random decision-making can expose you to valuable experiences you would never have intentionally chosen. Derek Sivers takes this idea to the extreme in *How to Live*, advocating for a fulfilling life built entirely around randomness. Because you did not choose your career or your style, for example, you no longer base your identity or self-worth on them. Sivers argues further that random decision-making lets you live entirely without ego: you never need to worry about whether you are making the responsible choice or doing everything possible to secure a good future. Instead, you are free to live wholly in the present, enjoying life as it is rather than as it could be.

*Sources and further reading:* [Quantum-computing prospects](https://www.wired.co.uk/article/quantum-computing-explained) · [How to Live](https://sive.rs/h)

<!-- atlas:block-content b-9bb22a84ccdc8a82:end -->
<!-- atlas:block-content b-8ab760f957a44971:start -->
### Design interactions, not just individual moves

<a id="b-8ab760f957a44971"></a>
<a id="u-algorithms-to-live-by-01632ac9d1aa"></a>
<a id="u-algorithms-to-live-by-ca921ba26a5f"></a>
<a id="u-algorithms-to-live-by-1232983dbc2b"></a>
<a id="u-algorithms-to-live-by-29f318bf3e4a"></a>

<a id="v2-222-0"></a>

**Mechanism design: equilibrium is not collective welfare**

At a Nash equilibrium, given the other players’ strategies, no player can improve their own payoff by changing strategy alone. It need not maximize total welfare, benefit everyone, be unique, or be reached in practice. Mechanism design asks how rules and incentives can produce desired behavior under specified information and constraints.

In the fishing example, individually attractive catches can deplete a shared fish stock and harm future earnings. Catch limits, rights or penalties aim to make sustainable behavior individually worthwhile. Their success depends on monitoring, enforcement, incentives and biological conditions; announcing a penalty does not by itself establish a sustainable equilibrium.

<a id="b-52de4d2a133a5baf"></a>
<a id="u-algorithms-to-live-by-8eac30b84f4a"></a>

<a id="v2-233-0"></a>

**Shortform perspective: Nash equilibria: imprecise calculation, useful framework**

Shortform emphasizes limits on calculating and applying equilibria. Difficulty can arise from computational demands, uncertain payoffs, incomplete information, multiple equilibria and behavior that departs from the model; it is too strong to say equilibria generally cannot be computed by any algorithm. Even without an exact solution, ask whether the rule gives people incentives to comply or profitable ways around it. That question helps policy design without guaranteeing everyone benefits or predicting the exact outcome.

<a id="b-b6c8bf44d48e6cce"></a>
<a id="u-algorithms-to-live-by-1c85c21336ba"></a>
<a id="u-algorithms-to-live-by-5fc6c3e280cf"></a>
<a id="u-algorithms-to-live-by-a87c43e84462"></a>
<a id="u-algorithms-to-live-by-6daa37c8e1de"></a>

<a id="v2-223-0"></a>

**Acknowledgment: receipt is part of communication**

In networking protocols, acknowledgments report receipt or protocol progress; they do not by themselves prove a permanently stable connection or understanding of content. Christian and Griffiths use this as an analogy for listening. Brief backchannels—such as an appropriate acknowledgment or clarification—help a speaker know whether a message is arriving without surrendering the floor. Silence alone may leave that unclear, though the right amount of feedback depends on the conversation.

<a id="b-7ca82771e11ae011"></a>
<a id="u-algorithms-to-live-by-32e562be0462"></a>

<a id="v2-232-0"></a>

**Shortform perspective: Active listening: why silent attention is not enough**

Shortform contrasts active feedback with a simplified reading of Carnegie’s emphasis on attentive listening. These need not conflict: checking understanding differs from interrupting, taking over or rushing to offer solutions. Briefly restate the situation and, where appropriate, ask whether you have understood the speaker’s feelings before giving advice. Such responses can clarify both people’s thinking, but should leave room for correction; no listening style guarantees appreciation.

<a id="b-a4531e9eb60f9ba9"></a>
<a id="u-algorithms-to-live-by-e7304ed232fd"></a>

<a id="v2-235-0"></a>

**Shortform perspective: The brain-as-computer metaphor: debate over limits and utility**

The brain-as-computer metaphor, which underpins Christian and Griffiths’s argument, is far from new: researchers have drawn the comparison for decades, and today it sits at the center of a fierce debate in neuroscience. Some experts argue that the metaphor is limiting more than it helps at the cutting edge of the field. They point out numerous ways in which the brain does not behave like a computer, and warn that such inaccuracies will lead researchers astray with misguided assumptions. Other experts counter that the metaphor has not yet outlived its usefulness. In their view, it does not matter whether the brain operates exactly like a computer; what matters is that it accomplishes many of the same functions—intaking, processing, and exporting information. Because the brain computes, they argue, it is a computer.

*Sources and further reading:* [Limits of equilibrium models](https://www.newyorker.com/news/john-cassidy/the-triumph-and-failure-of-john-nashs-game-theory) · [Attentive listening](https://short.zaynjarvis.com/books/how-to-win-friends-and-influence-people/) · [Active listening](https://hbr.org/2016/07/what-great-listeners-actually-do) · [Brain–computer metaphor debate](https://www.theguardian.com/science/2020/feb/27/why-your-brain-is-not-a-computer-neuroscience-neural-networks-consciousness) · [Computational view of the brain](https://www.nytimes.com/2015/06/28/opinion/sunday/face-it-your-brain-is-a-computer.html)

<!-- atlas:block-content b-8ab760f957a44971:end -->
<a id="theme-science-incentives-and-reliability"></a>

## How to judge scientific claims

<!-- atlas:block-content b-3cde76f739a25f37:start -->
### Follow the rewards before blaming every researcher

Research reliability depends on incentives, methods and checking. Fabrication, careless error and unconscious bias require different diagnoses even when all can make a paper misleading.

<a id="b-3cde76f739a25f37"></a>
<a id="u-science-fictions-bc7a361a677d"></a>
<a id="u-science-fictions-3b77d16a4ada"></a>
<a id="u-science-fictions-55add9c3cf4d"></a>
<a id="u-science-fictions-23b9ba31ee4c"></a>
<a id="u-science-fictions-dd5b2032bd11"></a>
<a id="u-science-fictions-0bb869134e92"></a>
<a id="u-science-fictions-b6abe9218fde"></a>
<a id="u-science-fictions-b0e9b0e1eaab"></a>
<a id="u-science-fictions-56982d47fe81"></a>
<a id="u-science-fictions-8e1055783957"></a>
<a id="u-science-fictions-7a9980afa4af"></a>
<a id="u-science-fictions-ac9944240a03"></a>
<a id="u-science-fictions-adb981864530"></a>
<a id="u-science-fictions-b2045e52f83a"></a>
<a id="u-science-fictions-bf2fdcc1360a"></a>
<a id="u-science-fictions-817d43f8cccd"></a>
<a id="u-science-fictions-9a3772da2970"></a>
<a id="u-science-fictions-1d7e3d593140"></a>
<a id="u-science-fictions-0135dda2b1e5"></a>

<a id="v2-237-0"></a>

**Why Publication Pressure Selects for Poor Science**

In *Science Fictions*, Stuart Ritchie treats unreliable research primarily as a system problem rather than a collection of isolated bad acts. Academic hiring and funding make scientists feel that they must publish to survive, while publication quantity becomes a proxy for success. The decades-long growth in paper volume therefore cannot be read as a simple measure of progress: under these rewards, speed and output can displace rigor and produce work that contains more errors or adds little useful knowledge.

Goodhart’s principle explains the mechanism: once a metric becomes a target, it stops being a useful metric. When publication counts, journal prestige, and citation scores determine careers and resources, researchers and publishers have reason to optimize the numbers rather than the quality of the knowledge. The result can range from rushed, attention-seeking work to outright scientific misconduct.

Smaldino and McElreath’s model illustrates how selection could reward laboratories whose methods yield more publishable findings, allowing poor practices to spread even without a deliberate plan by every researcher. It makes the incentive mechanism explicit; a simulation alone does not establish the extent or causal history of poor science in real institutions. Ritchie uses it to explain why changing individual intentions may be insufficient when the reward structure remains.

<a id="b-0a3c005b6112ced9"></a>
<a id="u-science-fictions-8e75484163f2"></a>
<a id="u-science-fictions-372d47f49e4a"></a>
<a id="u-science-fictions-7571bec8bee9"></a>
<a id="u-science-fictions-d364e39f2345"></a>
<a id="u-science-fictions-1efcf741a286"></a>
<a id="u-science-fictions-93e1532595f1"></a>
<a id="u-science-fictions-fee311fe9765"></a>
<a id="u-science-fictions-5abb4b1d4ffd"></a>
<a id="u-science-fictions-ce34b3aeac72"></a>
<a id="u-science-fictions-4847f3821855"></a>
<a id="u-science-fictions-db3066141f8f"></a>
<a id="u-science-fictions-d31f89b651ad"></a>
<a id="u-science-fictions-6156f240427f"></a>
<a id="u-science-fictions-7a3fbe6e8484"></a>
<a id="u-science-fictions-80757aec9566"></a>
<a id="u-science-fictions-f24441226edf"></a>
<a id="u-science-fictions-e5cb3ea09d13"></a>
<a id="u-science-fictions-0a7178a02c15"></a>
<a id="u-science-fictions-2d7d59484538"></a>
<a id="u-science-fictions-44ebd7dd4a63"></a>
<a id="u-science-fictions-2e50c5f9b778"></a>
<a id="u-science-fictions-d2a3cc6c3734"></a>
<a id="u-science-fictions-4147e5069499"></a>
<a id="u-science-fictions-60703b45c4b3"></a>
<a id="u-science-fictions-4d4240be0258"></a>
<a id="u-science-fictions-e707e7d8e225"></a>
<a id="u-science-fictions-ed720aaf9d53"></a>
<a id="u-science-fictions-80ab333677ff"></a>
<a id="u-science-fictions-809486ea07b3"></a>
<a id="u-science-fictions-bc83c8b20fd6"></a>
<a id="u-science-fictions-7761a7943543"></a>
<a id="u-science-fictions-477d85550ad7"></a>
<a id="u-science-fictions-d01193038398"></a>
<a id="u-science-fictions-5f450c89b5df"></a>

<a id="v2-238-0"></a>

**From Metrics to Gaming: Citation Markets, Publisher Rents, and Fake Review**

Once output and prestige metrics become career targets, gaming follows. Journal impact factor—the average number of citations received by a journal’s articles—is arbitrary enough, and manipulable enough, that editors can end up protecting the metric rather than the reliability of what they publish. The h-index similarly began as an attempt to represent a researcher’s contribution but became another target. Coercive citation pressures authors to add particular references; citation cartels turn the same tactic into coordinated reciprocal citation among authors or journals.

Researchers can also game the count itself. “Salami slicing” divides one study into multiple papers, inflating productivity, obscuring the work’s coherence, and unfairly altering employment comparisons. Direct cash bonuses for publication, especially in prestigious journals, sharpen the financial incentive and weaken Merton’s principle of impartiality. Self-recycling reuses previously published text without proper attribution, creating an appearance of originality and productivity. Predatory journals exploit the same demand by selling rapid review and guaranteed acceptance through outlets with thin editorial standards and questionable peer review.

The publishing market adds another layer. Ritchie describes exorbitant subscription charges as a market failure and a form of rent-seeking: commercial publishers can extract profits far beyond the value of their editorial services because prestige is concentrated and competition is weak. His comparison of nonprofit National Academies journals with Elsevier finds dramatically different prices for nearly identical services, even though the system relies heavily on scientists’ unpaid reviewing labor.

Fake peer review shows how far the pressure can reach. Some authors nominate acquaintances as supposedly independent reviewers; others create bogus email accounts in real researchers’ names and review their own papers. Ritchie highlights Hyung-In Moon’s use of fake accounts and notes that even a prominent journal such as *Tumor Biology* struggled with this fraud. The practice subverts the process that is supposed to protect research quality.

<a id="b-5e5a19f5e42f3bb2"></a>
<a id="u-science-fictions-31463bbf62d1"></a>
<a id="u-science-fictions-5037ba6e7c68"></a>
<a id="u-science-fictions-a881d4ea8402"></a>
<a id="u-science-fictions-2e5506e022ce"></a>
<a id="u-science-fictions-134dc070796d"></a>
<a id="u-science-fictions-dc24973d4c61"></a>
<a id="u-science-fictions-73e1700018b7"></a>
<a id="u-science-fictions-7603d5da4851"></a>
<a id="u-science-fictions-d3930460573a"></a>
<a id="u-science-fictions-6f33abbcc15c"></a>

<a id="v2-239-0"></a>

**Shortform perspective: Where the Incentive Account Needs Qualification**

Shortform’s supplementary material connects these incentives to the reproducibility crisis: career and funding rewards can tilt toward prolific, attention-grabbing work even though a journal-level citation average says little about the quality or significance of an individual paper. Competition and gaps in research-ethics training can normalize corner-cutting, which helps explain how output-maximizing practices spread.

That mechanism is a risk, not a claim that all research is compromised. Collaboration can distribute pressure while preserving rigor, and many studies still follow strict standards. Government and industry laboratories may place less weight on prestigious-journal publication and more on practical or long-term work. Less flashy fields can remain valuable because of their foundational contributions. Direct publication bonuses also vary by country, discipline, and institution; where they exist, they may ease researchers’ financial pressure rather than necessarily corrupting impartiality.

<a id="b-3e4d59bcd430e2f5"></a>
<a id="u-science-fictions-9e4164bd8a52"></a>
<a id="u-science-fictions-1ee2f8181a05"></a>

<a id="v2-240-0"></a>

**Financial and Career Incentives That Distort Science**

Stuart Ritchie argues that financial interests can conflict with researchers' goals in ways that go beyond direct institutional funding. Scientists who have built successful careers around a specific theory—for example through book deals or paid lectures—may be strongly motivated to publish only studies that affirm that theory. To address this broader class of conflict, Ritchie calls for greater transparency, suggesting that current financial-disclosure requirements should be expanded to cover situations where a researcher's professional standing depends on particular results being validated.

<a id="b-4d921e50498c2a66"></a>
<a id="u-science-fictions-d23a9b23b458"></a>

<a id="v2-241-0"></a>

**Shortform perspective: Goodhart's Principle and Perverse Incentives in Science and Beyond**

Goodhart's principle, formulated by economist Charles Goodhart in 1975 in the context of economic policy, holds that targets become ineffective when they are used as control mechanisms. The principle extends well beyond economics: it applies to fields such as education, healthcare, and business, where performance metrics can incentivize gaming the system rather than genuine improvement.

<a id="b-f93ec238e1c71f4b"></a>
<a id="u-science-fictions-e6c72a570710"></a>

<a id="v2-242-0"></a>

**Shortform perspective: The Matthew Effect: funding concentration and structural inequity in science**

The Matthew Effect—the concentration of resources and recognition among already established scientists—shapes careers and the direction of research through several mechanisms.

Established scientists typically have extensive professional networks that generate further collaborative opportunities and entrench their status, while newer researchers struggle to build comparable connections. Senior investigators also tend to have better access to institutional resources such as lab space and administrative support, which are critical for conducting research effectively.

Early-career researchers often bring strengths of their own: strong command of the latest technologies and methodologies, and training across multiple disciplines that allows them to integrate concepts across fields and generate novel solutions. Even so, over time, funding concentration can homogenize research topics and methods, slowing progress and reducing a field's adaptability to new challenges. The Matthew Effect can also create a self-reinforcing career barrier: without recognition and funding, younger scientists find it hard to build a reputation, publish their work, and secure academic positions.

Diverse teams, by contrast, are more likely to pose a broader range of research questions and address issues that a more uniform group might miss. The peer review system itself can compound the problem by favoring studies that align with established theories, making it harder for unconventional ideas to gain a foothold.

<a id="b-333ef11bbbb38bd1"></a>
<a id="u-science-fictions-d315d54874f5"></a>

<a id="v2-292-0"></a>

**Shortform perspective: Impact Factor and Alternative Metrics**

Critiques of the impact factor as an arbitrary prestige metric that encourages unethical behavior should be tempered by two counterpoints. First, alternative metrics such as article-level metrics or altmetrics have their own limitations and can also be manipulated, meaning the problem is not unique to the impact factor. Second, the correlation between impact factors and citation rates does not necessarily imply causation: editors and publishers may focus on improving journal visibility and accessibility, which could naturally raise citation rates without compromising research quality.

<!-- atlas:block-content b-3cde76f739a25f37:end -->
<!-- atlas:block-content b-d306b9c2a0a59ca6:start -->
### How analytical freedom can select a result

<a id="b-d306b9c2a0a59ca6"></a>
<a id="u-science-fictions-7177b474aaaa"></a>
<a id="u-science-fictions-4f07e7d2c01c"></a>
<a id="u-science-fictions-7e0b19c4aa1f"></a>
<a id="u-science-fictions-14b984b3e051"></a>
<a id="u-science-fictions-3a7d3096023e"></a>
<a id="u-science-fictions-e320f5555d5d"></a>
<a id="u-science-fictions-ea0dc8f79994"></a>
<a id="u-science-fictions-58bf78e6a5f2"></a>
<a id="u-science-fictions-8ebf84e2115e"></a>
<a id="u-science-fictions-aeeacdb2d804"></a>

<a id="v2-243-0"></a>

**P-Hacking, Analytic Flexibility, and the Garden of Forking Paths**

Ritchie examines p-hacking as a pervasive form of data manipulation in which researchers adjust their analyses to push a result below the conventional p < 0.05 significance threshold. Tactics include dropping data points, recalculating results within specific subgroups, switching between statistical tests, or continuing data collection until a statistically significant effect appears. He illustrates the logic with the Texas sharpshooter analogy: a shooter fires randomly at a barn, then paints a bullseye around the tightest cluster of holes, presenting chance clustering as deliberate accuracy.

Running numerous unplanned, post-hoc tests is a central p-hacking method; when a desired result fails to appear, researchers cycle through analysis strategies in the hope of crossing the 0.05 line. Each additional test raises the probability of a false positive—mistakenly concluding an effect exists when it does not—turning analysis into a statistical fishing expedition in which more tests make meaningless patterns increasingly likely. A related variant is HARKing, or Hypothesizing After the Results are Known: researchers explore data without a predefined hypothesis, then construct a hypothesis around whichever results turn out significant, again presenting incidental statistical fluctuations as genuine, predicted effects and hiding the exploratory nature of the work.

Ritchie frames the broader structural problem as the "garden of forking paths": when many analytic choices are available, researchers tend to select the path that yields the most favorable outcome. This frequently leads to overfitting, in which a model fits the existing data perfectly but fails to generalize to new data or to reality. As an analogy, he contrasts a gently curved line approximating annual rainfall data—a useful predictor for future years—with a highly complex line that passes through every single data point, which merely captures the random noise of that particular dataset.

A key driver of these biases, Ritchie argues, is scientists' desire to uncover impressive-looking outcomes that support their own theories. Quoting Gould, he notes that the scientific community rewards "clean and unambiguous discoveries," creating pressure on researchers to simplify and overstate their findings. This emphasis on compelling narratives often comes at the cost of scientific rigor and obscures the inherent complexity of scientific investigation.

<a id="b-390b72705b44c5ff"></a>
<a id="u-science-fictions-3efaa0b00d71"></a>
<a id="u-science-fictions-b53c9ae6da67"></a>
<a id="u-science-fictions-80b2e963ba92"></a>
<a id="u-science-fictions-4147ce361ee8"></a>
<a id="u-science-fictions-a08cb1325d6c"></a>
<a id="u-science-fictions-252878897144"></a>
<a id="u-science-fictions-edd816617e85"></a>
<a id="u-science-fictions-efe7279524c4"></a>

<a id="v2-244-0"></a>

**Analytic Biases and Outcome-Switching in Research**

Beyond p-hacking, Ritchie identifies a broader set of analytic biases that skew results because the many choices made during data interpretation can nudge findings toward a researcher's preferred conclusions. A key example is outcome-switching in medical research: investigators test multiple outcomes upfront, but when their primary, preferred outcome fails to reach statistical significance, they reframe the study around a different outcome that did produce a significant result. He argues this practice obscures potentially valuable insights from the original non-significant outcome and inflates the false-positive rate, since the analysis is effectively run multiple times without statistical correction.

Ritchie also describes a "meaning well" or beneficial-intent bias: researchers may unconsciously skew their work because they genuinely want their studies to produce beneficial results and make a difference in the world. This desire can lead them to unconsciously favor findings that align with their hopes and expectations.

<a id="v2-239-1"></a>

**Shortform perspective: Questionable Research Practices That Favor a Result**

Other incentive-driven practices make findings look stronger or cleaner than they are. *P-hacking* selectively reports data that support a hypothesis while ignoring contradictory data. Peer review may miss the selection when reviewers cannot see the full dataset. Running many tests also increases the chance of finding at least one conventionally significant result by chance—the multiple-comparisons problem—and treating p < 0.05 as a magic threshold compounds the error. Data snooping searches for significant relationships without a prior hypothesis, raising the risk of spurious correlations.

HARKing blurs exploratory and confirmatory research by presenting a hypothesis devised after seeing the data as though it had been specified before collection. The resulting theory may rest on a chance pattern rather than a robust test.

Outcome-switching—changing a study’s focus after an insignificant result—is more conditional. It can be legitimate when a secondary analysis was pre-planned and documented, when an iterative project transparently refines its questions, or when all outcomes, including non-significant ones, are reported. Null findings can guide better designs and more precise hypotheses. Reanalysis is also defensible when a new hypothesis follows emerging evidence or theory rather than a data-driven fishing expedition.

<a id="b-2ddbb49e761532f0"></a>
<a id="u-science-fictions-0694f88a4400"></a>
<a id="u-science-fictions-f2bc9d495dc8"></a>
<a id="u-science-fictions-c2bf20dd66a9"></a>
<a id="u-science-fictions-9f9763651abe"></a>
<a id="u-science-fictions-179f362c4771"></a>
<a id="u-science-fictions-91d3991b2d5e"></a>
<a id="u-science-fictions-8434cffb2fee"></a>
<a id="u-science-fictions-bdd00067a9cd"></a>
<a id="u-science-fictions-92045af9fa15"></a>

<a id="v2-245-0"></a>

**Statistical Reform: Rethinking P-Values and Significance**

Ritchie argues that the heavy emphasis on clearing the arbitrary 0.05 p-value threshold has fueled p-hacking and a distorted view of what counts as good research. He advocates shifting the focus toward practical significance—assessing the real-world implications of an effect—and urges scientists to prioritize transparent reporting of uncertainty, error margins, and effect sizes rather than fixating solely on statistical significance.

At the same time, he cautions against completely abandoning the concept of statistical importance: although p-values have clear limitations, removing this objective benchmark entirely could introduce even more subjectivity into analysis and potentially worsen existing bias problems.

<a id="b-9dd8a7c2a9ade453"></a>
<a id="u-science-fictions-cee6edfc9390"></a>
<a id="u-science-fictions-9a8e8c4d93c9"></a>
<a id="u-science-fictions-ee25d711603d"></a>
<a id="u-science-fictions-32c4ceb61e42"></a>
<a id="u-science-fictions-cada002021e4"></a>
<a id="u-science-fictions-cfcff4e112f0"></a>
<a id="u-science-fictions-6092b4bbc959"></a>

<a id="v2-246-0"></a>

**Statistical and Methodological Reforms**

Ritchie discusses a proposed statistical reform: tightening the standard threshold for statistical significance from *p* < 0.05 to *p* < 0.005. The goal is to force researchers to rely on more robust evidence to support their hypotheses, which would reduce the rate of false-positive findings. He notes an important trade-off, however: achieving the same statistical power under the stricter threshold would require larger sample sizes, which would raise research costs and add logistical complexity.

Bayesian analysis combines an explicit prior with a likelihood for the observations and normalizes to obtain a posterior. Its probability statements are conditional on that model and prior; they are not assumption-free probabilities of truth. This permits incorporation of relevant prior knowledge, while making sensitivity to those choices another part of the analysis. The numerical updating mechanism is explained in the probability section.

Independent or blinded analysis can limit opportunities to favor an expected result. A statistician may work with masked treatment labels or without knowing the preferred hypothesis, while still needing enough design, sampling and measurement information to analyze correctly. Ritchie notes costs and possible disagreements over interpretation. Plan who sees what and when; simply withholding all study conditions can make the analysis worse.

<a id="b-c78366b3f386f19d"></a>
<a id="u-science-fictions-c594f28fadcb"></a>
<a id="u-science-fictions-6af9951f46c8"></a>
<a id="u-science-fictions-ce61b7ba2951"></a>

<a id="v2-247-0"></a>

**Multiverse Analysis and Transparent Analytic Practice**

Ritchie describes multiverse analysis as an alternative to p-hacking that embraces rather than hides analytic flexibility. Instead of picking a single analytical path that can be tweaked to produce a desired result, researchers run every statistically justifiable analysis on a dataset and present the overall pattern of results across all those variations. Ritchie argues that this approach acknowledges the many decision points involved in data analysis and reduces bias by exposing the full range of possible outcomes—making the presentation of evidence more honest and transparent.

<a id="b-8ec97716a7fb8945"></a>
<a id="u-science-fictions-cdad7d567268"></a>

<a id="v2-248-0"></a>

**Shortform perspective: Practice alternatives with small exercises**

<a id="v2-278-2"></a>

Shortform’s informal exercises help make alternatives visible. They are analogies for inquiry, not substitutes for multiverse analysis of one dataset:

- For a major choice, list plausible scenarios in a spreadsheet, assign tentative probabilities and calculate weighted outcomes. Consider at least three explanations for an event: a promotion might reflect your performance, company needs and market conditions. These are possible contributions to investigate, not three established causes.
- Track mood, productivity and interactions, then make a weekly visual map with colored pens or a digital tool. A pattern such as productive days following exercise suggests a question; it does not establish that exercise caused the difference.
- In a “flexible data challenge,” friends predict sports results or local weather and discuss why forecasts differed. In decision role-play, each person shares a choice and the group acts out possible outcomes. Both expose assumptions and perspectives.
- Try a modest change for a week—for example, order tasks by estimated duration instead of priority—and record what happens. A personal before/after trial teaches observation but cannot isolate every other change. A separate “Simplicity Challenge” asks a group to find a straightforward solution to a budget vacation or household problem; simplicity still has to meet the requirements.
- For a week, let a coin or random-number generator choose small, inconsequential things, such as the next topic to read. Compare those outcomes with what you would have chosen to notice the influence of expectations and desires.
- At the start of a meeting, state your goals and concerns and invite others to do the same. This “no hidden agenda” practice supports candid discussion without guaranteeing agreement or trust.

<!-- atlas:block-content b-d306b9c2a0a59ca6:end -->
<!-- atlas:block-content b-459d1bf6df9c046c:start -->
### Which completed studies become visible?

<a id="b-459d1bf6df9c046c"></a>
<a id="u-science-fictions-a15482a37475"></a>
<a id="u-science-fictions-b260b14774f0"></a>
<a id="u-science-fictions-f38c2e413e7c"></a>
<a id="u-science-fictions-9422e2bf9647"></a>
<a id="u-science-fictions-799e624437d6"></a>
<a id="u-science-fictions-f0f688a02f02"></a>
<a id="u-science-fictions-0b4ce07ef225"></a>
<a id="u-science-fictions-18c62318d7d4"></a>
<a id="u-science-fictions-d35dccdbeaa9"></a>
<a id="u-science-fictions-4605c7bbe04c"></a>

<a id="v2-249-0"></a>

**How Publication Bias Distorts the Evidence**

Ritchie discusses publication bias—the tendency of scientists and journals to favor publishing positive results that support a hypothesis over null results that find no effect. Journals often prioritize exciting, novel findings that capture readers' and reviewers' attention, leading to a disproportionate share of published positive findings. This preference for favorable results creates a strong disincentive for researchers to submit studies with null results, further reinforcing the bias in the published literature.

Funnel plots can suggest small-study effects or missing evidence in a meta-analysis by relating study precision or sample size to estimated effect size. An absence of small studies with minor or null effects is compatible with selective publication, but is not proof of it: heterogeneity, study methods, chance and other small-study effects can also create asymmetry. Selective publication can inflate a pooled effect and mislead users of the synthesis; the diagnostic needs its own assumptions checked.

A more direct approach is to follow completed projects into writing and publication. Ritchie cites Annie Franco and colleagues’ study of preapproved social-science projects. The local summary reports that strong findings made up 41% of completed studies and 53% of published papers. Those are compositions with different denominators—not publication probabilities conditional on a strong result. They remain figures reported by this summary, pending verification in the underlying tables. The primary abstract separately reports that strong-result studies were 40 percentage points more likely to be published and 60 points more likely to be written up than null-result studies; those contrasts answer a different question and cannot replace the 41%/53% figures. [Franco and colleagues, Science (2014)](https://pubmed.ncbi.nlm.nih.gov/25170047/).

<a id="b-93fe7bfe811221ed"></a>
<a id="u-science-fictions-5da97a5abec0"></a>

<a id="v2-252-0"></a>

**Shortform perspective: Publication-bias diagnostics need their own checks**

Shortform mentions Egger’s test, trim-and-fill and sensitivity analyses as supplements to funnel plots. Each relies on assumptions; a small set of heterogeneous studies can make asymmetry hard to interpret, and an imputed set of “missing” studies is not recovered data. Comparing conclusions under plausible missing-evidence scenarios can expose fragility, but cannot guarantee correction. Training in critical appraisal helps readers ask these questions; it does not make professionals immune to an exaggerated pooled effect.

<a id="b-9e748e06b9abd56a"></a>
<a id="u-science-fictions-8da8d15c38a1"></a>
<a id="u-science-fictions-f07704b19cee"></a>

<a id="v2-251-0"></a>

**Shortform perspective: How Large and Durable Is Publication Bias?**

The incentives behind publication bias are real, but Shortform’s supplementary material identifies several trends that may mitigate them. Growing support for open science practices such as pre-registration and open data encourages publication of all research outcomes. Journals may prioritize methodological rigor and scientific significance regardless of whether results are positive or negative. Funding agencies and institutions may require publication of all results as a condition of grants. Some fields may naturally have more positive findings because of the kinds of questions they ask or their stage of research maturity.

Shortform presents education, advocacy, results-neutral publication and replication outlets as possible ways to change the culture. The existence of such outlets does not measure how far publication bias has declined. A null result may lose relevance as a question changes, just as a positive result can; its original contribution to the evidence record does not disappear merely because the field moves quickly.

<a id="b-7a7cf03f57a990d8"></a>
<a id="u-science-fictions-9655780eb6fb"></a>
<a id="u-science-fictions-5817c0741406"></a>
<a id="u-science-fictions-d4520d7e9524"></a>
<a id="u-science-fictions-ea029bb2f158"></a>
<a id="u-science-fictions-7e97c78ac8a5"></a>
<a id="u-science-fictions-3cf15aacc1f0"></a>
<a id="u-science-fictions-97cdd88e9c9b"></a>
<a id="u-science-fictions-8808784a7b84"></a>
<a id="u-science-fictions-bff51b1be587"></a>

<a id="v2-250-0"></a>

**Bringing Null Results Out of the File Drawer**

The "file drawer" phenomenon describes a core problem in scientific publishing: studies that yield null or ambiguous results are often hidden away in researchers' file drawers or abandoned entirely, rather than being published. Ritchie argues that this bias against publishing null results is a major issue, because null findings can be just as informative as positive ones. They help clarify the overall state of knowledge in a research area and prevent other teams from wasting resources on unnecessary duplication of work. Publication bias, in his view, is a major factor behind the misleadingly positive picture of research findings presented in the literature.

<a id="b-c11a5c85f1fb56dd"></a>
<a id="u-science-fictions-5f2298ce8922"></a>
<a id="u-science-fictions-12d8943ede4f"></a>
<a id="u-science-fictions-fa857a2f8af4"></a>
<a id="u-science-fictions-6b68d7cadae9"></a>
<a id="u-science-fictions-81bdbc9381e7"></a>
<a id="u-science-fictions-7e38ae5d5dec"></a>

<a id="v2-253-0"></a>

**Publishing Reforms to Counter Bias**

Journals that specialize in publishing null results have had only limited success, according to Ritchie. Because these venues are associated with low prestige, researchers seeking career advancement are discouraged from submitting their work there. As a result, such journals often end up as repositories for research that other outlets have rejected, which reinforces the perception that null findings are inherently less valuable than positive ones.

Mega-journals—large online publications that accept all methodologically sound studies regardless of their results or perceived impact—represent one promising model for reducing publication bias. Ritchie explains that these journals have no novelty requirement, and they create a venue where replications and results that do not confirm a hypothesis can be published alongside positive findings. He notes, however, that mega-journals still struggle with lower prestige compared to more selective, high-impact outlets.

<a id="v2-250-1"></a>

**Reforming Publishing Incentives and Replication**

Within Ritchie’s broader “Transparency and Open Science” reform agenda, countering publication bias and improving reliability requires a cultural shift: science must come to value all high-quality research regardless of its results, and build a more open and transparent publication system. A key part of this broader effort is revealing hidden findings—that is, bringing unpublished null results out of the file drawer and making them visible to the research community.

Some journals and funders are beginning to recognize the value of replication work and are explicitly inviting submissions that aim to verify earlier findings. A "you damage it, you pay for it" mentality is also taking shape: journals are being encouraged to publish replications of papers they originally ran, especially high-profile or controversial ones, as a way to take responsibility for the literature they curate.

<a id="b-9e4151ae9e24f216"></a>
<a id="u-science-fictions-328d0284cac9"></a>
<a id="u-science-fictions-10effa8bdc4d"></a>
<a id="u-science-fictions-7138a07b8c6b"></a>

<a id="v2-254-0"></a>

**Shortform perspective: Advocating for Transparency and Reducing Publication Bias**

Shortform proposes several distinct forms of advocacy:

- Write to editors of journals you read about Registered Reports. Explain that results-neutral acceptance depends on the agreed protocol and quality conditions, rather than promising unconditional publication. Letters or reasoned blog posts about a particular controversial study offer another way to participate in public scrutiny.
- Give a community talk or workshop on why null and inconclusive findings matter. Comparing natural and store-bought cleaning products is a possible low-stakes demonstration; an inconclusive outcome needs honest interpretation.
- Ask university boards or research committees for independent panels to handle conflicts of interest. Prepare a personal misconduct-reporting plan: identify contacts, steps and relevant whistleblower protections before a problem arises.
- Discuss citation ethics in professional forums, study groups or informal meetings. These conversations should strengthen relevant, honest citation rather than create reciprocal citation deals.

To learn about peer review, seek a supervised activity or a clearly defined public or patient contribution suited to your experience. Volunteering without subject expertise does not make someone qualified to judge a paper’s technical validity. Disclose competence and conflicts; formal editors must match the reviewer to the task.

<a id="b-529b39ef10cb12e9"></a>
<a id="u-science-fictions-76b83c776cbe"></a>
<a id="u-science-fictions-e8a1208cef81"></a>
<a id="u-science-fictions-865d24b51b00"></a>
<a id="u-science-fictions-cb5cc61dea61"></a>

<a id="v2-255-0"></a>

**Shortform perspective: Valuing negative and null results in research and daily life**

When assessing a health claim, include databases and evidence syntheses that contain null findings, rather than selecting only significant studies. The shared-record practices below apply the same visibility principle to failed projects and ordinary outcomes; a personal log remains different from a controlled study.






Mega-journals, which accept a broader range of solid studies without requiring novelty, help democratize the scientific process by giving a voice to researchers from diverse backgrounds who might otherwise struggle to publish in high-impact journals. However, they often rely on article processing charges (APCs) paid by authors or their institutions, which can create a barrier for researchers from underfunded institutions or countries.

<!-- atlas:block-content b-459d1bf6df9c046c:end -->
<!-- atlas:block-content b-2b619df79c0f34cb:start -->
### Separate fraud, error and a numerical flag

<a id="b-2b619df79c0f34cb"></a>
<a id="u-science-fictions-52020768de15"></a>
<a id="u-science-fictions-5511efb92a48"></a>
<a id="u-science-fictions-69e36ea1e032"></a>
<a id="u-science-fictions-fc3b5a17c7dc"></a>
<a id="u-science-fictions-7d764cd49188"></a>
<a id="u-science-fictions-7ecfaed3aeb1"></a>
<a id="u-science-fictions-bd21723e2b7d"></a>
<a id="u-science-fictions-d6bb1d15eaad"></a>
<a id="u-science-fictions-403d98ed817c"></a>
<a id="u-science-fictions-516594615f54"></a>

<a id="v2-256-0"></a>

**Problematic Research Practices: Misconduct, Negligence, and Unconscious Bias**

Ritchie frames problematic research practices as a prevalent, systemic issue that cuts across many scientific disciplines and undermines the reliability and trustworthiness of research findings. He emphasizes that scientific studies are not always distorted by intentional fraud; much of the problem stems from unconscious biases that shape study design, analysis, and interpretation, producing misleading results. These biases are subtle and pervasive, affecting many elements of the research process, and even when unintentional they can strongly influence scientific writing and warp understandings of complex phenomena.

Beyond these subtler forms of bias, Ritchie identifies two additional broad categories of problematic practice that erode scientific reliability: intentional scientific misconduct, and negligence in day-to-day work. Outright fraud—where researchers deliberately fabricate or falsify data or findings to produce desired outcomes—inflicts severe harm on both the scientific community and the public. Negligence, meanwhile, takes many forms, from basic numerical mistakes to fundamentally flawed experimental designs; carelessness or oversight can seriously erode the reliability and validity of scientific findings.

<a id="b-3f7bc253f6a7d8eb"></a>
<a id="u-science-fictions-cdcf65d62738"></a>
<a id="u-science-fictions-269838cb520c"></a>
<a id="u-science-fictions-db12813586f4"></a>
<a id="u-science-fictions-963b7e3955a5"></a>

<a id="v2-264-0"></a>

**Data Fabrication and Its Detection**

Data fabrication—inventing or altering numerical data to fit a theory—is a common form of scientific fraud. Ritchie stresses that generating convincing fake datasets is harder than it may appear, because real data carry inherent noise from measurement error and sampling. That natural noisiness is precisely what helps expose fraud: fabricated datasets often look too clean, lack natural variability, or show suspiciously similar patterns across different groups. Ritchie illustrates this with two high-profile cases: that of Diederik Stapel, whose fabricated data were remarkably free of inconsistencies, and that of Michael LaCour, whose data appeared to have been generated by simply adding random noise to an existing dataset.

One arithmetic check is GRIM, explained below in “What GRIM can and cannot flag”. It examines whether a reported mean can arise from the stated discrete data and sample size; it flags inconsistency with those assumptions, not “statistical improbability” and not guilt.

<a id="b-320d2e38869cda6a"></a>
<a id="u-science-fictions-3220b4d800ed"></a>
<a id="u-science-fictions-b619f9363618"></a>
<a id="u-science-fictions-0500b48c5545"></a>
<a id="u-science-fictions-9de0aafb0a68"></a>
<a id="u-science-fictions-890f829808e2"></a>
<a id="u-science-fictions-258012263a99"></a>
<a id="u-science-fictions-54fb5b51f78e"></a>

<a id="v2-265-0"></a>

**Scientific Misconduct: Fraud and Fabrication**

Image manipulation is a frequent method of scientific deceit, especially in biomedical research. Ritchie describes how researchers can alter images—such as western blot results and micrographs—using tools like Photoshop to make them more convincing or to support a preferred hypothesis. He details two prominent cases: the Woo-Suk Hwang scandal, in which identical images were presented as coming from different cell lines, and the Haruko Obokata case, in which images were spliced, resized, and color-corrected to fabricate evidence for a new stem-cell technique.

Ritchie maintains that greater transparency is essential for combating research fraud. He proposes that publicly exposing individuals found to have engaged in research misconduct would not only deter potential fraudsters but also hold academic institutions and publications accountable for addressing such issues more effectively. He further argues that universities should not be responsible for investigating misconduct among their own researchers, because institutional self-interest creates inherent conflicts of interest that can compromise the integrity of inquiries. He cites the Karolinska Institute as a cautionary example: the institute initially defended surgeon Paolo Macchiarini, who had performed a series of experimental windpipe surgeries, before being forced to acknowledge his fraud. Ritchie frames independent, third-party investigations as a necessary reform to protect the credibility of research oversight.

<a id="b-a5f33d06b5d01eb1"></a>
<a id="u-science-fictions-3c0d2edd016d"></a>

<a id="v2-266-0"></a>

**Shortform perspective: Data Fabrication Cases and Institutional Responses**

Beyond the cases discussed by Ritchie, another major example is the Schön scandal in physics, in which fabricated data resulted in multiple retracted papers. Past fabrication cases have raised broader awareness and spurred the development of better detection methods, with high-profile scandals prompting institutions to adopt stricter data management and verification protocols. Advanced software tools and algorithms can now analyze datasets for telltale signs of manipulation, such as unusual clustering of data points or an absence of expected randomness. After his fraud was exposed, Stapel lost his academic position and suffered major professional and personal consequences. LaCour's study had claimed that a brief conversation with a gay canvasser could substantially shift people's views on same-sex marriage; the result was hailed as groundbreaking and received widespread media coverage.

<a id="b-30f50ba405533ccc"></a>
<a id="u-science-fictions-f10f7ff2d5ae"></a>
<a id="u-science-fictions-85bfcab0d8b6"></a>
<a id="u-science-fictions-c1acb0cd17ec"></a>
<a id="u-science-fictions-9616cfd31aa3"></a>
<a id="u-science-fictions-0b3265b364f2"></a>
<a id="u-science-fictions-18d1717026b7"></a>
<a id="u-science-fictions-7cc639448047"></a>
<a id="u-science-fictions-6c6621d92d34"></a>
<a id="u-science-fictions-e2ad2a9fff2c"></a>
<a id="u-science-fictions-d59f29c675ec"></a>
<a id="u-science-fictions-1c24c515ae59"></a>
<a id="u-science-fictions-111de7b03197"></a>

<a id="v2-257-0"></a>

**Negligence and Methodological Shortcomings in Research Practice**

Numerical errors are widespread in scientific papers, often caused by inattention, negligence, or simple carelessness. Such errors can distort results and lead to wrong conclusions, and they are especially worrying in high-stakes fields such as medical research. Ritchie emphasizes that attention to detail is essential in science: even seemingly minor oversights, such as typos or copy-paste mistakes, can have significant consequences. The pressure to publish quickly, combined with the complexity of modern data analysis, can lead researchers to skip proper checks and balances, introducing errors into the literature that can be very hard to correct later.

Cell line misidentification is a long-standing, solvable problem in biological research that the scientific community has been slow to address. Researchers can mistakenly use the wrong cell type, and contamination from faster-growing cell lines can spread widely, undermining the validity of many studies—especially in fields like cancer research. According to Ritchie, this issue has been well documented for more than fifty years, yet the response has been inadequate. He frames the persistent inaction as a cultural blind spot: a tendency to prioritize exciting results over methodological rigor allows a systemic flaw to continue, wasting resources and producing a steady stream of misleading, error-prone publications.

Underpowered animal research—studies that use too few animals to reliably detect the effect being investigated—poses both scientific and ethical problems. Drawing on work by Malcolm Macleod and others, Ritchie notes that only a small fraction of animal studies adequately address statistical power. He argues that this practice is scientifically unsound and, more sharply, ethically problematic: when sample sizes are too small to reliably pick up real effects, the harm and death inflicted on the animals risk being rendered meaningless. For animal research to be ethically justified, in his view, studies must be carefully designed with enough subjects to produce meaningful results. Ritchie calls for a broader cultural shift in science, one that puts methodological rigor ahead of chasing flashy findings—especially when sentient beings are sacrificed in the process.

<a id="b-791386165df46059"></a>
<a id="u-science-fictions-0384dd65bda2"></a>

<a id="v2-259-0"></a>

**Shortform perspective: Cell line contamination: a case study of neglected methodological error**

The long-running problem of cell line contamination illustrates how even solvable methodological issues can persist for decades with serious consequences.

The issue first came to wide attention in the 1960s, most notably through the work of Walter Nelson-Rees, who identified its widespread presence in cancer research. The financial costs are substantial: unreliable data can force entire studies to be repeated or discarded.

Although better authentication practices—such as short tandem repeat (STR) profiling—have been called for, they have not been universally adopted. Technological advances, including DNA barcoding and commercial authentication services, can help verify cell line identities, but realizing their benefit requires broad uptake and consistent implementation across laboratories. At the core of the issue, using the wrong cell line can invalidate findings entirely, because the biological traits under study may not belong to the intended cell type at all.

<a id="b-249f022db280bfe6"></a>
<a id="u-science-fictions-18551046ab7a"></a>
<a id="u-science-fictions-ee162fff5820"></a>

<a id="v2-258-0"></a>

**Shortform perspective: Negligence, Errors, and Underpowered Research**

Carelessness and inattention to detail remain serious problems in science despite technological advances. Emphasizing attention to detail in scientific education and training helps produce meticulous researchers who are better able to carry out high-quality studies. Human oversight remains critical: simple mistakes such as mislabeling data or entering it incorrectly can have major repercussions. Modern research also relies heavily on complex software and statistical methods, and errors in code or analysis scripts can be subtle and very hard to catch without thorough review.

Animal research ethics includes the 3Rs: Replacement, Reduction and Refinement. Reducing numbers without preserving the ability to answer the question can waste animals as well as information. Statistical power is the probability of rejecting the null under a specified alternative effect and analysis; it depends on effect size, variability, design, sample size and threshold. Low power means a greater risk of missing that specified effect. The objective is an informative, ethically justified design, not merely the smallest or largest possible study.

<a id="b-7c8d79bfb340795d"></a>
<a id="u-science-fictions-898be78325ab"></a>
<a id="u-science-fictions-a98a91d34a2f"></a>
<a id="u-science-fictions-69b9b538a9b3"></a>
<a id="u-science-fictions-f1a242a445b0"></a>
<a id="u-science-fictions-591c231b71c1"></a>
<a id="u-science-fictions-c253383a9250"></a>

<a id="v2-260-0"></a>

**Technological and Automated Tools to Improve Reliability**

Ritchie highlights statcheck, an algorithm that applies statistical rules to detect numerical inconsistencies within papers, as a valuable defense against negligence. As an illustration, he notes that when applied to a large sample of psychology papers, it uncovered a substantial number of errors. He recommends that journals and researchers integrate statcheck into their standard workflows to catch potential errors before publication, thereby avoiding the embarrassment and harm to scientific credibility that come from publishing inaccurate findings.

Linking analysis code to tables, figures and prose can reduce manual transcription errors and make updates traceable. It does not eliminate mistakes in code, data, labels or interpretation. Automated checks such as statcheck flag particular inconsistencies among reported statistics; their scope and assumptions limit what they can detect, and a flag needs examination rather than an automatic misconduct verdict.

Ritchie also outlines the potential of automated algorithms to identify fabricated data or images in scholarly articles. These algorithms scan for telltale signs of manipulation, such as suspiciously uniform distributions, implausible results, or contradictions among related statistics. He proposes that academic publications could deploy such algorithms as an initial filtering step to flag papers that merit closer, more detailed examination.

<a id="b-b4dd0bcd8f3f99db"></a>
<a id="u-science-fictions-19db1bcab3e9"></a>
<a id="u-science-fictions-deaa68fd20d3"></a>

<a id="v2-263-0"></a>

**Shortform perspective: What GRIM can and cannot flag**

GRIM—the Granularity-Related Inconsistency of Means test—checks arithmetic compatibility. Given equally weighted observations on a known discrete scale, the sample size, number of items when applicable, and the reported rounding precision, only certain means are possible. For 20 integer-valued single-item responses, the unrounded mean lies on a 0.05 grid; a reported mean of 3.48 to two decimals is incompatible with those assumptions. This illustration applies the method described by Brown and Heathers (DOI: 10.1177/1948550616673876).

GRIM does not assume a uniform probability distribution. Rare sampling variation cannot produce an arithmetically incompatible mean. A flag instead calls for checking the actual denominator, missing responses, weighting, scale increments, transformations, number of items, rounding and transcription. An error in that description can explain a flag; a flag is not proof of fraud. Conversely, a compatible mean does not validate the data, and GRIM cannot detect every kind of manipulation.

Shortform’s separate caution about transparency remains useful: exposing misconduct after the event cannot replace prevention, and public shaming can discourage admission of honest errors. Separate an established finding of misconduct from a numerical anomaly that still needs explanation.

<a id="b-fd300585db710fdc"></a>
<a id="u-science-fictions-1ac04f050a08"></a>
<a id="u-science-fictions-515a50dd9f4e"></a>
<a id="u-science-fictions-9d536a52ed4f"></a>
<a id="u-science-fictions-13aba76a69e0"></a>
<a id="u-science-fictions-97a75a02a578"></a>
<a id="u-science-fictions-72bb93138867"></a>
<a id="u-science-fictions-9ed7a04258aa"></a>
<a id="u-science-fictions-e6d20ea9a252"></a>
<a id="u-science-fictions-4152631a6e0c"></a>

<a id="v2-262-0"></a>

**Combating Fraud and Reducing Negligence**

Efforts to improve the reliability of science include targeted strategies against both fraud and carelessness. To address fraud specifically, Ritchie emphasizes the need for transparency, accountability, and technology-assisted verification, supported by mechanisms such as independent investigations and automated fraud detection algorithms that identify manipulated data or images. For the unintentional mistakes that erode accuracy—negligence in day-to-day work—solutions include new automated software and a stronger emphasis on attention to detail.

<a id="b-7a885562ad28265d"></a>
<a id="u-science-fictions-005999c55eab"></a>

<a id="v2-261-0"></a>

**Shortform perspective: Detecting Numerical Errors and Strengthening Research Rigor**

Readers can strengthen their critical evaluation skills by actively checking numerical consistency in the papers they read. For instance, when a paper reports a percentage, one can verify whether the stated numerator and denominator actually produce that figure; this practice sharpens attention to detail and deepens understanding of how results are presented. Another practical step is to form a peer-review buddy system with colleagues, exchanging drafts before submission and cross-checking statistics and findings; this mutual support catches errors and fosters a collaborative culture of research quality.

<a id="b-c9ae449f722ecd5d"></a>
<a id="u-science-fictions-808aef744b95"></a>
<a id="u-science-fictions-82b22df11ada"></a>
<a id="u-science-fictions-9177eba79910"></a>

<a id="v2-267-0"></a>

**Shortform perspective: Text recycling, image manipulation, and research integrity**

Questionable research practices include self-recycling of previously published text and the manipulation of images or data. A range of strategies can help uphold integrity, both individually and collectively.

- Check drafts against others’ work and your earlier publications; investigate matches and supply needed citations.
- Use peer review to decide whether reuse is justified, permitted, attributed and clearly distinguished from a new contribution. A shared archive is not a pool of text to republish without permission.
- Collaborate across disciplines to contribute new questions, methods and expertise.
- A “truth or fiction” image-and-story game can prompt discussion of manipulation cues; follow it with source checking rather than treating a game score as expertise.
- Adopt sharing rules such as checking a dataset’s origin and using reverse image search before reposting.
- A voluntary data-integrity pledge or badge can remind a team of its commitment, but does not replace verifiable records and review.

Shortform notes that image fabrication is only one integrity problem; plagiarism, falsification and undisclosed conflicts also matter. Image software can be used legitimately, but intent alone does not decide acceptability. The relevant standards include whether processing is disclosed, preserves the evidence, and follows the field’s reporting rules.

<a id="b-508ad8fab90cd32a"></a>
<a id="u-science-fictions-ed5e9b9030a8"></a>
<a id="u-science-fictions-d7829a0915ff"></a>

<a id="v2-268-0"></a>

**Shortform perspective: Fake peer review: risks, countermeasures, and counterperspectives**

Fake peer review—including manipulated reviewer identities and self-review schemes—undermines the integrity of the publication process, though the scale and implications of the problem are debated.

On the practical side, two tips help build a more robust review culture. First, educate your network about the importance of email verification through social media or blog posts, sharing guidance on identifying and reporting suspicious email activity so the community can collectively guard against deceptive practices. Second, use legitimate peer feedback platforms—such as writing groups, professional networks, or field-specific online forums—to exchange honest critiques instead of resorting to self-review, thereby improving your work through constructive feedback while upholding ethical standards.

Shortform offers qualifications rather than evidence that reviewer fraud is negligible. Knowing an author does not automatically make a reviewer dishonest, and author suggestions can identify relevant expertise; journals still need independent identity and conflict checks. An ORCID identifier helps distinguish researchers and connect records, but possession of an identifier does not itself verify expertise or independence. Existing safeguards can reduce opportunities for deception without proving that most reviews in a particular sample were sound.

<!-- atlas:block-content b-2b619df79c0f34cb:end -->
<!-- atlas:block-content b-d5f5bdbb37d5480c:start -->
### Funding, shared beliefs and the limits of extrapolation

<a id="b-d5f5bdbb37d5480c"></a>
<a id="u-science-fictions-a8ff1f337a95"></a>
<a id="u-science-fictions-4a2308da01b8"></a>
<a id="u-science-fictions-315907dd615a"></a>
<a id="u-science-fictions-25e11c1822cb"></a>
<a id="u-science-fictions-daba277e83c3"></a>
<a id="u-science-fictions-5d47a19478af"></a>
<a id="u-science-fictions-daf1aed524c2"></a>
<a id="u-science-fictions-f9b1a1cf16d1"></a>
<a id="u-science-fictions-675246c5a20e"></a>
<a id="u-science-fictions-5be2a8d07014"></a>

<a id="v2-269-0"></a>

**External Sources of Bias: Funding, Ideology, and Gender**

Ritchie identifies several external sources of bias that can skew research and misrepresent scholarly publications: funding sources, researchers' ideological commitments, and gendered patterns in science.

One well-documented concern is industry-funded research, especially clinical trials, where funders have a direct financial stake in the success of a drug or treatment. Ritchie cites meta-scientific evidence showing that trials with industry funding are more likely to report positive results than trials funded from other sources, raising concerns about bias and the possible suppression of unfavorable findings.

Political prejudices can also shape scholarly work. Ritchie explores how researchers' political ideologies may influence which scientific topics they prioritize and how they interpret evidence. As an illustration, he points to the liberal skew in psychology research, noting concerns that the field's lack of conservative viewpoints may weaken peer review's ability to detect biases.

Sexism in scientific research manifests in multiple forms, from the underrepresentation of women across scientific fields to biases embedded in study design and interpretation. Ritchie argues that unconscious biases often lead researchers to overlook the importance of studying both sexes—particularly in animal research—potentially narrowing the generalizability of findings and slowing progress in fields such as medical treatment. The "feminist science" movement, which centers women's perspectives and experiences in research, has been proposed as a corrective to this bias, though Ritchie acknowledges concerns that explicitly injecting political ideologies into scientific work could prove counterproductive.

<a id="b-6a8fa73a8765da1d"></a>
<a id="u-science-fictions-841d5a169afe"></a>
<a id="u-science-fictions-d93c7153716c"></a>

<a id="v2-272-0"></a>

**Shortform perspective: Funding, Industry Ties, and Financial Conflicts of Interest**

Industry-funded trials have been associated with a higher share of positive results, raising concerns about bias. However, collaboration between industry and academia can drive advances in knowledge and technology that benefit public health. Some industry-funded studies include built-in safeguards such as independent data monitoring committees to help ensure objectivity. Industry funding can also support research that would otherwise go unfunded, contributing to progress in under-resourced areas.

Shortform notes that ideology, culture and personal commitments can create bias alongside money. Ethical commitments may motivate honest work but cannot prevent unconscious bias. Disclosure also has costs, and rules need a clear account of which interests are relevant and how they will be assessed. Those trade-offs call for proportionate, enforceable disclosure, not an assumption that professional integrity makes conflicts harmless or that disclosure alone removes them.

<a id="b-132c278f34734fa7"></a>
<a id="u-science-fictions-af14db838ffb"></a>
<a id="u-science-fictions-8ef57f21c288"></a>
<a id="u-science-fictions-bb6244ff7aa0"></a>
<a id="u-science-fictions-52118043a1ed"></a>
<a id="u-science-fictions-147fb3ab244d"></a>
<a id="u-science-fictions-c2401d23dd28"></a>
<a id="u-science-fictions-68b273b94d31"></a>
<a id="u-science-fictions-49237a63d792"></a>
<a id="u-science-fictions-bcb35ae6f525"></a>

<a id="v2-270-0"></a>

**Groupthink, Funding, and Ideological Biases**

Ritchie analyzes groupthink within research communities, a dynamic that emerges when a field becomes overly homogeneous in its perspectives and background assumptions, eroding the critical scrutiny on which scientific progress depends. Shared ideological commitments, he argues, can lead researchers to overlook weaknesses in theories they favor and to resist new ideas that challenge the prevailing consensus. As a concrete example, he cites the amyloid cascade hypothesis in Alzheimer's disease research, where proponents of the theory were accused of acting as a "cabal" that suppressed dissenting voices and slowed progress by refusing to engage with alternative explanations. Research exhibiting gender bias is also identified as a category of bias relevant to the influence of funding and ideology.

Under the heading of accumulated advantage, Ritchie discusses the Matthew Effect in science: the tendency for already well-funded and established researchers to secure additional funding more easily than newer or less well-known investigators. This bias concentrates resources in the hands of a relatively small group, putting early-career researchers at a structural disadvantage even when they have innovative ideas and the potential to make important contributions. By limiting the diversity of perspectives and discouraging work on unconventional or risky ideas, this concentration of funding can, in Ritchie's view, slow overall scientific progress.

<a id="b-85577561e9c06065"></a>
<a id="u-science-fictions-21804f27d1cb"></a>

<a id="v2-271-0"></a>

**Shortform perspective: Homogeneity is a risk factor, not a diagnosis**

Shortform offers possible explanations for reviewer agreement: a mature field may share well-tested conclusions; norms of criticism and openness to dissent can preserve scrutiny; and shared background may sometimes support candid collaboration. These possibilities do not establish that groupthink is absent. In the Alzheimer’s example, alternative hypotheses and debate need to be examined alongside incentives favoring established pathways. Structural resistance and deliberate suppression are different explanations; neither follows merely from reviewers’ agreement or disagreement.

<a id="b-7cc99332b5e71e2a"></a>
<a id="u-science-fictions-b4e69ff1f24f"></a>

<a id="v2-273-0"></a>

**Shortform perspective: Sex, representation and the limits of extrapolation**

Shortform presents several distinctions in debates about gender and science. Equal opportunity and equal representation are not identical goals, and a representation gap alone does not identify its cause. Replication and critical testing can reduce individual bias, but their existence does not demonstrate that a particular bias has been removed. Debate also continues over how explicitly political commitments should enter research, and whether an emphasis on a researcher’s gender can distract from methodological quality.

A single-sex study can answer a question about that studied sex and may be appropriate for a specified design. Controlling variability within it does not establish generalizability to an unstudied sex. Relevant sex differences need evidence, not a presumption that rigorous controls make them disappear. Clinical trials and regulation can evaluate safety and efficacy before approval; post-market surveillance operates afterward. None guarantees identical effects or absence of risk across all populations. The scientific task is to make the population, question and limits of inference explicit.

<a id="b-235daa650c1e0af3"></a>
<a id="u-science-fictions-39d1e1043323"></a>
<a id="u-science-fictions-bf07f3dfc73f"></a>

<a id="v2-274-0"></a>

**Shortform perspective: Mitigating personal bias and ideological influence**

Researchers' beliefs and ideological commitments can subtly shape their work, and similar dynamics operate in everyday judgment. Several practices can help counteract these influences.


- Broaden your media consumption to include underrepresented perspectives. For instance, seek out conservative perspectives on psychology by subscribing to a conservative psychology newsletter or following conservative-leaning thought leaders on social media.
- Practice blind assessment in everyday decisions to strengthen critical thinking. For example, when choosing a product or service, cover up brand names and focus solely on features and benefits. This strategy mirrors blind analysis in research, where the analyst is unaware of group assignments, thereby minimizing bias.
- Create a "blind feedback box" for any group projects or events you participate in, where participants can submit anonymous suggestions or observations. This could be a physical box at an in-person event or an online anonymous survey. The key is that you do not know who provided the feedback, which encourages unbiased interpretation and application of the suggestions.

<!-- atlas:block-content b-d5f5bdbb37d5480c:end -->
<!-- atlas:block-content b-266ec3d16d2fdc44:start -->
### What the headline leaves out

<a id="b-266ec3d16d2fdc44"></a>
<a id="u-science-fictions-e787ba9981fc"></a>
<a id="u-science-fictions-5cd204f9e38a"></a>
<a id="u-science-fictions-40b14d8faeb9"></a>
<a id="u-science-fictions-06191497e197"></a>
<a id="u-science-fictions-2040f5691442"></a>
<a id="u-science-fictions-ff71e79f95d0"></a>
<a id="u-science-fictions-7640a1de1c21"></a>
<a id="u-science-fictions-22b85a776620"></a>
<a id="u-science-fictions-f3b350a72a64"></a>
<a id="u-science-fictions-0e74053fad22"></a>
<a id="u-science-fictions-9fe615804f20"></a>
<a id="u-science-fictions-22e029114788"></a>
<a id="u-science-fictions-1f0d8f567a2f"></a>
<a id="u-science-fictions-101bb841cf09"></a>
<a id="u-science-fictions-431049afcd7d"></a>
<a id="u-science-fictions-5d5bef34628a"></a>
<a id="u-science-fictions-1ffcc3944d73"></a>
<a id="u-science-fictions-20b2b0ba06ad"></a>

<a id="v2-275-0"></a>

**Hype, Spin, and the Distortion of Science Communication**

Ritchie frames hype, overstatement, and distortion as pervasive problems in the communication of scientific findings—both to the public and among scientists. Driven by the pressure to attract attention and funding, these practices erode the trustworthiness of science, inflate public expectations, and obscure the nuanced, incremental nature of scientific progress.

Press releases can turn a study’s limited findings into stronger advice. Sumner and colleagues’ 2014 retrospective content analysis examined health-related press releases issued in 2011 by 20 UK universities and compared them with the associated papers. About 40% contained advice more direct or explicit than in the associated journal article; this is a proportion in that study’s material, not a current rate for all health press releases. Ritchie uses the result to show how exaggeration can enter before news reporting, including when researchers participate in writing releases. See [Sumner et al., BMJ (2014)](https://doi.org/10.1136/bmj.g7015).

Another recurring pattern is the exaggeration of early animal research. Ritchie notes that the great majority of findings from animal studies do not translate to humans, yet press releases routinely hype such results in ways that obscure their limited human relevance—creating a misleading impression of importance for human health. He cites the Twitter account @justsaysinmice, which appends "IN MICE" to the conclusion of sensational headlines about animal research, as a humorous illustration of how small linguistic choices can hide critical contextual information about a study.

Hype and spin also infiltrate scientific articles themselves. Scientists deliberately frame their findings as more impactful or definitive than the evidence warrants. In the medical literature, this can take the form of downplaying null outcomes and casting them as favorable, using weasel wording, selectively reporting only favorable outcomes, or glossing over limitations in study design—all of which can mislead clinicians and patients about the true efficacy of a treatment or diagnostic test.

Ritchie traces much of this behavior to distorted incentives within academia: constant pressure to win funding and publish in prestigious journals pushes researchers to exaggerate the significance of their work, creating a feedback loop that rewards flashy, attention-grabbing research over rigorous investigation and a balanced reading of the evidence.

<a id="b-84036893199a31f1"></a>
<a id="u-science-fictions-83952cde5fb8"></a>
<a id="u-science-fictions-c42bcb44a432"></a>
<a id="u-science-fictions-a926cb5d9e35"></a>
<a id="u-science-fictions-b759ce4ba1ce"></a>
<a id="u-science-fictions-aa0f4b2be6ae"></a>
<a id="u-science-fictions-d5223034b6e7"></a>
<a id="u-science-fictions-6e292d19fd6e"></a>
<a id="u-science-fictions-55d63e64c10c"></a>
<a id="u-science-fictions-cfceb3809549"></a>
<a id="u-science-fictions-89f0e785a585"></a>
<a id="u-science-fictions-ad577a9c4cd6"></a>
<a id="u-science-fictions-e1d98d765d10"></a>
<a id="u-science-fictions-77b5fcb6c79f"></a>
<a id="u-science-fictions-36c3624021ff"></a>
<a id="u-science-fictions-20d4e311db52"></a>

<a id="v2-276-0"></a>

**Hype and Misrepresentation in Science Communication**

Ritchie identifies a culture of hype that has spread through academic publishing over recent decades, even though it falls short of outright fraud. One visible marker is a dramatic rise in positive, subjective terms such as "innovative," "robust," and "unprecedented" in paper abstracts. Ritchie frames this as a form of "spin": scientists use favorably toned words and phrases to amplify the perceived impact of their findings, which in turn obscures the genuine uncertainty that surrounds research results.

Beyond journal articles, Ritchie examines exaggeration in popular science books, many of which are written by scientists themselves. To appeal to a broader readership, authors can overstate or misrepresent scientific conclusions, sometimes building attention-grabbing arguments on scant evidence. Ritchie is particularly critical of this tendency in fields such as psychology, education, and health, where authors motivated by book sales often prioritize crafting a compelling narrative over fidelity to nuance—oversimplifying complex scientific concepts or overstating real-world practical implications in the process of marketing science-based ideas.

In more severe cases, some popular science books cross the line from exaggeration into outright misinterpretation and misrepresentation of findings. Ritchie documents examples where authors present distorted versions of graphs, omit inconvenient results, or make bold assertions about data that are demonstrably inaccurate. He also flags common overstatement patterns in public science communication such as conflating correlation with causation and cross-species extrapolation that outpaces the evidence.

Ritchie warns that hyping scientific results carries long-term damaging consequences. Exaggerated claims inevitably fail to live up to their promises, creating a repeated cycle of hype followed by disillusionment that erodes public confidence in scientific work and makes it harder to communicate solid, reliable findings effectively.

<a id="v2-239-2"></a>

**Shortform perspective: Scientific Hype and Its Consequences**

Scientific hype exaggerates the significance, potential, or certainty of findings to attract attention, funding, or public interest. Social media can spread the claim before a correction or retraction, and repeated exaggeration can make policymakers less willing to trust scientific advice.

Press releases sometimes recommend action from scant evidence, though the severity varies: some advice is broad guidance consistent with public-health goals rather than a claim warranted by that study alone. Media outlets and audiences share responsibility for how the information is propagated and interpreted. The incentives are also changing. Funders and journals increasingly demand transparent methods, explicit expected outcomes, and replicability, and researchers who overpromise can jeopardize future support.

<a id="b-46d02539dcbb67d6"></a>
<a id="u-science-fictions-78bfeb9a4624"></a>
<a id="u-science-fictions-5a27f97dad90"></a>

<a id="v2-277-0"></a>

**Shortform perspective: Simplification can be responsible; enthusiasm still needs evidence**

Shortform qualifies Ritchie’s critique of popular science. Expert authors can simplify accurately, editors and publishers share responsibility for checking claims, and a book may aim to provoke useful debate as well as sell copies. Accessible explanations can lead readers toward deeper study. These are reasons not to infer dishonesty from readability; they do not excuse a conclusion that outruns the evidence.

Its practical suggestions can be used with the same constraint. In a presentation, show enthusiasm by specifying a strategy’s promise and the evidence behind it, rather than asserting “unprecedented growth” without support. Compare moderately positive and more emphatic versions of a resume with trusted peers, asking which claims remain credible and substantiated. Practice translating research into a plain-language blog without hyperbole. The exercise is to make accurate content engaging, not to test how much exaggeration an audience will accept.

<a id="v2-255-1"></a>

**Shortform perspective: Recognizing hype, spin, and misrepresentation in science communication**

Science communication is frequently distorted by hype, misleading language, and overstated conclusions. Practical steps can help readers and consumers navigate this landscape.

- Audit vague or inflated wording; Shortform’s “weasel-word jar” is a playful reminder, but accurate uncertainty words such as “may” are not defects.
- Learn research design well enough to distinguish random recruitment from random assignment. Recruitment affects representation; assignment helps causal comparability within a study.
- Prepare questions for a clinician about alternatives, risks, benefits, and the chance a test or treatment produces useful information or improvement.

Some apparent distortions in popular science may have less blameworthy explanations. Distorted graphs could result from simplification intended for a lay audience, designed to highlight trends without the technical complexity that might confuse non-expert readers. Omissions of certain results could be oversights rather than deliberate choices. Additionally, data interpretation can be inherently subjective, and different experts may legitimately disagree about the conclusions that can be drawn from the same dataset.

<a id="b-9d2324b4e0838a52"></a>
<a id="u-science-fictions-e3b18023cd75"></a>
<a id="u-science-fictions-51c3d60866a4"></a>
<a id="u-science-fictions-d1b856167a9d"></a>

<a id="v2-278-0"></a>

**Shortform perspective: Critically Evaluating Scientific Claims and Media Coverage**

Shortform suggests checking communication against its evidence: compare a press release with the original paper and its limitations, then compare coverage across news outlets for omitted context or sensationalism. An environmental-science blog or a “Narrative vs. Evidence” journal can use that same comparison to document a specific discrepancy, rather than judge a story by how compelling it sounds.

Two further exercises have different purposes. Suitable human-focused citizen-science projects on platforms such as Zooniverse or SciStarter can expose participants to research procedures, although participation alone does not establish expertise in translating animal findings to humans. A monthly “Science in Fiction” discussion chooses one concept from a science-themed novel or film, researches it and shares the findings. Both can cultivate curiosity and source checking; the quality of the evidence still decides the claim.

<a id="b-c3b22ba71e863859"></a>
<a id="u-science-fictions-40b244343ca4"></a>
<a id="u-science-fictions-d3b65cc09658"></a>

<a id="v2-286-0"></a>

**Shortform perspective: Moving Beyond P-Values to Assess Research Quality**

Shortform suggests examining effect size and uncertainty as well as p-values. For a weight-loss study, distinguish the estimated average treatment effect, its confidence interval, and the distribution of individual outcomes: these answer different questions. Ask whether a “20% faster” phone improves the tasks you actually perform. Its five-dimension check covers ethics, relevance, replicability, reporting transparency and impact. The regular discussion practice below provides a place to apply these questions, explain the finding plainly and return to its design and numbers.

<!-- atlas:block-content b-266ec3d16d2fdc44:end -->
<!-- atlas:block-content b-746fe7a4ac1891dd:start -->
### Commit before results; share what can responsibly be checked

<a id="b-746fe7a4ac1891dd"></a>
<a id="u-science-fictions-83e83c8196bd"></a>
<a id="u-science-fictions-1e7f9bc4a90d"></a>
<a id="u-science-fictions-abf4cd989c22"></a>
<a id="u-science-fictions-1af6ecda3a0f"></a>
<a id="u-science-fictions-7b42e4a1643a"></a>
<a id="u-science-fictions-f88bf1768ccd"></a>

<a id="v2-280-0"></a>

**Preregistration and Registered Reports do different jobs**

Preregistration creates a dated record of questions, hypotheses, outcomes and planned analyses before the relevant results are known. Ritchie recommends it to make HARKing, selective analysis and outcome-switching easier to detect. Researchers can still deviate or explore, but should disclose what changed, why, and which results are exploratory. Registration by itself neither reviews the design nor guarantees publication.

Registered Reports add review before results: a journal evaluates the research question and protocol, then can grant in-principle acceptance. At the second stage it checks execution, adherence to the accepted protocol, declared deviations, interpretation and prespecified quality conditions. Under the Center for Open Science’s description, acceptance is independent of whether the main finding is positive, null or novel **provided those conditions are met**; a study can still be rejected for relevant quality failures. This reduces selection based on the direction or significance of the result. It does not eliminate every source of publication bias, noncompletion, poor design, or misconduct.

See [Center for Open Science: Registered Reports](https://www.cos.io/initiatives/registered-reports).

<a id="b-8bb6262467f75dc8"></a>
<a id="u-science-fictions-a9e60b6d8bac"></a>
<a id="u-science-fictions-46079ee13d9e"></a>
<a id="u-science-fictions-296b344fda1a"></a>

<a id="v2-281-0"></a>

**Team science expands capacity when the design uses it**

Ritchie advocates collaboration across laboratories and disciplines to pool samples, skills and resources, coordinate replications and expose assumptions to varied perspectives. Larger teams do not automatically produce adequate power or low bias. Clear protocols, appropriate sample size, data quality, responsibilities and analysis plans determine whether the added capacity improves inference.

<a id="v2-267-1"></a>

**Shortform perspective: Collaboration, team science, and interdisciplinary work**

Cross-disciplinary work can add skills and perspectives. Turning that capacity into informative samples and reproducible results requires suitable design, data quality, coordination and analysis, as described above. A shared project is one way to begin:

- Use collaborative online tools such as Google Docs or Trello to organize a crowd-sourced research project. Set up a shared document or board where anyone can contribute data, resources, or ideas related to a specific question or problem. This approach draws on a diverse range of contributors and can yield a more comprehensive set of data or solutions than an individual researcher could compile alone.


<a id="b-08547e59f29091ca"></a>
<a id="u-science-fictions-1da93bbe6f24"></a>
<a id="u-science-fictions-b74c11181e6b"></a>
<a id="u-science-fictions-4101847be8b1"></a>
<a id="u-science-fictions-bb43eff291ab"></a>

<a id="v2-282-0"></a>

**Open Science and the Case for a Multifaceted Reform Agenda**

Open science makes methods, code, analyses and shareable data available for scrutiny, supporting replication and error detection. Ritchie argues that visibility can also deter misconduct. Sharing is constrained by consent, privacy, legal duties and legitimate restrictions; sensitive raw data may require controlled access or a justified alternative. Transparency increases the opportunity to check work, not a guarantee that anyone checked it or that its conclusions are correct.

Ritchie's proposed response to the broader set of problems is a multifaceted reform agenda. Improving scientific practice and fixing the academic reward system, he argues, will require both broad structural changes from universities, funders, and publishers, and grassroots efforts from working scientists, all aimed at realigning incentives to reward honesty, transparency, and the pursuit of reliable knowledge.

<a id="b-0b974319dba12d7c"></a>
<a id="u-science-fictions-f78558f43ff4"></a>
<a id="u-science-fictions-a8ffea31a015"></a>
<a id="u-science-fictions-f221679f708f"></a>
<a id="u-science-fictions-5b8eadab1c7a"></a>

<a id="v2-283-0"></a>

**Shortform perspective: Practical Actions for Supporting Research Integrity and Open Science**

Individuals can take a range of practical steps to support more reliable, open, and accountable science.

#### Replication and access
- Support credible replication projects, including suitable crowdfunding; the value depends on the design and execution.
- Share research through appropriate open-access journals or preprint servers such as arXiv and bioRxiv. A preprint server is not a journal, and posting there does not itself constitute peer review.
- Encourage institutional evaluation of rigor, relevance, accessibility and public contribution rather than journal prestige alone.

#### Engaging with scientific discourse and accountability
- Run small-scale personal replications of studies you read, and document the attempts on a blog to deepen your understanding of research methods.




#### Upholding citation and research ethics
- Keep a research journal reflecting on the purpose of your work—its contribution to knowledge rather than its citation potential—to anchor your practice in ethical values.

- Cross-check cited articles against databases such as Google Scholar or Web of Science to gauge their real influence beyond potentially inflated closed-group citation counts.
- Diversify your reading across authors, journals, countries, institutions, and career stages to avoid relying on a narrow set of sources.

#### Public engagement and everyday accountability
- Build a simple browser extension that alerts users to retractions and corrections in scientific papers they view online, raising public awareness of misconduct.

- Write reviews of books from lesser-known authors and publishers across multiple platforms to increase their visibility and create a more competitive market.
- Practice a “repair, don’t replace” ethic with household items to cultivate a sense of accountability for the things you own.

<a id="b-e11b9fde4eb314a3"></a>
<a id="u-science-fictions-a7331a877b8e"></a>

<a id="v2-284-0"></a>

**Shortform perspective: Keep plans, outcomes and explanations traceable**

<a id="v2-278-1"></a>

<a id="v2-254-2"></a>

<a id="v2-254-1"></a>

A dated record can serve several purposes without requiring a separate journal for each one. Shortform’s suggestions become more useful when the record states what is being checked:

- **Outcomes and learning.** Record the objective, expected outcome, actual outcome and date for work proposals, personal goals and community projects. Include favorable and unfavorable results. For weight management, record days the intended eating plan was followed and missed, plus possible contributing factors; for fitness, use credible exercise evidence and track progress rather than follow a trend. After a task, note what worked, what did not and what to change.
- **Motives and bias.** Before a major learning or education choice, record both personal benefits and the hoped-for contribution to knowledge. Review the pattern later. A weekly entry about topics that attract you can reveal neglected viewpoints—for example, an interest in free-market policies can prompt examination of research on regulation.
- **Plans before results.** Record development goals, predictions and evaluation criteria before acting. A monthly accountability group might compare a reading list, its expected benefit and the book summaries actually discussed. A public environmental-footprint plan can invite feedback. Commitment does not forbid justified changes: date and explain them. A decision app or a simple pro/con tool can lock the original criteria before displaying the comparison so later revisions remain visible.
- **A traceable public explanation.** A learning blog or vlog can explain coding concepts, resources and projects with links to open-source code. A portfolio can collect the work alongside data, methods and tools. Version control adds dated change notes to writing or design and allows recovery of earlier versions. Share a project summary with specific questions to invite error checking; infographics, short videos or social-media threads can reach audiences directly, but should preserve evidence and limitations.
- **Shared accountability.** A workplace “failure resume” or newsletter can record unsuccessful projects and lessons. A household or group ledger can make transactions visible to participants. A personal “No-Failure Friday” can make room for a new sport, hobby or DIY activity, with attention to learning rather than a successful result every time.

These are records and discussion practices, not proofs that personal observations identify causes or that public commitment makes a plan correct.













<!-- atlas:block-content b-746fe7a4ac1891dd:end -->
<!-- atlas:block-content b-8dfd23724ee68ded:start -->
### Practical routines for reading and doing research

The following Shortform exercises range from research procedures to informal analogies. Logs, games, badges and discussion groups can prompt attention; they are not established substitutes for training, sound design or independent review.

<a id="b-8dfd23724ee68ded"></a>
<a id="u-science-fictions-0adde11a5a9a"></a>
<a id="u-science-fictions-a91227dc3fcb"></a>
<a id="u-science-fictions-d0ccdc8963b5"></a>
<a id="u-science-fictions-0051cf3f84fd"></a>
<a id="u-science-fictions-a3317aa7ee8e"></a>

<a id="v2-285-0"></a>

**Shortform perspective: Navigating the volume of scientific literature and building research habits**

Several practical strategies help readers and researchers stay oriented amid the proliferation of scholarly papers and build more rigorous habits.

- Use a citation management tool such as Zotero or Mendeley to organize publications. Save articles as you encounter them, tag them with keywords, and build a curated personal library so you can navigate the literature without feeling overwhelmed.
- Set up keyword alerts on research databases such as Google Scholar or ResearchGate to create a personal knowledge filter. For example, someone interested in renewable energy can subscribe to alerts for terms like "solar power efficiency" or "wind turbine advancements" to receive a keyword-matched stream of new papers that still need appraisal.
- Volunteer for a citizen science project. Hands-on participation in research can help you develop a clearer sense of what constitutes meaningful scientific progress, as distinct from the mere proliferation of publications.
- Run one regular journal club rather than a new group for every exercise. Rotate selection of a high-quality paper, read its full methods and findings, discuss applications, and try explaining it without jargon before checking the design and numbers again. A monthly meeting is one possible rhythm. An interdisciplinary virtual book club serves a different purpose: books from varied fields and participants with different expertise can expose assumptions and suggest collaborations.
- Develop the habit of reading full articles rather than only abstracts or conclusions when conducting research, so that you can properly assess the context and quality of work and make informed decisions about which sources to cite.
- Before finalizing your own work, run it through a personal checklist of common errors found in scholarly papers—such as citation mistakes, logical fallacies, or data misinterpretation—and review each item systematically.
- Use the summary’s “Innovation Scorecard” to record personal interest: score novelty, applicability and clarity from 1 to 10, considering what differs from existing technology, how it might be used, and how well you understand it. These ratings assess perceived promise and communication, not research reliability. Check design, evidence and replication separately before trusting the finding.

<a id="v2-285-1"></a>

**Shortform perspective: Citation metrics, publication ethics, and predatory journals**

Incentives tied to publication and citation counts can distort research behavior, from self-citation practices to salami slicing and submission to predatory outlets. Practical guidance can help researchers navigate these pressures ethically.

- Participate in substantive professional discussion and cite work because it is relevant and used. Visibility may follow; reciprocal citation arrangements should not become the objective.
- Before submitting, verify the journal’s editorial board, affiliations, publication history, fees and review practices. A recognizable name alone is not proof of quality.
- Use publishing forums to learn about suspicious venues, then check claims rather than treating a recommendation as verification.

<a id="v2-285-2"></a>

**Shortform perspective: Automation, technology, and reducing human error in research and life**

Automated tools and software can reduce carelessness and human error in science, and analogous practices can improve personal routines and work.

- Try using a habit-tracker app to monitor and improve your daily routines. Habit trackers help you maintain consistency in activities such as exercising or practicing a language, creating a personal system that reduces the chance of forgetting tasks—much like an automated workflow does in research.
- Improve your personal finance tracking by building a custom dashboard that combines narrative and numerical analysis. Use a simple database program to log your income and expenses, then write weekly reflections on your spending habits and financial goals within the same tool. Seeing numbers and written reflections side by side makes it easier to spot patterns and make adjustments.
- Adopt a data visualization tool for dynamic reporting. Tools like Tableau or Power BI can connect directly to your data sources, so your visualizations stay up to date without manual updates. If you run a small business, for example, you can link your sales data to one of these tools to see live updates on a dashboard.

Beyond everyday applications, automated fraud-detection algorithms are increasingly used to identify manipulated data or images in scientific publications, acting as a technological check on misconduct.













<a id="b-c64f2d7fef91c22e"></a>
<a id="u-science-fictions-0efd41ae084a"></a>
<a id="u-science-fictions-f359fa8920b0"></a>

<a id="v2-287-0"></a>

**Shortform perspective: Building Intuition for Statistical Thresholds and Evidence Strength**

Shortform’s everyday exercises serve different purposes. A forecast log should record the stated chance of rain and whether rain occurred in the specified place and period—not whether carrying an umbrella kept you dry. Comparing many such forecasts with outcomes teaches calibration, not the meaning of a significance test. Constructively play devil’s advocate by asking what evidence supports a claim such as a product “revolutionizing” an industry. A discretionary waiting period, such as 24 hours, can create room to recheck a nonurgent finding. Before a career change, consulting people with different jobs, stages and experiences can expose omitted perspectives; it is not literally a larger random sample or a substitute for representative evidence.

<a id="v2-287-1"></a>

**Shortform perspective: Collaborative Resource Sharing and Cost Efficiency**

Practical tips address the high costs of academic publishing and research services by encouraging collective action. Optimize your library usage by asking your local or university library to subscribe to more cost-effective journals. Present librarians with a proposal including data on cost differences between journals and the potential savings, and emphasize that reallocating funds to more affordable options can expand the library’s offerings for all patrons. Create a shared resource with peers to track and share information on service fees, using a collaborative tool such as Google Sheets to record what different providers charge for similar services; this can help your network make informed decisions and potentially negotiate better deals using collective data. Collaborate with peers to share research costs and resources more broadly: seek out online forums, local meetups, or community groups where you can partner with others interested in similar topics. Pooling resources can make materials or services that are too expensive for one person accessible; for example, a group researching a niche topic could split the cost of a specialized journal subscription or a rare book. Simplify your research workflow by using a single multi-functional tool instead of switching between separate software for data collection, analysis, and storage. For example, a spreadsheet program that lets you enter data, use built-in statistical functions, and create visualizations all in one place reduces context switching and makes complex information easier to manage.

<a id="b-b07ea74f2a32a851"></a>
<a id="u-science-fictions-34661764f667"></a>

<a id="v2-293-0"></a>

**Shortform perspective: Shortform’s Cross-Domain Applications of “Salami Slicing”**

A Shortform practical-tips sidebar repurposes the idea of dividing work into smaller units in three deliberately different domains:

- Break a large personal goal into micro-projects to create manageable mini-goals and a stronger sense of achievement—for example, declutter one drawer or shelf per day instead of an entire room.
- Start a blog or podcast about employment inequality: share stories, interview experts, and offer actionable advice to job seekers and employers to raise awareness and encourage fairer hiring practices.
- Map a scientist’s major discoveries, publications, and collaborations on a visual career timeline to see how the work developed and connected over time, beyond raw publication counts.

<a id="b-54da1fb98924cea9"></a>
<a id="u-science-fictions-13db7cc98c98"></a>
<a id="u-science-fictions-f19fbc3e9cbd"></a>
<a id="b-03b4b94f19d7bb0a"></a>
<a id="u-science-fictions-312f9a25a786"></a>
<a id="u-science-fictions-f61948c9834a"></a>
<a id="b-16678b3ff53a7e43"></a>
<a id="u-science-fictions-73e1bfd3a3fa"></a>
<a id="u-science-fictions-aba0f74070ea"></a>

<a id="v2-279-0"></a>

**Shortform perspective: Further perspectives on misleading claims**

Shortform points to Bergstrom and West’s Calling Bullshit (2020) for misleading evidence, especially data used to persuade; Ben Goldacre’s Bad Science for distortions across study design, pharmaceutical claims and media reporting; and Charles Wheelan’s Naked Statistics for accessible explanations connecting statistical meaning with medical, policy, education and business decisions. These are further-reading routes, not additional books fully covered here.

<!-- atlas:block-content b-8dfd23724ee68ded:end -->
<a id="theme-durable-judgment"></a>

## Living and planning with uncertainty

<!-- atlas:block-content b-8e00c8a74fc2e8f1:start -->
### Plan around uncertainty rather than an exact script

Housel’s stable-pattern approach and probabilistic forecasting serve different purposes. Broad motivations can inform scenarios while new evidence changes their probabilities.

<a id="b-8e00c8a74fc2e8f1"></a>
<a id="u-same-as-ever-24360c928047"></a>
<a id="u-same-as-ever-23f7f19fe53b"></a>
<a id="u-same-as-ever-2f1bb02b09c9"></a>
<a id="u-same-as-ever-8335b4e30451"></a>
<a id="u-same-as-ever-336834ea1f96"></a>
<a id="u-same-as-ever-81b36fd33677"></a>
<a id="u-same-as-ever-0828ebd0f772"></a>
<a id="u-same-as-ever-0023ddb85980"></a>

<a id="v2-294-0"></a>

**The Unpredictability of the Future and the Limits of Foresight**

In Same as Ever, Morgan Housel emphasizes limits on predicting specific events and exact paths while asking what remains more stable: human motivations and recurring patterns. His book presents 23 lessons. The Shortform summary organizes selected ideas into why foresight fails and how to act despite uncertainty; those organizing principles below are the summary’s structure, not a claim that Housel wrote only three rules.

A core premise of Housel’s framework is that the future cannot be predicted with certainty. Even though people crave clarity about what lies ahead, Housel emphasizes that we can never know for certain what will happen. He identifies three major blind spots that limit foresight: misunderstandings about innovation, risk, and ramifications.

Unanticipated events can be especially damaging because preparations omit them; this does not mean every major risk is unforeseeable or its damage necessarily grows exponentially. Housel’s practical response is resilience: leave resources and flexibility for surprises rather than assume the plan lists every hazard. In financial planning, he illustrates this with saving beyond the central estimate of need and borrowing below maximum capacity. The amount of slack should follow plausible losses, commitments and recovery needs.

Stability can encourage complacency and rising risk-taking, which can make later shocks more damaging. That is a feedback mechanism, not an inevitable timetable from every calm period to crisis. Housel warns against treating success as permanent: maintain skills and attention to warning signs even when recent results have been good.

<a id="b-50b203b7d07b9d43"></a>
<a id="u-same-as-ever-153ea1c74835"></a>
<a id="u-same-as-ever-c10c0bd33fd3"></a>
<a id="u-same-as-ever-3dbe93fd8ce0"></a>
<a id="u-same-as-ever-c9b479bb6da6"></a>

<a id="v2-295-0"></a>

**Why the future is inherently hard to predict**

Housel points to two difficulties. Innovations often combine small advances whose eventual uses are not obvious: ARPANET, funded by the US Defense Department’s research agency in the 1960s to connect computers and share resources, became part of the history leading to the internet. It was not simply a system for managing Cold War secrets. Chance intersections also matter. The summary uses the Lusitania’s voyage and scheduling as an illustration of how small changes can alter exposure to danger. The ship was sunk in 1915, killing more than 1,000 people; the sinking influenced opinion but did not by itself cause US entry into the war in 1917. The precise counterfactual chain from a speed decision to the later political outcome is not established by the anecdote. Both cases illustrate why foresight must allow for interactions and alternative paths.

<a id="b-cbd167dccbfae1de"></a>
<a id="u-same-as-ever-8613c38b14f2"></a>
<a id="u-same-as-ever-5e46e0912632"></a>
<a id="u-same-as-ever-bdcab403e994"></a>
<a id="u-same-as-ever-4ac3877a4d59"></a>
<a id="u-same-as-ever-e863f90b5cbf"></a>
<a id="u-same-as-ever-42360cec8967"></a>
<a id="u-thinking-in-bets-b791798dcbf4"></a>
<a id="u-same-as-ever-e7265c50dc2f"></a>
<a id="u-same-as-ever-12a9b6de6dd3"></a>
<a id="b-76b4eddc7a994ff9"></a>
<a id="u-same-as-ever-a9f7c2619b12"></a>
<a id="connections"></a>

<a id="v2-296-0"></a>

**Three Principles for Navigating an Uncertain World**

Shortform organizes Housel’s guidance under three principles: understand human behavior, prepare the mind for uncertainty, and make plans more resilient. Its body provides five behavioral observations, five mindsets and five planning practices. The summary’s introductory claim of six behavioral rules is a counting error; no sixth rule appears in the supplied source.

Rather than trying to predict specific events, Housel argues that you should redirect your efforts to anticipating reactions. Keep an open mind about the specifics of what the future holds, and as events unfold, try to anticipate how people will react based on fundamental motivations such as fear and greed. Housel’s insights on human psychology can be organized into five rules about behavior to help with this anticipation.

*Sources and further reading:* [Same as Ever](https://www.penguinrandomhouse.com/books/672339/same-as-ever-by-morgan-housel/) · [The Psychology of Money](https://short.zaynjarvis.com/books/the-psychology-of-money/)

<!-- atlas:block-content b-8e00c8a74fc2e8f1:end -->
<!-- atlas:block-content b-0798112d877a9c8e:start -->
### Five recurring observations about behavior

<a id="b-0798112d877a9c8e"></a>
<a id="u-same-as-ever-c9c9dc44868e"></a>
<a id="u-same-as-ever-053a6944cd9c"></a>
<a id="u-same-as-ever-882fda39200f"></a>

<a id="v2-297-0"></a>

**Incentives as the Most Powerful Driver of Human Behavior**

Housel argues that incentives—such as the promise of profit or of belonging to a group—are the most powerful force in the world. They can drive otherwise good, honest people to justify and defend nearly anything. While financial incentives can be powerful, cultural and social incentives are often more compelling, because people fear social rejection and will take extreme actions to protect their connection to their community. Understanding incentives helps you predict what others are likely to do, and also gives you insight into your own decision-making and behavior. To examine how incentives influence you, ask which of your current views would change if your incentives were different; if you believe none would change, you are likely blinded by your current incentives. For example, you might support a work project only because it comes with a bonus, and oppose it if the bonus were removed.

<a id="b-97ded1a8b48cf9c0"></a>
<a id="u-same-as-ever-555bbead4e1a"></a>
<a id="u-same-as-ever-533571b18c2a"></a>
<a id="u-same-as-ever-66277266a684"></a>

<a id="v2-298-0"></a>

**The Power of Stories Over Facts and Logic**

Housel argues that compelling stories triumph over facts and logic in capturing attention and driving success, because they simplify complex ideas and unite diverse audiences. As a result, the most persuasive ideas are not always the most accurate or original, but those paired with the most engaging narratives. Conversely, many strong ideas and products fail to reach their potential because their proponents do not tell their stories compellingly. Housel advises using this insight to avoid blindly following attractive stories: ask whether any of your beliefs stem from good marketing rather than accuracy, or whether you might be missing important truths simply because they come from lackluster storytellers. For example, you might realize you favor a popular diet not because it is scientifically proven, but because its promoters tell more compelling stories than the advocates of more effective alternatives.

<a id="v2-304-2"></a>

**Shortform perspective: Why Stories Shape Judgment and How to Craft Compelling Narratives**

If people follow the best stories, a natural question is how to craft compelling narratives so your own ideas prevail. In *Wired for Story*, Lisa Cron argues that human brains evolved to absorb information through stories. Narratives developed as a way to keep us safe from potential dangers without our having to experience them directly, so we have evolved to crave them. Cron recommends writing in a way that satisfies the brain’s built-in expectations of stories. The basic elements of an engaging story, according to Cron, are the protagonist, the protagonist’s goal, the plot, and what the story is truly about—the underlying lesson. These are the elements our brains crave in order to anticipate what will happen next, and this anticipation lies at the heart of our evolutionary need for story.

<a id="b-f4efd1cb0fe375de"></a>
<a id="u-same-as-ever-ad0d9c51be43"></a>
<a id="u-same-as-ever-37c3edce063e"></a>

<a id="v2-301-0"></a>

**Firsthand experience, empathy, and the limits of theory**

Housel argues that firsthand experience changes your understanding and behavior in ways theoretical knowledge cannot. This is why plans often collapse when they meet the emotional reality of direct experience. To narrow the gap between how you think you or others will react in a given situation and the actual reaction, do not overestimate your future preferences and responses. He adds that major traumatic events leave lasting psychological effects that shape people’s behaviors and worldviews, and past life experiences more broadly influence current beliefs. Keeping this in mind helps you approach conflicts with greater empathy: rather than assuming others are misinformed or irrational, ask yourself what they might have experienced that leads them to think or act differently from you.

<a id="b-c424206c3fad0cac"></a>
<a id="u-same-as-ever-954a0241c618"></a>
<a id="u-same-as-ever-5b00951d7502"></a>

<a id="v2-302-0"></a>

**People are not one-dimensional**

Housel warns against expecting role models to possess only admirable traits. Persistence may shade into stubbornness and discipline into rigidity; one characteristic can have different consequences in different contexts. Seeing the whole person makes admiration more realistic. It does not require accepting harmful conduct, copying every trait, or abandoning boundaries and accountability.

<a id="v2-299-0"></a>

**People hide their struggles**

Housel’s Rule #5 holds that people hide their struggles. We tend to think others have it easier or better, but people’s public personas typically mask their struggles and difficulties; we see only the polished, filtered version of others’ lives, businesses, and careers, while being acutely aware of our own challenges. In reality, everyone deals with problems they do not disclose. For example, a successful entrepreneur’s social media may highlight achievements but rarely reveal the setbacks and failures she faced along the way. Recognizing this tendency can help you be more forgiving of yourself and others, and understand that most accomplishments come from ordinary people who persevered—not from people who never struggled.

<a id="b-253bf8622852d228"></a>
<a id="u-same-as-ever-153e2fd90dc2"></a>
<a id="u-clear-thinking-dcf3da206e19"></a>
<a id="u-same-as-ever-d2b14072cf0b"></a>

<a id="v2-316-0"></a>

**Shortform perspective: Envy, Hidden Struggles, and the Psychology of Ambition**

Envy and ambition are two powerful, often hidden forces that shape how we judge others and set goals for ourselves. People universally hide their struggles, and envy is itself one of those hidden struggles, as Robert Greene notes in *The Laws of Human Nature*. No one wants to consciously admit to envy, because doing so would mean acknowledging feelings of inferiority—admitting that someone else has something you want implies they are superior. As a result, envy rarely presents itself as envy, even to the person feeling it; instead it morphs into other negative emotions to stay hidden. People often reframe the envied person’s advantage as the product of luck, ambition, or underhandedness, concluding the person does not actually deserve it, and then feel anger or resentment at the perceived unfairness. This not only distracts from one’s own goals and progress, but in extreme cases can motivate harm toward the person triggering the envy.

Shortform brings in Cardone’s 10X Rule as an ambitious-goal perspective: larger aspirations can create excitement and persistence. Low goals do not guarantee disappointment, and multiplying a target by ten is a motivational prescription rather than a calibrated optimum. Incentive design separately asks whether benefits are salient and relevant; material financial incentives should not all be called nudges. Wheelan’s four trade-offs help examine individual motives: present versus future needs, work versus leisure, self versus others, and physical versus psychic needs. For example, habitually favoring work over leisure may explain unnecessary weekend work. The exercise is to identify the trade-off, not assume everyone responds to the same reward.

*Sources and further reading:* [Wired for Story](https://www.shortform.com/app/book/wired-for-story/1-page-summary) · [Narrative and anticipation](https://www.shortform.com/app/book/wired-for-story/1-page-summary#the-elements-of-a-story) · [Hidden envy](https://www.shortform.com/app/book/the-laws-of-human-nature/law-5) · [The 10X Rule](https://www.shortform.com/app/book/the-10x-rule) · [Fear of ambitious goals](https://www.shortform.com/app/book/the-10x-rule/chapter-1#set-10x-goals) · [Ambition and motivation](https://www.shortform.com/app/book/the-10x-rule/chapter-9) · [Influencing choices](https://www.shortform.com/app/book/nudge/techniques-3-and-4#offering-incentives) · [Naked Economics](https://short.zaynjarvis.com/books/naked-economics/)

<!-- atlas:block-content b-0798112d877a9c8e:end -->
<!-- atlas:block-content b-579c22fcc84a2eee:start -->
### Five ways to prepare the mind

<a id="v2-299-mind"></a>

**Five mindsets in the summary’s organization**

Shortform groups Housel’s advice about the mind into five practices: manage expectations; make the most of hardship while remaining wary of ease; combine optimism and pessimism; give up the demand for certainty; and understand compounding. These concern how to engage with an uncertain future, rather than predict exactly what happens when. The five-part grouping belongs to the summary, not the chapter count or complete structure of Housel’s book.

<a id="b-579c22fcc84a2eee"></a>
<a id="u-thinking-fast-and-slow-a42181ae6e98"></a>
<a id="u-thinking-fast-and-slow-327b4685882d"></a>
<a id="u-thinking-fast-and-slow-576cb4af8e60"></a>
<a id="u-thinking-fast-and-slow-8988d7ddffa6"></a>
<a id="u-thinking-fast-and-slow-df3e04d646f6"></a>
<a id="u-thinking-fast-and-slow-8eb043503c2c"></a>

<a id="v2-312-0"></a>

**Experience and remembered evaluation**

Kahneman distinguishes moment-to-moment experience from retrospective evaluation. The experiencing self is a way to discuss pleasure and pain while they occur; it is not a literal mental accountant maintaining a conscious running sum. The remembering self constructs an evaluation using selected moments and interpretation.

Two patterns can affect such evaluations: the peak-end rule gives substantial weight to an intense moment and the ending; duration neglect gives length less weight than a sum of lived experience would. A musical with a memorable song and strong finale illustrates why a retrospective review can differ from the quality of most minutes. These are context-dependent patterns, not laws saying everyone remembers only two moments or that a one-month and six-month injury must feel identical in retrospect.

<a id="b-4746da554607ecb6"></a>
<a id="u-thinking-fast-and-slow-738e80723a75"></a>
<a id="u-thinking-fast-and-slow-a1a5e4e3f840"></a>
<a id="u-thinking-fast-and-slow-10580a6e9396"></a>
<a id="u-thinking-fast-and-slow-820840c60917"></a>
<a id="u-thinking-fast-and-slow-11835295b70c"></a>

<a id="v2-313-0"></a>

**Balancing the Two Selves for Durable Well-Being**

Both lived experience and remembered meaning matter. A person might endure years of unpleasant work for an imagined gratifying retirement, or avoid an enjoyable activity because its ending may disappoint. These are trade-offs worth examining, not a reason to identify the experiencing self with impulsivity or the remembering self with all future planning. Future choices can aim to improve future daily experience as well as future memories. Ask how a plan changes the life actually lived and how it will be evaluated afterward.

<a id="b-530f8c56cb762627"></a>
<a id="u-thinking-fast-and-slow-0bcad928d227"></a>
<a id="u-thinking-fast-and-slow-dc97ab840fa4"></a>
<a id="u-thinking-fast-and-slow-fabd32357c20"></a>
<a id="u-thinking-fast-and-slow-eda164a8415a"></a>
<a id="u-thinking-fast-and-slow-047aa190a6bc"></a>
<a id="u-thinking-fast-and-slow-abc7263df570"></a>
<a id="u-thinking-fast-and-slow-83a52bcc2737"></a>
<a id="u-thinking-fast-and-slow-c4e73b2a6c85"></a>
<a id="u-thinking-fast-and-slow-e4e6237343bb"></a>

<a id="v2-20-0"></a>

**The Two Selves: How Memory Distorts Happiness and Decision-Making**

Memories guide later choices but need not represent the total experience well. The distinction above explains why a satisfying story and a satisfying daily life can diverge; it is not an identity between the remembering self and System 1. Kahneman’s practical concern is to give both forms of well-being weight.

Three applications make the distinction concrete:
- A boxer may accept difficult training and painful fights for a memorable victory. That meaning can matter, but it does not erase the lived cost or establish that every sacrifice is worthwhile.
- The focusing illusion gives whatever is currently salient excessive weight in an overall evaluation. A couple asked about their marriage during a conflict may answer differently from when attention rests on other aspects.
- People can overestimate the lasting emotional effect of a change because they imagine thinking about it constantly. Adaptation to greater wealth illustrates why expectations and attention can change; adaptation varies and need not restore everyone to an identical baseline.

<a id="b-4a36d330eb322daa"></a>
<a id="u-thinking-fast-and-slow-74f174471cdd"></a>

<a id="v2-21-0"></a>

**Shortform perspective: Adaptation is uneven**

The hedonic-treadmill idea describes a possible cycle: pursue an improvement, enjoy it, adapt, and pursue the next one. It helps explain why forecasts exaggerate how constantly a change will occupy attention. Adaptation varies across people and events; serious negative experiences can have enduring effects, and neither positive nor negative changes necessarily return everyone to a fixed emotional baseline.

<a id="b-b8d198c59745f962"></a>
<a id="u-thinking-fast-and-slow-f9373431fd08"></a>

<a id="v2-22-0"></a>

**Shortform perspective: Memory, Trauma, and the Body’s Record of Experience**

Shortform adds that a past event can affect present well-being through consequences even when its details are not currently recalled. An old injury, for example, can still hurt. Trauma-related symptoms and memory can be complex, but unexplained tension or bodily symptoms do not by themselves establish a forgotten traumatic event. The relevant distinction is between remembered evaluation and ongoing consequences, not a method for inferring hidden memories from symptoms.

<a id="b-3742d92675f866a9"></a>
<a id="u-thinking-fast-and-slow-eb5dc1d86d17"></a>
<a id="u-thinking-fast-and-slow-f728bad19ac0"></a>

<a id="v2-336-0"></a>

**Shortform perspective: Memory, the Peak-End Rule, and Duration’s Indirect Effect**

Shortform applies peak-end thinking to relationships: memorable moments can disproportionately shape an overall impression. Deliberately sharing varied experiences—trying a new place rather than always repeating the same outing—can add material to a relationship’s memory. More novelty does not automatically make the overall memory accurate or repair a pattern of harmful daily experience.

Shortform cites a 2020 study suggesting duration can affect memory indirectly by changing how salient moments are recalled. Its injury comparison is an explanatory hypothetical, not a reported experiment proving that six months always feels worse than one month. Longer recovery might alter fatigue, frustration or adaptation, and those changes could alter recollection in different directions. This qualifies an overly universal reading of duration neglect; it does not show that remembered evaluation simply equals total elapsed time.

<a id="b-e24ad1804c7644b5"></a>
<a id="u-same-as-ever-a0e282fb8ab6"></a>
<a id="u-same-as-ever-792b7718348c"></a>
<a id="u-thinking-fast-and-slow-ab93b96e124b"></a>
<a id="u-same-as-ever-62813e21f13f"></a>

<a id="v2-314-0"></a>

**Expectations, happiness, and the two selves**

Housel emphasizes how rising expectations can absorb improvements in circumstances. The direction matters: falling short of what one expects differs from life exceeding expectations; happiness is not simply the absolute distance between two quantities. Managing expectations alongside circumstances may help someone appreciate steady career progress rather than demand instant success. Kahneman’s career example adds a separate question: would a prestigious, well-paid job impose many unhappy days for the sake of a future life story or retirement? Compare those daily costs with the remembered and anticipated benefits, rather than letting one measure of success stand for well-being.

<a id="b-927b8788ea02bcd6"></a>
<a id="u-thinking-fast-and-slow-98f932c6c5c9"></a>
<a id="u-same-as-ever-c918319d782d"></a>
<a id="u-thinking-fast-and-slow-386c1de4440f"></a>

<a id="v2-315-0"></a>

**Shortform perspective: Balancing Present Contentment and Future Orientation**

Kahneman’s framework of the two selves connects to a broader set of ideas about balancing present-moment experience and future-oriented striving. Kahneman’s advice to balance the needs of the experiencing and remembering selves echoes what Daniel Z. Lieberman and Michael E. Long argue in *The Molecule of More*: lasting happiness requires balancing excitement about future possibilities with contentment with your present circumstances. The pursuit of “more”—money, possessions, extreme experiences—delivers a pleasurable dopamine rush but keeps you locked on future rewards; present-focused brain activity is needed to enjoy what you already have. Lieberman and Long suggest the simplest path to balance is a career or hobby that demands full present attention while also offering future goals to work toward, such as painting, where each brushstroke absorbs you and the finished piece provides a forward-looking milestone.

People tend to focus heavily on the remembering self and neglect the experiencing self. One way to become more attuned to the experiencing self is to practice mindfulness meditation, which trains you to accept each moment-to-moment experience as it arises and then let it fade naturally—you do not judge experiences as “good” or “bad,” and therefore you do not try to consciously hold onto good ones or block out bad ones.

The focusing illusion, another concept from *Thinking, Fast and Slow*, helps explain why expectations so often outrun reality. When judging overall life satisfaction, people typically substitute an easier question—such as “what is my mood right now?”—and fixate on salient events or recurring concerns, giving rise to the illusion that “nothing in life is as important as you think it is when you are thinking about it.” This illusion causes you to overestimate how much a change like a promotion will affect your future happiness, because you overestimate how much you will still be thinking about it once you have adapted. To narrow the gap between circumstances and expectations, three antidotes help: build a weighted life rubric covering the factors you value (such as health and relationships) to evaluate satisfaction more broadly and avoid fixating on single sore points; seek out absorbing, pleasurable activities like social interaction or team sports that hold your focus, rather than passive activities like television; and reflect on past experience to identify what you are actually grateful your past self did—then do more of that.

<a id="b-ae7a1b628a318116"></a>
<a id="u-same-as-ever-492b2f3b8d14"></a>
<a id="u-same-as-ever-69b70f63db58"></a>
<a id="u-same-as-ever-9bdf32a9db75"></a>

<a id="v2-307-0"></a>

**Hardship, crisis, and the cost of worthwhile achievement**

Housel’s practical theme is to accept the unavoidable costs of a worthwhile goal—uncertainty, effort, delay and some discomfort—rather than demand a frictionless path. A marathon illustrates training effort and setbacks making achievement meaningful; injury and exhaustion are reasons to reassess, not proof that training is valuable. Crisis can sometimes focus attention, accelerate collaboration and loosen bureaucratic barriers, but innovation also occurs in prosperity, and severe stress can destroy the capacity to create. Distinguish productive difficulty from avoidable harm, and change or abandon a method when its costs no longer serve the goal.

<a id="b-263caec57300ea50"></a>
<a id="u-same-as-ever-6a1425af15a7"></a>

<a id="v2-308-0"></a>

**Shortform perspective: Stoic Strategies for Turning Hardship Into Progress**

An editorial sidebar connects views of hardship to Stoic thought via Ryan Holiday’s *The Obstacle Is the Way*. Holiday agrees that life constantly tests us with obstacles, and our choice is to give up or to turn the obstacle upside down and use it to advance our purpose. Framing obstacles in an empowering way is a central tenet of Stoicism, an ancient Greek philosophy founded in the third century BCE. Three Stoic strategies help make the most of hardship: 1) Learn from crises—rather than being distracted by fear, treat crises as learning opportunities, observe them dispassionately, and learn from others’ mistakes to avoid repeating them. 2) Focus on what you can change—distinguish between what is within your control and what is not, and direct your effort toward the former; you cannot control the economy, weather, disasters, or other people’s feelings and actions, but you can control your attitude, emotions, judgments, and responses. 3) Turn obstacles upside down—every situation or obstacle contains a hidden opportunity you can use to your benefit, if you learn to flip your perspective; for example, being dumped by a long-term partner can be reframed as a chance to find someone new with whom you are even more compatible.

<a id="b-ed6fa6e5663c4e96"></a>
<a id="u-same-as-ever-3b109a311a18"></a>
<a id="u-same-as-ever-62d339412500"></a>

<a id="v2-309-0"></a>

**Optimistic Pessimism as a Balanced Future Mindset**

Morgan Housel argues that optimism and pessimism work better together than in isolation, just as hardship and progress are intertwined. Pessimism helps us prepare for risks, while optimism sustains hope for the future even when evidence is uncertain. An optimistic pessimist recognizes that life is full of unavoidable, constant setbacks that do not rule out eventual progress. For example, when learning a new skill such as playing an instrument, expecting setbacks and mistakes (the pessimistic side) helps you prepare and stay patient, while maintaining hope that you will improve over time (the optimistic side) keeps you motivated to keep practicing.

<a id="b-85b7eec967825159"></a>
<a id="u-thinking-in-bets-6272f8ccf9cf"></a>
<a id="u-same-as-ever-6819badf5f1c"></a>

<a id="v2-310-0"></a>

**The Stockdale Test: Hope Without a Deadline**

Jim Stockdale’s captivity shows what Housel’s balanced stance demands when the horizon is unknown. The Vietnam War prisoner of war and admiral embodied what Housel calls optimistic pessimism. During captivity, Stockdale knew he would have to endure torture for many months or even years, yet he held unwavering faith that he would eventually be released. After he was freed, when asked who had the hardest time in the camps, he replied that it was the optimists: they naively believed they would be freed soon, and their spirits were crushed with each passing day that did not bring liberation. Stockdale’s approach balanced absolute faith that things would eventually improve (optimism) with acceptance of the long, difficult nature of his situation (pessimism). A similar long-term orientation appears in the mental time travel framework from *Thinking in Bets*, which encourages you to keep long-term goals in mind even when deciding on matters whose benefits or consequences are not immediately obvious—such as relocating or changing careers.

<a id="b-db2b0913229618aa"></a>
<a id="u-same-as-ever-0e2cc033128a"></a>

<a id="v2-311-0"></a>

**Shortform perspective: The Stockdale Paradox: Balancing Realism and Hope in Leadership**

Admiral Stockdale’s balanced approach to hardship has inspired other authors, including Jim Collins in *Good to Great*. Collins agrees that Stockdale endured because he could confront and manage the brutal reality of daily camp life without losing hope, and he argues that this psychological duality is a defining trait of great business leaders. He calls it the “Stockdale Paradox”: stoic acceptance of business realities paired with unwavering faith in eventual triumph. To cultivate a leadership style that balances an unvarnished pursuit of facts and good-faith efforts to address them, Collins offers four guiding principles:
- Run meetings focused on dialogue and inquiry: ask your team for information and insights, and pose tough questions to surface assumptions. This deepens understanding and helps you spot risks outside your direct line of sight.
- Foster an atmosphere where dissent and debate actually shape decisions, so people feel encouraged to speak up when they disagree.
- When setbacks occur, conduct detailed post-mortems without finger-pointing to prevent the same mistake from happening twice.
- Prepare for unavoidable hardships by building alarm bell mechanisms: procedures that let team members escalate urgent concerns directly to decision-makers.

<a id="b-242872a252717454"></a>
<a id="u-same-as-ever-8840f46b2de2"></a>
<a id="u-same-as-ever-555e6f20fd75"></a>
<a id="u-same-as-ever-c5f2a24e9bb3"></a>

<a id="v2-300-0"></a>

**The Illusion of Certainty and the Value of Probabilistic Thinking**

Morgan Housel argues that humans value certainty over accuracy. In pursuit of certainty about what will happen, we delude ourselves with definitive but inaccurate predictions. He recommends abandoning the demand for certainty, recognizing it as an illusion, and practicing probabilistic thinking instead. To illustrate, imagine a group asking two experts whether a natural disaster will occur in the coming year: Expert 1 declares there definitely will be one, while Expert 2 estimates a 45% chance. Most people will follow Expert 1 because of the definitive tone, even though Expert 2’s estimate is likely more accurate.

<a id="b-18eaeafc0f48cafe"></a>
<a id="u-same-as-ever-6ddea7eb3806"></a>
<a id="u-same-as-ever-b2db6f232e87"></a>

<a id="v2-306-0"></a>

**Compounding, Impatience, and Long-Term Progress**

One strategy for future-proofing your mind is to understand how compounding works. Seemingly small changes, accumulated over time, create massive outcomes, yet humans often lack the patience to wait for them. Housel writes that progress is gradual and often invisible, while setbacks are sudden and command attention; grasping the tension between compounding dynamics and natural human impatience helps people stay focused on long-term goals.

*Sources and further reading:* [Hedonic adaptation](https://thedecisionlab.com/reference-guide/psychology/hedonic-treadmill) · [Lasting injury effects](https://www.txfoot.com/post/could-i-have-an-ankle-sprain-without-knowing-it) · [Trauma and memory](https://www.kvc.org/blog/children-may-not-remember-trauma-but-their-bodies-do/) · [Relationship memories](https://www.psychologytoday.com/us/blog/social-instincts/202510/3-ways-to-beat-the-peak-end-rule-in-your-relationship) · [Duration and recollection](https://www.sciencedirect.com/science/article/abs/pii/S0010027721001694) · [The Molecule of More](https://www.shortform.com/app/book/the-molecule-of-more/1-page-summary) · [Present and future satisfaction](https://www.shortform.com/app/book/the-molecule-of-more/1-page-summary#happiness-requires-both-what-if-and-what-is-chemicals) · [Dopamine](https://my.clevelandclinic.org/health/articles/22581-dopamine) · [Mindfulness meditation](https://www.shortform.com/app/book/the-master-guides-choosing-a-meditation-practice/1-page-summary#styles-of-buddhist-meditation) · [The Obstacle Is the Way](https://www.shortform.com/app/book/the-obstacle-is-the-way/1-page-summary) · [Reframing obstacles](https://www.shortform.com/app/book/the-obstacle-is-the-way/part-1) · [Good to Great](https://www.shortform.com/app/book/good-to-great/1-page-summary) · [Stockdale paradox](https://www.shortform.com/app/book/good-to-great/chapter-4#the-stockdale-paradox)

<!-- atlas:block-content b-579c22fcc84a2eee:end -->
<!-- atlas:block-content b-528fa7247b2b6eba:start -->
### Five ways to make a plan more resilient

<a id="b-528fa7247b2b6eba"></a>
<a id="u-same-as-ever-65aaa7703de1"></a>
<a id="u-same-as-ever-3e3a9d9149a9"></a>
<a id="u-same-as-ever-abc5a46db7a1"></a>

<a id="v2-317-0"></a>

**Future-proofing plans: five strategies against four vulnerabilities**

Future-proofing plans means limiting four vulnerabilities: size, complexity, efficiency, and complacency. Housel lays out five strategies to guard against these risks: think long term, keep it simple, keep it less than perfect, keep it manageable, and don’t let down your guard.

<a id="v2-299-long"></a>

**Think long term, with review and exit conditions**

Long horizons can let learning, relationships and returns accumulate, but only if the activity remains viable. Housel stresses patience through short-term difficulty and flexibility about timing. In a new business, slow customer growth may call for experimentation and an adjusted strategy rather than immediate abandonment. Persistence is not a guarantee: monitor demand, resources and predetermined review or exit thresholds. A long-term goal can survive a change of method, and evidence can justify ending a failing plan.

<a id="b-3cbca35c98e0a89c"></a>
<a id="u-same-as-ever-585f038a386c"></a>
<a id="u-same-as-ever-7dbe77a7f8f1"></a>
<a id="u-same-as-ever-bbcaf041995d"></a>

<a id="v2-318-0"></a>

**Keep it simple: core drivers over comforting complexity**

The “keep it simple” strategy rests on Housel’s observation that people often reach for complex approaches when simpler ones would work better. Complexity feels comforting, he argues, because it gives a sense of control and signals that effort was invested. Yet in most fields, he believes, outcomes are driven by a small number of core principles rather than endless details. To avoid being drawn into complexity that does not improve results, start by identifying the few variables that drive most outcomes in a given situation, then adopt simple, effective solutions built on those proven drivers. For example, a small business owner might focus on strong customer relationships and cash flow management—two core factors—rather than getting lost in overly detailed marketing strategies or complicated financial models.

<a id="v2-304-0"></a>

**Shortform perspective: Why Modern Environments Overload Our Decision-Making and Fuel Complexity Bias**

Shortform connects Housel’s simplicity advice with the conditions for learning: feedback that is delayed, inconsistent or hard to interpret can teach the wrong lesson. A workplace where managers rarely give clear feedback illustrates the problem. Its contrast with supposedly simple hunter-gatherer life is a broad evolutionary interpretation, not a demonstrated division between two uniformly different worlds. Barry Schwartz’s *The Paradox of Choice* adds a distinct burden: abundant market choices can consume time and attention and sometimes leave people overwhelmed. Check feedback quality and choice demands in the actual setting.

This same dynamic helps explain the appeal of complexity itself. While Housel frames complexity as comforting because it creates a sense of control and effort, Shortform notes it may also be comforting because it releases people from effort. Citing Shane Parrish, author of *The Great Mental Models*, Shortform explains that complexity bias—the tendency to treat simple things as complicated—functions as a mental shortcut: labeling something as too complex excuses us from trying to understand it. Parrish recommends using Occam's razor—a preference for fewer unsupported assumptions when explanations account for the evidence comparably well—not a rule that the shortest explanation is true—as a way to sidestep this bias.

<a id="b-48d7ed88a7a65be8"></a>
<a id="u-same-as-ever-e1401e6ecf3a"></a>
<a id="u-same-as-ever-942f2ac93247"></a>
<a id="u-same-as-ever-eb21bd821076"></a>

<a id="v2-319-0"></a>

**Keep it less than perfect: deliberate slack over extreme efficiency**

The “keep it less than perfect” strategy warns against extreme efficiency. Housel argues that pursuing flawless efficiency often creates unexpected vulnerabilities and undermines adaptability. He uses the metaphor of a perfectly tuned machine in which every cog fits with zero slack: it may seem ideal, but a single jammed or broken cog will seize the whole system. A machine built with a little flexibility—some slack between its parts—can absorb shocks and keep running even when one component fails. The same logic applies to personal systems: building deliberate inefficiency into your schedule creates room for adaptability and resilience. Resisting the cultural pressure to be constantly productive and making time for unstructured thinking is not wasted effort; it is essential for solving complex problems and generating creative solutions. For example, regularly setting aside agenda-free time to read and reflect can pay off in unexpected ways.

<a id="b-0b3a4bdd91175d6a"></a>
<a id="u-same-as-ever-d5812b2deef9"></a>

<a id="v2-320-0"></a>

**Shortform perspective: Flexibility Over Perfect Efficiency: How “Wasted” Time Can Pay Off**

Shortform uses Steve Jobs’s college experience to illustrate flexibility rather than perfect efficiency. After leaving formal study, he lived frugally and attended subjects that interested him, including calligraphy. The connection to the Macintosh’s typography emerged roughly a decade later—from his early-1970s college period to the 1984 Mac—not several decades later. The case shows a possible value of exploration; a successful retrospective example does not mean every apparently impractical activity will pay off.

<a id="v2-331-1"></a>

**Compounding, sustainable growth, and natural limits**

Two ideas from Housel reinforce the value of gradual, sustainable progress over forced speed. First, to leverage compounding, he advises prioritizing projects that are sustainable over time and maintaining perspective during challenging periods. As an example, consistently dedicating a small amount of time each day to learning a skill—even when progress feels slow—can yield remarkable improvement over the years through steady, compounding effort. Second, his “keep it manageable” strategy holds that pushing good ideas beyond their natural limits often turns them into failures. There is an optimal scale and pace of growth, beyond which things break down. To keep good ideas manageable, let them grow at their natural pace and to their natural limit, rather than using artificial methods to force rapid, exponential expansion. For example, someone starting a fitness routine would do better to gradually increase workouts over months than to jump immediately into intense daily sessions; this prevents injury and burnout and makes progress more sustainable. The mental time travel framework from *Thinking in Bets* complements this by encouraging you to evaluate present choices against their long-term, cumulative effects rather than only their immediate payoffs.

<a id="v2-331-2"></a>

**Continuous learning, mistakes, and avoiding complacency**

Sustained good judgment depends on continuous learning and adaptation rather than resting on past outcomes. In *Clear Thinking*, Parrish frames this as Tip #3: Learn From Every Mistake—treating every mistake as a learning opportunity. High achievers distinguish themselves by taking responsibility for their irrationality and errors instead of pretending they do not happen, and by working to avoid repeating mistakes. In *Same as Ever*, Housel makes a parallel argument from the direction of success under Strategy #5: Don’t Let Your Guard Down. No success is permanent enough to justify complacency; people often let their guard down after achieving some success, assuming it will last, which leads them to stop refining their skills and become less alert to potential problems. To avoid crises born of complacency, you should never rest on past achievements, but continue learning and evolving to keep pace with changing times and secure new successes—for example, if you stop updating your skills after a promotion, you may be unprepared when industry standards shift. The mental time travel approach from *Thinking in Bets* reinforces this orientation by making active learning from past outcomes a routine part of present decisions, rather than waiting for failure to force reflection.

<a id="b-2d2056c1d1133d7a"></a>
<a id="u-same-as-ever-2fa55206e46d"></a>

<a id="v2-321-0"></a>

**Shortform perspective: Complacency and Inflation: Why Good Times Can Sour Without Vigilance**

Shortform draws an analogy from Charles Wheelan’s Naked Economics: if demand expands faster than an economy’s capacity to supply goods and services, prices can rise. Easy credit can contribute in some conditions; supply shocks, expectations and policy also matter. Inflation is a sustained rise in the general price level, not necessarily a rapid rise in nearly every price. The example in which $100 buys four shirts one month and only two the next depicts an extreme price increase, not ordinary inflation’s definition. Complacency can delay a response, but is not a general single cause of inflation or crisis.

*Sources and further reading:* [Learning-environment feedback](https://www.psychologytoday.com/us/blog/tech-happy-life/202009/whats-the-root-cause-many-the-worlds-problems#:~:text=The%20legacy%20of%20our%20hunter-gatherer,and%20technological%20evolution.%20The) · [Choice overload](https://www.shortform.com/app/book/the-paradox-of-choice/part-2#how-expanded-choice-complicates-decision-making) · [Complexity bias](https://fs.blog/complexity-bias/) · [Jobs’s college experience](https://news.stanford.edu/stories/2005/06/youve-got-find-love-jobs-says)

<!-- atlas:block-content b-528fa7247b2b6eba:end -->
<!-- atlas:block-content b-92dd3d8fac9bcce8:start -->
### Leave room for an unexpected connection

<a id="b-92dd3d8fac9bcce8"></a>
<a id="u-same-as-ever-61d8f44d9659"></a>

<a id="v2-341-0"></a>

**Shortform perspective: Innovation can come from an unexpected collaborator**

Hedy Lamarr, an Austrian-born Hollywood actor, worked with composer George Antheil on a frequency-hopping communication system for radio-controlled torpedoes. The National Inventors Hall of Fame describes her earlier marriage to a munitions manufacturer as a source of weapons-system knowledge. The technical mechanism was synchronized changes of radio frequency at transmitter and receiver, reducing opportunities for detection or jamming while retaining communication. Their patent was issued in 1942.

This was guidance communication, not a remote device for locating submarines. Lamarr and Antheil’s work is recognized as an important development in wireless communication; that is more precise than saying one patent directly became Wi-Fi and Bluetooth. The story supports looking beyond conventional professional categories for collaborators and ideas. It does not establish an unverified personal motive or a single inevitable path from wartime invention to modern networks. [National Inventors Hall of Fame](https://www.invent.org/inductees/hedy-lamarr).

*Sources and further reading:* [Lamarr’s early marriage](https://www.womenshistory.org/education-resources/biographies/hedy-lamarr) · [Frequency hopping](https://science.howstuffworks.com/innovation/famous-inventors/hedy-lamarr.htm) · [Unexpected inventors](https://harbingersmagazine.com/articles/how-five-overlooked-female-inventors-changed-our-lives-for-the-better-by-maria-mitko/)

<!-- atlas:block-content b-92dd3d8fac9bcce8:end -->