Book summary “How to Decide”: summary

Introduction

You often judge your choices by how they turn out in the end. When a gamble pays off, you assume you made a brilliant choice. When misfortune strikes, you blame your own poor judgment. This habit binds your evaluation to luck rather than to the actual quality of your thinking.

You can radically improve your life by separating decision quality from luck and using structured probabilistic tools to navigate uncertainty.

Annie Duke presents a practical framework to upgrade your decision-making process. You will learn to recognize outcome bias, map alternative futures, quantify uncertainty, and leverage outside perspectives. The book also provides mental tools for anticipating failure and executing long-term goals.

• How do you stop judging decisions by their results?
• What are the hidden costs of certainty?
• How can you translate gut feelings into precise probabilities?

1. Resulting and the Shadow of Luck

Life outcomes are determined by only two factors: decision quality and luck. While luck is entirely out of your control, decision quality can be improved. However, human tendency is to rely on simple shortcuts like resulting, making it difficult to accurately evaluate whether a decision was actually good or bad.

People mistakenly judge a decision entirely by how it turns out.

Annie Duke calls this tendency resulting or outcome bias. Because you cannot easily see decision quality after the fact, you rely on outcome quality as a proxy. This distorts learning, causing you to repeat bad decisions that happened to work out and abandon good decisions that faced bad luck.

An individual quits a job to take a new position. If it leads to a promotion, it is labeled a great decision. If it leads to a layoff, it is called a bad decision, despite identical initial information.

Outcome Bias

Resulting is a mental shortcut where the quality of an outcome is used to determine the quality of a decision. This compromises your ability to learn from experience.

The Luck Matrix

The interaction between decision quality and outcome quality creates four distinct categories: Earned Reward, Bad Luck, Dumb Luck, and Just Deserts. Good decisions can lead to bad outcomes due to bad luck, and poor decisions can lead to good outcomes due to dumb luck. Resulting causes you to ignore the shadow of luck and draw false conclusions about your choices.

A driver runs a red light and gets through unscathed. This is categorized as dumb luck rather than proof that running red lights is a sound strategy.

A structured checklist helps evaluate whether resulting is clouding your judgment by actively examining alternative outcomes and uncontrollable factors:

• How much is the outcome clouding your judgment about the quality of the decision?
• Even if bad decisions preceded a bad outcome, can you identify some good decisions made along the way?
• Even if good decisions preceded a good outcome, can you identify some ways the decision could have been better?
• What are the factors outside the control of the decision-maker, including the actions of other people?
• What are the other ways things could have turned out?

2. The Decision Multiverse

Exploring counterfactuals and mapping the tree of possible outcomes prevents hindsight bias.

Once an outcome is revealed, post-outcome knowledge creeps into the memory of what was known beforehand. This leads to false retrospective certainty that a result was inevitable. Our brains lack computer file timestamps. Once an outcome is known, the new information reflexively creeps into our before-the-fact knowledge box, creating hindsight bias and causing undue self-blame or judgment of others.

United Artists, Universal, and Disney famously passed on Star Wars, and subsequent commentators labeled these passes as colossal blunders in hindsight. The original Star Wars film cost $11 million to make. It generated a total box office of $775 million, reaching $10.3 billion in worldwide box office across twelve films as of early 2020. Disney paid $4 billion to buy the franchise rights in 2012.

Decision Forestry

When the future unfolds, our minds use a cognitive chain saw to hack away all unrealized possibilities, making the past feel inevitable. Ahead of us, possibilities look like a tree of branches of varying thicknesses. Once an outcome happens, we forget the other branches on the ground, falsely believing the actual result was the only possible path.

Consider the 2016 US presidential election where Hillary Clinton lost Michigan, Pennsylvania, and Wisconsin by tiny margins, leading to widespread post-election pundit critique of campaign strategy despite pre-election polls showing her ahead. Out of 14,000,000 votes across those three key states, the margin was only 80,000 votes. Pre-election polls projected an electoral victory of 278 to 260, while Donald Trump achieved an actual win of 306 to 232.

To counteract this distortion, use a Knowledge Tracker:

• Stuff you knew before the decision: The sum of your knowledge and beliefs at the time.
• Stuff you know after the outcome: Including new information that revealed itself after the future unfolded.
• Take the time to deliberately reconstruct what was known before versus what was revealed after.

Journaling the information known during the decision-making process acts as a vaccine against hindsight bias. Because it is difficult to accurately recall previous knowledge once an outcome is known, writing down key facts provides a concrete reference point that establishes a clear time stamp and prevents memory distortion. Follow these practices to protect your judgment:

• Journal the stuff you knew before the decision while you are in the process of making it.
• Refer back to your journal record to see what was reasonably knowable at the time.

To rebuild the decision tree, take these steps:

• Identify the decision.
• Identify the actual outcome.
• Create a tree with other reasonable outcomes that were possible at the time of the decision.
• Explore the other possible outcomes to understand better what is to be learned.

The Three Ps

A rigorous six-step process evaluates future decisions and past choices by mapping reasonable outcomes, preferences, and probabilities. Outcomes are only informative when unexpected, regardless of whether they are awesome or terrible. A repeatable process prevents the haze of resulting and hindsight bias.

All guesses are educated guesses, and decision-makers must adopt the archer's mindset rather than viewing estimates as right or wrong. People often avoid probability estimates by claiming they are merely guessing. However, some knowledge always exists, allowing us to narrow ranges and score points like an archer taking aim at a target rather than playing blindfolded pin the tail on the donkey.

Imagine guessing the weight of a bison blocking traffic in Yellowstone National Park. You can safely avoid absurd thresholds like 100 pounds or 10,000 pounds, leaving a realistic narrowed range of 800 to 3,500 pounds for a bison's weight.

Improve your estimates by following these practices:

• Figure out what you already know that will make your guess more educated.
• Figure out what you can find out that will make your guess more educated.

Apply the complete analytical framework using these six steps:

• Step 1—Identify the reasonable set of possible outcomes.
• Step 2—Identify your preference using the payoff for each outcome—to what degree do you like or dislike each outcome, given your values?
• Step 3—Estimate the likelihood of each outcome unfolding.
• Step 4—Assess the relative likelihood of outcomes you like and dislike for the option under consideration.
• Step 5—Repeat Steps 1–4 for other options under consideration.
• Step 6—Compare the options to one another.

3. Taking Dead Aim

Converting vague natural language terms into precise percentages and bounded ranges uncovers hidden biases.

Natural language terms like "real possibility" or "fair chance" are ambiguous and have vastly different meanings to different people. This ambiguity allows biases to hide and prevents corrective feedback. Converting them into percentages based on surveys brings clarity.

President Kennedy approved the Bay of Pigs invasion because his military advisers told him the plan had a fair chance of success. The advisers meant 25 percent while Kennedy thought it meant something much higher.

Precision and Ranges

You can improve your estimates by removing ambiguity. Take these steps to refine your guesses:

• Take the Mauboussins' survey on the meaning of common terms for probability.
• If you are uncomfortable making specific estimates, use the term that comes to mind and convert it to a specific bull's-eye estimate by referring to your answers from the survey.
• In addition, make a target estimate, comprising a range of the reasonable upper and lower bounds.
• Test the reasonableness of your upper and lower bounds with the shock test.
• Ask yourself, What information could I discover that would make me change my mind?
• If the information is available, go find it.
• If it isn't, keep an eye out for it in the future.

Setting a range around a bull's-eye estimate using the shock test helps signal uncertainty and prevents overconfidence. A wide or narrow range reflects how much knowledge you possess. The shock test requires setting the narrowest range where you would be pretty shocked if the correct answer fell outside it, aiming for roughly a 90 percent success rate across estimates.

When tested on trivia like Meryl Streep's 21 Academy Award nominations or Prince's age of 57 at his death on April 21, 2016, people typically miss the correct answer because they severely oversell their knowledge.

You can apply this to your own trees. Use these practices to test your ranges:

• Take one of the decision trees you've already developed in this book and give your bull's-eye estimate for the likelihood of each potential outcome.
• Include a lower and upper bound around each bull's-eye estimate using the shock test.
• Ask yourself what information you could seek out that would help narrow the widest ranges or check for overconfidence on narrow ranges.

The Inside View

The inside view distorts decision-making through personal bias and identity protection, whereas the outside view looks at what is objectively true of the world. People naturally evaluate situations through their own unique experiences and beliefs, leading to cognitive biases like confirmation bias, availability bias, and overconfidence. Stepping outside one's perspective allows an objective assessment, much like how it is easier to solve someone else's problems than your own.

A friend repeatedly dates jerks and blames bad luck or a curse, completely blind to their own pattern of behavior, whereas an outside observer easily spots the issue.

Examine your own thinking with these reflection steps:

• Identify a situation where a friend, family member, or coworker was caught in the inside view.
• Reflect on past situations where you got caught in the inside view and how it affected your decisions.

4. Turning Decisions Outside In

People consistently overestimate their relative skill levels compared to the general population due to the inside view. Individuals live inside their own experience without access to a full population poll, leading them to assume they belong to the upper tiers of ability in almost everything.

Base rates and outside perspectives discipline the overconfidence of the inside view.

Over 90% of professors rate themselves as better-than-average teachers, 90% of Americans rate their driving ability as better than average, and 1% of students think their social skills are below average.

Base Rates

Base rates provide an objective statistical center of gravity that disciplines the overconfidence of the inside view. By discovering the statistical likelihood of outcomes in situations similar to yours, you establish an outside view baseline that keeps your forecasts grounded.

The divorce rate for first marriages in America is between 40% and 50%, 25% of adult American deaths are caused by heart disease, and 60% of new restaurants fail within the first year. Other relevant baselines include the fact that 8% of the American population lives in a city with a population over 1 million, and 63.1% of high school graduates will attend college without taking any time off. Additionally, 82% of gym members go once a week or less, according to a survey cited by Zachary Crockett, with 77% of gym memberships going to the gym once a week or less going completely unused.

Perspective tracking systematically forces a decision-maker to record both the outside view and the inside view to improve future feedback loops. Writing down external realities and internal perceptions separately prevents memory creep, reduces self-deception regarding luck versus skill, and creates accountability.

Examine your own situation using these steps:

• Describe your situation entirely from the outside view, including applicable base rates and perspectives provided by other people.
• Describe your situation entirely from the inside view.
• Find the intersection of the outside view and the inside view to land on a more accurate narrative.

High Stakes and Fast Decisions

The Happiness Test identifies low-impact decisions by assessing whether an outcome will significantly affect long-term happiness. If a decision outcome does not alter your happiness over a week, month, or year, it is low impact, allowing you to trade accuracy for speed and save time for major choices.

Evaluate your choices with these steps:

• Ask yourself if the outcome of your decision, good or bad, will likely have a significant effect on your happiness in a year.
• Repeat for a month and a week.

Freerolling involves decisions with high asymmetric upside and limited downside where you cannot end up worse off than before. Because potential losses are insignificant, you can speed up decision-making without fearing the standard penalty of sacrificing accuracy. A man is offered a trivia question where a correct answer earns ten dollars and a wrong answer costs ten dollars or nothing, illustrating how a zero-downside scenario permits immediate action.

When a decision is difficult due to closely matched options, it is actually easy because the difference in potential payoffs is negligible. People agonize over close calls because they focus exclusively on the downside, treating a sheep in wolf's clothing as a dangerous beast when both choices offer nearly identical utility. Choosing between a vacation in Paris and a vacation in Rome creates immense anxiety, whereas choosing between Paris and a trout cannery is effortless, driven by a mere 0.2% difference in potential happiness increase between the two choices in a hypothetical scenario.

Effective decision-making requires heavy investment in sorting options by personal values, followed by rapid picking among valid choices. The main value of decision-making time lies in sorting the world into things you like and dislike, because once options pass your criteria, spending extra time picking yields virtually no added accuracy.

Apply this approach with these exercises:

• For the next week, practice applying the Only-Option Test whenever you are at a restaurant.
• Look through the menu and figure out which items you would be happy with if they were your only option.
• After sorting the menu in this way, decide among the options that pass the Only-Option Test by flipping a coin.

5. The Power of Negative Thinking

Combining mental contrasting, premortems, and backcasting prepares you for roadblocks and goal execution.

Positive visualization mimics success and feels good, but it fails to prepare you for roadblocks. Imagining how you might fail acts like a navigation app to reveal obstacles and prepare you to react when bad luck strikes.

Program participants trying to lose fifty pounds who imagined ways they might fail lost twenty-six pounds more on average than those using positive visualization alone. Twentytree percent of New Year's resolutions are abandoned within a week, and ninety-two percent of people never achieve their goal.

Premortems

A premortem has you imagine a future failure to list internal and external reasons for it, breaking team groupthink. A backcasting exercise works backward from a successful outcome, and combining both builds a complete map. Combining mental time travel and mental contrasting produces thirty percent more reasons for why something might fail.

A medical device company planning to sell a device overseas identified through a premortem that new regulatory bans had a high probability of enactment.

Execute this process with these steps:

• Identify the goal or specific decision you are considering.
• Figure out a reasonable time period for the goal or decision to play out.
• Imagine it is the day after that period and you failed, listing up to five reasons due to your decisions and actions.
• List up to five reasons for failure due to things outside your control.
• Repeat independently for team members if doing this as a group, followed by discussion.

The Dr. Evil Game

An imagined evil genius makes you fail through small, easily justifiable exceptions that add up over time. Recognizing these categories allows you to replace reflexive choices with fixed rules or category decisions. Because willpower alone is insufficient, you can physically block poor decisions or raise and lower barriers based on your premortem insights.

Making occasional exceptions to eat junk food when stressed or working late slowly accumulates to destroy a nutritional goal. Ulysses tied himself to the mast and filled his crew's ears with beeswax to resist the Sirens, or kept an alarm clock across the room. Thirty-seven percent of college students transfer to a new school, and nearly half of those transfer multiple times.

Apply this approach with these exercises:

• Imagine a positive goal.
• Imagine Dr. Evil controls your brain to make decisions guaranteeing failure while avoiding detection.
• Write down decisions that have rationales good enough not to be noticed in isolation.
• Define category decisions as a precommitment to eliminate repeated one-time exceptions.
• Using your Decision Exploration Table, list up to three precommitments to physically prevent actions or raise/lower barriers for decisions.

Conclusion

You can radically improve your life by separating decision quality from luck and using structured probabilistic tools to navigate uncertainty.

The core concepts of decision-making form an integrated system for navigating an unpredictable world. Resulting clouds your learning by letting outcomes dictate your judgment, but you can correct this distortion by examining the hidden branches of the decision multiverse. Replacing rigid certainty with precise numerical ranges exposes your true level of knowledge and grounds your thinking in reality. Seeking outside views and historical base rates shatters the dangerous illusion that your current situation is entirely unique. Finally, mental contrasting, premortems, and backcasting steel your plans against foreseeable failures before you commit resources.

To put these ideas into practice, apply the following rules to your daily choices:

1. Never judge the quality of a past decision solely by whether it turned out well or poorly.
2. Map out alternative futures and counterfactual histories to neutralize hindsight bias and build genuine self-compassion.
3. Translate vague verbal beliefs into exact percentages and bounded ranges to force intellectual honesty.
4. Actively seek out outside perspectives and base rates to counteract the overconfidence of the inside view.
5. Anticipate obstacles in advance using premortems and backcasting to protect your goals from unforeseen shocks.

10 Key Ideas

1. People mistakenly judge a decision entirely by how it turns out, leading to outcome bias and distorted learning.
2. The interaction between decision quality and outcome quality creates four distinct categories that reveal the shadow of luck.
3. Exploring counterfactuals and mapping the tree of possible outcomes prevents hindsight bias from rewriting memory.
4. Converting vague natural language terms into precise percentages and bounded ranges uncovers hidden biases.
5. Base rates and outside perspectives discipline the overconfidence of the inside view.
6. The Happiness Test and freerolling allow you to trade accuracy for speed on low-stakes and zero-downside decisions.
7. Sorting options by personal values allows you to pick rapidly among valid choices without losing accuracy.
8. Combining mental contrasting, premortems, and backcasting prepares you for roadblocks and goal execution.
9. An imagined evil genius makes you fail through small, easily justifiable exceptions that add up over time.
10. You can radically improve your life by separating decision quality from luck and using structured probabilistic tools to navigate uncertainty.