Thinking, Fast and Slow

The model

Daniel Kahneman was a psychologist who received the Nobel Memorial Prize in Economic Sciences in 2002 for work done mostly with Amos Tversky, who had died in 1996. Thinking, Fast and Slow, published in 2011, was his synthesis of that work and of the research it inspired, written for a general reader.

The book’s organising image is two characters in the mind. Kahneman borrows the terms System 1 and System 2 from the psychologists Keith Stanovich and Richard West, and is careful to say they are useful fictions rather than parts of the brain. System 1 operates automatically and quickly, with little effort and no sense of voluntary control. It recognises faces, reads emotions, completes familiar phrases and produces impressions, intuitions and feelings. System 2 allocates attention to effortful mental activity, such as complex calculations, careful comparison and self-control. We identify with System 2, the conscious reasoning self, but most of the time it simply endorses what System 1 suggests.

System 1 is usually right. It is fast, efficient and finely tuned to familiar environments. But it relies on shortcuts, or heuristics, that produce predictable errors, biases, in particular situations, and System 2 often fails to correct them because effort is costly and System 1’s answers feel convincing.

The book then sets out the main patterns. People judge probability by how easily examples come to mind, the availability heuristic. They judge whether something belongs to a category by how closely it resembles a stereotype, ignoring base rates, the representativeness heuristic. Numerical estimates are pulled toward any number that happens to be present, even an irrelevant one, which is anchoring. When faced with a hard question, people often answer an easier one without noticing, which Kahneman calls substitution. And the mind builds the most coherent story it can from whatever information is available, without registering what is missing, which he summarises with the acronym WYSIATI: what you see is all there is.

Seeing with it

The model’s power lies in how widely it applies. It explains overconfidence: a coherent story feels true, and the less we know, the easier it is to make the story coherent. It explains the planning fallacy, the tendency to underestimate the time and cost of projects by focusing on the specific case and ignoring how similar projects have turned out. Kahneman’s remedy, the outside view, asks what happened to a reference class of comparable cases. The site’s idea page The Inside View and the Outside View traces it.

The book’s fourth part explains prospect theory, the account of decision-making under risk that Kahneman and Tversky published in 1979. People evaluate outcomes as gains and losses relative to a reference point, not as final states of wealth. Losses loom larger than gains of the same size, roughly twice as large in many experiments. And people overweight small probabilities and underweight large ones. These patterns explain why people buy both lottery tickets and insurance, why they hold losing investments too long and why framing the same choice as a gain or a loss can reverse preferences. The idea page Loss Aversion and the Pull of the Status Quo follows it into products and policy.

The last part distinguishes two selves: the experiencing self, which lives through each moment, and the remembering self, which keeps score. People judge past experiences mostly by their peak and their end, and pay little attention to duration. In one study, patients undergoing a painful medical procedure remembered it as less bad when it ended with a few minutes of milder discomfort, even though the total pain was greater. Decisions are made by the remembering self, which can choose experiences the experiencing self would not.

Where the model stops

Kahneman was candid about the limits of the two-system picture, which he presented as a simplification. Researchers continue to debate whether thinking really divides into two systems or rather lies along a continuum.

Gerd Gigerenzer and colleagues have long argued that heuristics are often good adaptations rather than flaws, and that many biases shrink when problems are presented in more natural formats, such as frequencies instead of probabilities. Kahneman engaged with this debate over many years. Gary Klein’s research on expert intuition, which Kahneman discusses in the book, shows that in environments with regular feedback, experts’ fast judgements can be highly accurate. The two men co-wrote a paper in 2009 setting out when intuition can be trusted.

The model also says little about how to fix the problems it describes. Kahneman himself was doubtful that individuals can learn to overcome their own biases, and placed more hope in organisations designing procedures that catch errors.

Since it was published

The replication crisis in psychology reached parts of the book. Its chapter on priming, describing experiments in which subtle cues seemed to change behaviour substantially, rested on studies that later failed to replicate or were found to be underpowered. When a 2017 analysis by Ulrich Schimmack and colleagues assessed the studies cited in the chapter, Kahneman responded publicly that he accepted its basic conclusions and had placed too much faith in underpowered studies. Research on ego depletion, which the book discusses, has also faced replication failures. The core findings on heuristics, biases and prospect theory have proved far more robust. Kahneman died in March 2024. His final book, Noise (2021), written with Olivier Sibony and Cass Sunstein, turned to a different problem: the unwanted variability in judgements that should be identical.

Questions it lets you ask

Is this an answer to the question we were asked, or to an easier one? What information are we not seeing, and what would the story look like if we had it? What number did we hear first, and how far has it pulled our estimate? What happened to other projects like this one? Are we framing the choice as a loss or a gain, and would we decide differently the other way round? And is this a situation where intuition has been trained by regular feedback, or one where it only feels confident?