Hindsight Bias: Meaning, Examples and the Research Behind It

By the IQ Revealed Team · Updated November 3, 2026

Hindsight bias is the tendency to see an event as more predictable after you know the outcome than it really was beforehand. Once we know how an election, a match, an investment or a medical case turned out, the result starts to feel obvious, and we remember our earlier expectations as closer to it than they were. It is also called the knew-it-all-along effect or creeping determinism, and it is one of the most studied cognitive biases.

Hindsight bias at a glance

What happens: knowing an outcome makes it seem, in retrospect, more likely and more foreseeable than it was.
Classic studies: Baruch Fischhoff, 1975, and Fischhoff and Ruth Beyth's study of predictions about President Nixon's 1972 trips, published the same year.
Other names: knew-it-all-along effect, creeping determinism, being wise after the event.
How strong: moderate on average (d ≈ 0.39 across 95 studies), larger with quiz-style questions than with real-world events.
Who shows it: people of every age tested, from 3 to 95, and experts including doctors, judges and investment bankers.

The original experiments

In 1975 Baruch Fischhoff, then at the Oregon Research Institute, gave people short accounts of historical events and asked how likely each of several possible outcomes had been. Some were told which outcome had actually happened. Those who were told rated that outcome as more likely than those who were not, and they did not realise how much the information had changed their judgement. His paper's title summed it up: hindsight is not equal to foresight.

In the same year, Fischhoff and Ruth Beyth, at the Hebrew University of Jerusalem, used a real event. Before President Nixon's 1972 visits to China and the Soviet Union, students estimated the chances of various things happening on the trips. After the visits, they were asked to recall their own estimates. They tended to remember having given higher probabilities to the events that had happened, and lower ones to events that had not. The paper was titled “I knew it would happen”.

Three kinds of hindsight bias

Later work showed that “I knew it all along” mixes several distinct errors. In a 2012 review, Neal Roese and Kathleen Vohs described three, which can occur in any combination:

Memory distortion: “I said it would happen.” Your remembered prediction drifts towards the outcome.
Inevitability: “It had to happen.” The outcome seems to have been objectively likely all along.
Foreseeability: “I knew it would happen.” You believe you personally could have predicted it.

Hartmut Blank and colleagues argued in 2008 that these components are partly independent: a person can show one without the others, so a single “hindsight score” can hide very different things.

Why it happens

Knowledge updating. Ulrich Hoffrage, Ralph Hertwig and Gerd Gigerenzer at the Max Planck Institute for Human Development proposed in 2000 that new information updates what we know, and we then rebuild our old judgement from the updated knowledge. Their model accounted for about 75% of cases of hindsight bias or its reverse, and helping people retrieve what they knew before reduced the bias. On this view it is a by-product of a useful process, not a simple failure.
Sense-making. Roese and Vohs list selective recall of information that fits the outcome, and building a story that makes it seem to follow naturally.
Ease of understanding. An outcome that is easy to explain afterwards can be mistaken for one that was easy to predict.
Motivation. People want the world to seem orderly and predictable, and they want to avoid being blamed for failing to see problems coming.

The memory side links hindsight bias to false memory: in both, what we learn later reshapes what we remember.

How strong is the effect?

A meta-analysis. In 2004 Rebecca Guilbault and colleagues pooled 95 studies and 252 effect sizes and found a mean effect of d = 0.39 (95% interval 0.36 to 0.42). Effects were larger for almanac questions (“how long is the Nile?”) than for real-world events or case histories.
A replication. A preregistered 2025 replication of Fischhoff's second experiment with 431 participants in Brazil, posted as a preprint by Leonardo Seda and colleagues at the University of São Paulo, found no reliable difference between the hindsight and foresight groups and only a small effect overall (d = 0.12).
Across the lifespan. Daniel Bernstein and colleagues tested 194 people aged 3 to 95 in 2011. Every age group showed the bias; preschoolers and older adults showed more than older children and younger adults, for different reasons.

The honest summary: hindsight bias is real and widespread, but its size depends on the task, and some classic demonstrations are smaller in modern replications.

Hindsight bias in medicine, law and money

Medicine. Hal Arkes and colleagues demonstrated it in 1981 among physicians weighing the likelihood of diagnoses. Knowing a patient's outcome can make an earlier decision look obviously right or obviously wrong.
Negligence. In Kamin and Rachlinski's 1995 flood study, three quarters of people judging in foresight decided a flood was too unlikely to justify extra precautions; most people judging the same case after a flood found that decision negligent. An instruction designed to reduce the bias made no significant difference.
Judges and lawyers. In two studies by researchers at Leiden and Utrecht, with 1,729 participants in total, published in 2021, legal professionals judged a director's conduct more harshly and saw a bankruptcy as more foreseeable when they knew it had happened.
Investing. Bruno Biais and Martin Weber found in 2009 that among 85 investment bankers in London and Frankfurt, those with more hindsight bias had lower performance, and that the bias led students to underestimate volatility.

Reverse hindsight bias

Surprising outcomes can push the other way. Bas Verplanken and Rik Pieters had 212 people in the Netherlands estimate the chance of many casualties from nuclear power in their country two months before the Chernobyl disaster in 1986, and again five months after it. In hindsight, people gave lower probabilities than they actually had beforehand, as if the disaster had been even more unthinkable than they had believed. People high in need for cognition did not show this reverse effect.

Does a high IQ protect you?

Not reliably. In a 2007 review of individual differences, Jochen Musch and Thomas Wagner listed intelligence, field dependence and concern with self-presentation as the traits most strongly associated with the size of the bias. They also warned that hindsight measures are sometimes unreliable and that many studies are too small to be sure. What is clear is that no group is immune: the bias shows up in children, older adults and trained professionals alike. Our pages on critical thinking and metacognition look at which thinking habits help with errors like this one.

Related biases

Confirmation bias works forwards, favouring evidence for what you already believe; hindsight bias works backwards, reshaping what you think you believed.
Outcome bias judges a decision by its result rather than by what was known at the time. It often travels with hindsight bias.
The sunk cost fallacy lets the past steer the next decision; hindsight bias distorts how the past itself is remembered.

How to reduce hindsight bias

Write predictions down. Record what you expect, with a date and your reasons, before the outcome is known. Then compare against the record, not your memory.
Consider the alternatives. Roese and Vohs report that encouraging people to think of other causal explanations for an outcome can reduce the bias.
Judge decisions by what was knowable. Ask what information the decision maker had at the time, not what is known now.
Do not rely on warnings alone. In the 2004 meta-analysis, manipulations meant to reduce the bias did not produce smaller effects on average, and the debiasing instruction in the flood study failed. In ten preregistered experiments published in 2023, Markus Kneer and Izabela Skoczeń found that some strategies helped and others did not.

Trust and scope notes

This page summarises published research for general information, with figures taken from the studies listed below. It is not legal, medical or financial advice. The IQ Revealed test is an online reasoning test of 40 original items for adults aged 18 and over; it does not measure hindsight bias or any other thinking bias. Our guide to how IQ tests work explains what an online test can and cannot tell you.

IQ Revealed is not affiliated with, endorsed by or connected to the researchers, universities, institutes, journals or publishers named on this page. Names are used only to identify the research being described.

Frequently asked questions

What is hindsight bias?＋

Hindsight bias is the tendency to see an event as more predictable after you know how it turned out than it really was beforehand. It is often called the knew-it-all-along effect. Researchers describe three parts that can occur together: misremembering your earlier prediction as closer to the outcome, believing the outcome was objectively more likely, and believing you personally could have foreseen it.

What is an example of hindsight bias?＋

Before President Nixon's 1972 trips to China and the Soviet Union, Baruch Fischhoff and Ruth Beyth asked students to estimate the chances of various possible outcomes. Afterwards, asked to recall their own estimates, the students tended to remember having given higher probabilities to the events that had happened. Everyday versions include saying "I knew it" after a match, an election or a stock-market fall, or judging a doctor's decision more harshly once you know the patient's outcome.

Who discovered hindsight bias?＋

The psychologist Baruch Fischhoff is usually credited. In 1975 he published "Hindsight ≠ foresight", showing that people told the outcome of a historical event judged that outcome more likely than people who were not told, and did not realise how much the information had influenced them. In the same year he and Ruth Beyth published the Nixon study, titled "I knew it would happen".

What is the difference between hindsight bias and outcome bias?＋

Hindsight bias is about predictability: after an outcome, it seems to have been more foreseeable than it was. Outcome bias is about evaluation: a decision is judged better or worse depending on how it turned out rather than on what was known when it was made. The two often travel together, for example when a bad outcome makes a risk look obvious and the person who took it look negligent.

How strong is hindsight bias?＋

Clear but moderate on average. A 2004 meta-analysis by Rebecca Guilbault and colleagues pooled 95 studies and 252 effect sizes and found a mean effect of about d = 0.39. Effects were larger with almanac-style knowledge questions than with real-world events or case histories. A preregistered 2025 replication of Fischhoff's 1975 experiment with 431 participants in Brazil found only a small effect, so the size depends heavily on the task.

Do smarter people show less hindsight bias?＋

The evidence is thin. A 2007 review by Jochen Musch and Thomas Wagner listed intelligence, along with field dependence and self-presentation concerns, among the individual traits most strongly associated with the size of the bias, but it also warned that hindsight measures are sometimes unreliable and many studies are too small. Everyone shows the effect: a study of 194 people aged 3 to 95 found it at every age.

Why is hindsight bias a problem?＋

It makes the past look simpler than it was, so people learn the wrong lessons and grow overconfident about predicting the future. It also leads to unfair blame. In a 1995 study by Kamin and Rachlinski, three quarters of people judging in foresight decided a flood was too unlikely to justify extra precautions, yet most people judging the same case after a flood found the decision negligent.

How can you reduce hindsight bias?＋

Write predictions down, with dates and reasons, before the outcome is known, so you can compare against a record rather than your memory. Before judging a past decision, ask what else could have happened and why. Judge decisions by what was knowable at the time. Simply warning people about the bias does not reliably work: the 2004 meta-analysis found that manipulations meant to reduce it did not produce smaller effects on average.