Hindsight Bias

Definition

Hindsight bias is the tendency, once an outcome is known, to believe it was predictable and that you knew it all along. It makes the past look more certain than it was, and quietly rewrites what people remember having predicted.

In plain terms

In Baruch Fischhoff’s 1975 experiments, people told how an event turned out judged it to have been far more probable than people who were not told, and they overestimated what they themselves would have predicted. Its effects are unfair blame (“how could they not have seen it?”), overconfidence about the future, and poor learning, because a surprise that is remembered as expected teaches nothing.

Remedies are practical: write predictions down with probabilities before the outcome; date them; keep them unaltered; and, when reviewing, list the outcomes that could have happened as well as the one that did. The data equivalent is Look-ahead Bias: judging the past with information that only exists today.

Example

A forecast made on Monday: “60% chance the rating count rises more than 10% this week.” It rises 14%.

Without the note, Friday’s memory has quietly turned a 60% guess into near-certainty and made the forecaster look more skilled than the record supports.

Common misunderstandings

• “It is arrogance.” It is a memory and judgment effect that affects careful people too.
• “Knowing about the bias protects you.” Awareness helps little; written, dated predictions help a lot.
• “Only novices are affected.” Experts show it, often more strongly, in their own field.