Hindsight bias is the tendency, once an outcome is known, to see it as having been more predictable than it actually was beforehand — and, more specifically, to misremember one's own past predictions as having been more accurate than they really were. This isn't simple forgetfulness; it's an active reconstruction where the past gets quietly edited to match what's now known to be true.
The practical cost of hindsight bias is that it makes past decisions look worse (or better) than the information available at the time actually justified, punishing good decisions that happened to produce bad outcomes and rewarding bad decisions that happened to get lucky — which corrodes the ability to fairly evaluate whether a decision-making process was actually sound.
Before President Nixon's diplomatic trips to China and the Soviet Union, psychologist Baruch Fischhoff and colleagues asked people to estimate the probability of various possible outcomes. After the trips concluded, the same people were asked to recall their original estimates — and they consistently misremembered themselves as having predicted the actual outcome with far more confidence than they'd really shown at the time. Kahneman applies the same critique to bestselling business books like Built to Last, which compared 'excellent' and 'average' companies and attributed the difference to specific leadership traits; when researchers later tracked the same 'excellent' companies' performance, it regressed sharply toward the average, showing much of the original comparison had mistaken luck for skill and then reframed it, after the fact, into a tidy causal story.