By page two of the postmortem, the failure looks so obvious that you start quietly wondering about the people who missed it.

Nobody missed it. You’re reading the story backward. In 1975, Baruch Fischhoff showed that once we know an outcome, we lose the ability to recover our earlier uncertainty. He called it creeping determinism: knowing the ending quietly rewrites the beginning. Safety researcher Sidney Dekker turned that finding into a rule for reviews: judge a decision by what people knew, saw, and were rewarded for at the time, not by what the outcome later revealed.

Reviews that skip the rule teach the wrong lesson twice. The fix targets the story (add an approval step) instead of the conditions. And the people involved learn that hindsight will be used against them, so the next review receives information already shaped for defense. Each cycle, the write-ups get cleaner and the organization learns less, while its confidence in its own foresight grows.

Months apart, the same decision often reaches me twice: first as a close call between reasonable options, later as an obvious mistake. Nothing changed but the memory.

In your next review, add one question before the timeline: what made the chosen path reasonable on the day it was chosen? Answer it seriously and the fixes change.

Think of the last failure your team dissected. Would the decision have looked wrong on the day it was made?