Log probability estimates before the outcome

Write down your confidence level and reasoning before you know how things turn out.

Why it works

Hindsight bias works by overwriting the memory of your prior uncertainty with the knowledge of the outcome. Writing the estimate down at decision time creates an external record that the bias cannot rewrite — you have objective evidence of what you actually believed, not what you now feel you must have believed.

How to do it

1. Before any significant decision or prediction, write one sentence on what you expect and assign a rough probability (e.g. "70% likely to succeed").
2. Note the top two reasons behind the estimate.
3. After the outcome, compare what you wrote with what you now remember believing — the gap is your hindsight bias in action.
4. Use the gap to calibrate, not to shame yourself.

Evidence

Pre-outcome recording is the most reliable counter to hindsight bias identified in the research literature, because it bypasses memory reconstruction entirely by externalizing the prior belief. Fischhoff’s foundational experiments established that outcome knowledge systematically distorts recalled judgment, and Roese and Vohs’ review confirms the effect is one of the most robust and persistent in the field. (observational)

Logging works to expose the bias after the fact; whether it reduces the bias in future predictions is less established. Calibration improves with practice, but hindsight bias is remarkably persistent even in people who know about it.

Sources

• Fischhoff (1975), "Hindsight ≠ Foresight: The Effect of Outcome Knowledge on Judgment Under Uncertainty", Journal of Experimental Psychology: Human Perception and Performance
• Fischhoff, B. (1975). Hindsight ≠ foresight: The effect of outcome knowledge on judgment under uncertainty. Journal of Experimental Psychology: Human Perception and Performance, 1(3), 288–299.
• Roese, N. J., & Vohs, K. D. (2012). Hindsight bias. Perspectives on Psychological Science, 7(5), 411–426.

Common mistake

Writing the estimate vaguely ("I thought it might work") so the record is just as malleable as memory. Specificity — a number and reasons — is what makes it useful.