Outcome Bias Resistance¶

Why It Matters¶

Hindsight bias makes past uncertainty invisible. After the outcome is known, the correct path seems obvious, and you judge past decisions as if that certainty existed. This creates perverse incentives: punishing good decisions that had bad outcomes, rewarding bad decisions that got lucky. Outcome bias resistance corrects the systematic error of conflating decision quality with result quality.

The Key Move¶

When evaluating a past decision (yours or others'), explicitly reconstruct: "What information was available at decision time? What were the probabilities then?" Judge the decision based on that state of knowledge, not on the outcome. Ask: "Given what was known then, was this a reasonable probability estimate and decision threshold?" A 70% probability of success that fails is still a good decision.

Classic Application¶

Medical morbidity and mortality conferences. A patient died - was it a bad outcome from a good decision or a bad decision? Outcome bias says "patient died, therefore something was wrong." Outcome bias resistance says "given the patient's presentation, symptoms, and available tests at the time, was the probability assessment reasonable and the chosen intervention appropriate?" Sometimes the answer is yes - a 5% risk event occurred, but the 95% bet was still correct.

Surprising Application¶

Venture capital and startup investing. A startup failed - was it a bad investment? Outcome bias judges by the failure. Outcome bias resistance asks: "Given the information available at investment time - market size, team quality, technology risk - was the probability estimate reasonable?" A startup with 20% success odds that fails is not necessarily a bad investment if it was part of a portfolio strategy accepting that hit rate.

Failure Modes¶

Becomes an excuse for negligence: "the decision was fine, we just got unlucky" when the decision actually was poor. The tool requires honest reconstruction of decision-time knowledge, which is difficult and subject to motivated reasoning. It fails when you can't actually reconstruct what was known (inadequate documentation, distant past). Also fails when applied to absolve accountability - sometimes bad outcomes indicate bad process even if individual decisions seemed reasonable.

Go Deeper¶

Baron, J. & Hershey, J.C. (1988). "Outcome Bias in Decision Evaluation." Journal of Personality and Social Psychology, 54(4), 569-579; Kahneman, D. (2011). Thinking, Fast and Slow, particularly Part 3 on hindsight and outcome bias.