Judging a decision's quality by how it turned out rather than by whether the reasoning was sound at the time.
Imagine you and your friend both studied the same amount for a test. Your friend guessed on the last question and got lucky — got it right. You guessed too but got it wrong. Now everyone says your friend is smarter and you made a bad choice, even though you both did the exact same thing. That's outcome bias: judging the choice by what happened after, not by whether it was a smart choice to begin with.
Outcome bias leads people to conflate the quality of a decision with the quality of its result, even when the result was largely determined by chance or factors beyond the decision-maker's control. A surgeon who made a statistically sound call to operate will be praised if the patient lives and condemned if the patient dies, despite the decision process being identical in both cases. This bias persists even when evaluators are explicitly told they have all the same information the decision-maker had and even when they acknowledge that outcomes should not influence their ratings. The effect is particularly insidious because it punishes well-reasoned risk-taking that happens to fail and rewards reckless choices that happen to succeed, thereby distorting the feedback loops people rely on to improve their judgment over time.
The same glitch looks different depending on the terrain. Finance, medicine, a relationship, a team — same mechanism, different costume.
Investors and fund managers are routinely evaluated based on portfolio returns rather than the soundness of their investment theses. A trader who made a highly leveraged, undiversified bet that happened to pay off is celebrated as skilled, while a disciplined risk manager whose diversified portfolio underperformed due to a black swan event is penalized — leading to systematic reward of reckless strategies and punishment of prudent ones.
Physicians and surgeons are frequently judged by patient outcomes rather than by whether their clinical reasoning was appropriate given the available evidence. A surgeon who opts for a procedure with an 80% success rate is evaluated differently depending on whether the patient lives or dies, despite the decision being identical. This can lead to defensive medicine, where doctors avoid risky-but-appropriate interventions to protect their reputation.
Jonathan Baron and John C. Hershey, 1988. Formalized in their paper 'Outcome Bias in Decision Evaluation' published in the Journal of Personality and Social Psychology (Vol. 54, pp. 569–579).
In ancestral environments, outcomes were often the only reliable feedback signal. If eating a berry made you sick, avoiding that berry was adaptive regardless of the underlying probability. Environments were stable enough that outcome-based learning ('that worked, do it again; that failed, avoid it') was a fast, low-cost heuristic that generally tracked real causal patterns. The brain did not evolve to handle domains where outcomes are dominated by randomness or delayed feedback.
Machine learning models trained on outcome data rather than decision-process quality can systematically learn to reward lucky strategies. Recommendation systems and reinforcement learning agents may over-fit to positive outcomes caused by random environmental factors rather than genuinely superior actions. In hiring algorithms, training on 'successful employee' outcomes can encode outcome bias by attributing success to the hiring decision rather than to post-hire environmental factors like team composition or market conditions.
This page taught you the name. The deck turns the name into reflex. 1,100+ swipeable scenarios, 1,100+ defenses, 650+ detection prompts — spaced-repetition Swipe Deck, unlimited Spot-the-Bias Quiz, Defense Playbook, Pre-Flight, My Blindspots, Cheat Sheets, Field Guide e-book. $39.53$59.
