Outcome Bias

Category: Decision Making

You judge the quality of a decision by how it turned out, not by whether it was smart when it was made.

How it works

A good decision is a bet made under uncertainty, and any bet can lose even when the odds were with you. Outcome bias is what happens when the result leaks backward into your evaluation of the process, so a decision that was correct at the moment it was made gets graded as stupid because a low-probability bad thing happened. The engine underneath is partly hindsight bias (once you know the ending, the whole story feels like it was always heading there) and partly the plain cognitive difficulty of holding "good process" and "bad result" in your head at the same time. Baron and Hershey showed the effect survives even when people explicitly agree that outcomes shouldn't matter, which means it is not a reasoning error you can talk yourself out of just by wanting to. You know the outcome, and knowing it rewrites how competent the decision looks.

Where you'll see it

  • **The Baron and Hershey (1988) surgery vignette.** Subjects read about a surgeon who correctly weighed a risky operation against the odds. When told the patient survived, they rated the decision as good. Told the identical decision led to the patient's death, they rated it worse and the surgeon as less competent. Nothing changed but the ending.
  • **Annie Duke, the World Series of Poker, and 'resulting.'** Duke built her book *Thinking in Bets* around the poker term for outcome bias: 'resulting,' judging a decision by whether the hand won. A great fold can lose you a pot you would have taken; a reckless all-in can hit a miracle card. Amateurs grade the pot, pros grade the play.
  • **Pete Carroll's goal-line pass, Super Bowl XLIX (2015).** With the Seahawks a yard from winning, Carroll called a pass instead of handing off to Marshawn Lynch. Russell Wilson's throw to Ricardo Lockette was intercepted by Malcolm Butler, and Carroll was branded the author of the worst play call in Super Bowl history. Interceptions on second-and-goal passes from the one had run around 2 percent league-wide, so the call was defensible. Because it failed, almost no one evaluated it as anything but idiocy.
  • **Gino, Moore, and Bazerman (2009), 'No harm, no foul.'** Across six studies, the same ethically questionable behavior was judged more unethical and more deserving of punishment when it produced a bad outcome than a good one, even when the outcome was explicitly determined by chance. Your moral judgment, not just your competence judgment, gets hijacked by the result.

Where it comes from

Named and demonstrated by Jonathan Baron and John C. Hershey in 1988 in "Outcome Bias in Decision Evaluation" (Journal of Personality and Social Psychology, 54(4), 569-579). Across five experiments they gave subjects descriptions of medical decisions and monetary gambles made under uncertainty, then varied only the outcome. Subjects rated the identical decision as better thought out, and the decision-maker as more competent, when the outcome happened to be favorable, and the effect persisted even among subjects who said outcomes should be irrelevant. Francesca Gino, Don A. Moore, and Max H. Bazerman later (2009, HBS Working Paper 08-080) extended it from competence judgments into ethical and moral judgments.

How to counter it

Freeze the information set. Before you grade any decision, ask literally: "What did they know, and what could they reasonably have known, at the moment they chose?" If a fact only became available after the choice, it is not allowed in the evaluation. This is the one move that actually neutralizes the bias, because it forces you back to the pre-outcome world.

Keep a decision journal. Write down the reasoning and the odds before you learn the result. When you review it later, you are comparing your logged reasoning against your remembered reasoning, which strips out the hindsight rewrite. Ray Dalio and Annie Duke both push this for exactly this reason.

Separate the two verdicts out loud. Say "good process, bad outcome" or "bad process, good outcome" as an explicit sentence. Naming the four-box grid (good or bad decision by good or bad outcome) breaks the reflex to collapse them into one score, and it stops you from firing the analyst who was right and promoting the one who got lucky.

The tell

You catch yourself using the word "obviously" about something you only know because it already happened, or you feel a surge of certainty that a call was dumb right after you learn how it ended. If your opinion of the decision flipped the instant you learned the result, that flip was the bias, not new insight.

Related biases

References

  1. Baron, J., & Hershey, J. C. (1988). Outcome bias in decision evaluation. Journal of Personality and Social Psychology, 54(4), 569-579
  2. Gino, F., Moore, D. A., & Bazerman, M. H. (2009). No harm, no foul: The outcome bias in ethical judgments. Harvard Business School Working Paper 08-080
  3. Aiyer, S., Kam, H. C., Ng, K. Y., Young, N. A., Shi, J., & Feldman, G. (2023). Outcomes affect evaluations of decision quality: Replication and extensions of Baron and Hershey's (1988) outcome bias experiment 1. International Review of Social Psychology, 36(1), 12
  4. Duke, A. (2018). Thinking in Bets: Making Smarter Decisions When You Don't Have All the Facts. Portfolio / Penguin
  5. Baron, J. (2008). Thinking and Deciding (4th ed.). Cambridge University Press