Annie Duke's 'Thinking in Bets' reveals why good decisions can have bad outcomes

In Thinking in Bets, former poker champion Annie Duke argues that we confuse decision quality with outcome quality. A doctor who correctly diagnoses a 95% survival-rate condition still loses 5% of patients. That doesn't make the diagnosis wrong.

This "resulting" bias — judging decisions by their outcomes — poisons everything from hiring to investing. We fire the manager whose well-reasoned bet didn't pan out and promote the one who got lucky. Over time, this selects for risk-aversion, not good judgment.

Duke's fix: think in probabilities, not certainties. Before any decision, assign confidence levels. "I'm 70% sure this feature will increase retention." Now you can evaluate the decision process separate from the result. If your 70% bets hit 70% of the time, your calibration is perfect — even though 30% "failed."