Thinking in Bets

How it works

Annie Duke, professional poker player turned decision strategist, argues that most people evaluate decisions by their outcomes. Good outcome? Good decision. Bad outcome? Bad decision. But this confuses two distinct things.

A decision made with 90% odds of success will still fail 10% of the time. That 10% failure doesn’t make it a bad decision — it was a good bet that happened to lose. Conversely, a reckless decision with 10% odds of success will occasionally pay off. That rare win doesn’t make it a good decision.

The practice: evaluate decisions at the time they were made, with the information available at that time, not by the outcome that happened to occur. This separates “resulting” (judging by results) from genuine decision quality assessment.

Case study: How Annie Duke won $4 million by separating decisions from outcomes

Professional poker player Annie Duke was playing in the 2004 Tournament of Champions when she faced a critical decision with a medium-strength hand. The mathematically correct play was to fold — the expected value of continuing was slightly negative. But she noticed tells from her opponent that shifted the probability calculation, and she called.

She lost the hand. By outcome-focused thinking, it was a bad decision. By process-focused thinking, it was a good decision — she correctly incorporated new information to update her probability estimate, and the expected value with the updated information was positive.

Duke went on to win the tournament and $2 million. Her key insight: judge decisions by the quality of the process, not the quality of the outcome. A good decision can produce a bad result. A bad decision can produce a good result. Over time, good process beats good luck.

Real-world examples

Poker. Going all-in with pocket aces is a good decision. If you lose to a lucky river card, it was still a good decision — the odds were heavily in your favour. Evaluating the decision by the outcome would teach you the wrong lesson: “don’t go all-in with aces.”

Hiring. You hired the candidate with the strongest profile, best references, and strongest interview. They didn’t work out. Was it a bad hiring decision? Not necessarily — you made the best bet with available information. Judging the decision by the outcome would make you second-guess good processes.

Business strategy. A company invests in a reasonable market expansion that fails due to an unpredictable pandemic. The strategy wasn’t wrong — the outcome was driven by an unforeseeable event.

When to use it

Apply “thinking in bets” whenever you’re tempted to judge a decision (yours or someone else’s) purely by its outcome. Use it to create better feedback loops: after every significant decision, record your reasoning and your probability estimate, then compare against reality over many decisions. Your calibration across many bets is what matters, not individual outcomes.

Common mistakes

The main mistake is using “it was a good bet that lost” as an excuse for genuinely poor decisions. The question isn’t just “did the outcome go my way?” — it’s “given what I knew, did I make the best bet available?” Sometimes the answer is no, and the outcome happening to be bad is also correct feedback. The key is to evaluate both the decision process and the outcome, independently.

Try it now

Think of a recent decision that turned out badly. Separate the decision from the outcome: given what you knew at the time, was it a reasonable bet? Now think of a recent decision that turned out well. Same question: was the outcome due to good decision-making or good luck?

Related models

These models complement Thinking in Bets — they address similar situations from different angles.