Thinking in Bets
Annie Duke
Thinking in Bets is Annie Duke's argument, from twenty years of professional poker, that every decision is a bet on an uncertain future and should be judged as one. Its central error is resulting, judging a decision by its outcome, which teaches you to repeat lucky mistakes and abandon right choices. Its tools are separating decision quality from outcome quality, replacing certainty with a probability, treating 'I'm not sure' as accuracy rather than weakness, and building a truth-seeking group that rewards you for being right about your confidence.
You call the decision that worked a good one and the one that failed a mistake. A poker professional explains why that habit is teaching you the wrong lessons, and what to do instead.
The book in essence
Duke's frame is that life is poker, not chess. In chess all the information is on the board; in life you decide with some of it hidden and then the future deals a card. So a good decision can lose and a bad one can win, and the habit of resulting, judging by the card, poisons learning. She proposes four boxes instead of two: decision quality and outcome quality, held apart. The rest of the book is about living with uncertainty honestly: stating beliefs as probabilities so that new evidence can move them, saying 'I'm not sure' out loud, and recruiting a small group committed to accuracy over agreement, who ask 'how sure are you and why?' rather than praising the call.
Published in 2018, after Duke's move from poker to decision consulting, it draws on Kahneman, Tetlock and behavioural economics but is written from the table, with hands, bad beats and the discipline of a player who had to separate skill from luck to survive.
Who it's for
- Anyone who reviews past decisions by whether they worked
- Managers and founders whose teams are judged on outcomes they only partly control
- Readers who want to be less certain and more accurate at the same time
Who it's not for
- — Readers looking for decision tools step by step; Duke gives a stance, the Heaths give the steps
- — Anyone allergic to poker; the examples come from the table
Key lessons
01
Separate the decision from the outcome
A good decision can lose and a bad one can win. Judge the decision by what you knew and could have known at the time, and judge the outcome separately. Four boxes, not two.
An architect who turned down a small firm that later thrived calls it her big mistake; at the time the firm had eight months of cash. Good decision, bad outcome.
02
Replace certainty with a number
'I'm sure' invites 'you're wrong'. 'I'm about 70 percent' invites 'why not 60?' The number makes room for being wrong without humiliation and lets evidence move you.
A manager who says he is 65 percent sure a hire will work can update to 55 after a bad month without having to admit he was wrong.
03
Build a truth-seeking group
Left alone, people seek approval. Recruit two or three people who reward accuracy over confidence and ask what would have to be true for your number to be lower.
A product manager sends his monthly probability and reasoning to two friends who only ask: what would have to happen for it to be 40?
Try this today
Pick one past decision you call a mistake or a win. Write what you knew at the time, then sort it into one of the four boxes: decision quality and outcome quality, separately.
Quick check
What is resulting?
Why say 'I'm about 70 percent' instead of 'I'm sure'?
Selected quotes
“What makes a decision great is not that it has a great outcome. A great decision is the result of a good process.”
“Thinking in bets starts with recognizing that there are exactly two things that determine how our lives turn out: the quality of our decisions and luck.”