Thinking in Bets
Decision quality depends on reasoning under uncertainty, not simply whether the latest outcome happened to be favourable.
Outcomes combine decisions with conditions outside the decision-maker's control.
A poor outcome can follow a careful decision, and a careless decision can get lucky. Duke asks us to judge the reasoning available beforehand rather than let hindsight rewrite it.
A good result does not prove a good process.
Beliefs can be held with degrees of confidence.
New evidence should change those degrees rather than threaten identity.
Alternative explanations counter the appeal of one coherent story.
A probability is not an admission of weakness. It makes uncertainty explicit, allowing new evidence to change a belief without requiring the person holding it to defend an entire identity.
A truth-seeking group can expose gaps in private reasoning.
Incentives and belonging can also bias group discussion.
Considering possible futures makes hidden risks more visible.
Reviewing decisions requires the information available at the time.
The central aim is calibrated learning rather than certainty about every result.