- Tetlock and Gardner tell the story of the Good Judgment Project, which dominated a U.S. intelligence-community forecasting tournament on world-affairs questions by out-predicting professional analysts with access to classified data [1].
- Granularity itself predicts accuracy: superforecasters gave finer-grained probabilities than ordinary forecasters, with a third of their estimates using single percentage points (3% rather than “about 5%”) [2]. Artificially rounding everyone’s forecasts to coarser numbers cost superforecasters accuracy even at the smallest rounding, while ordinary forecasters barely lost anything from rounding four times as coarse [2].
- The single strongest predictor of becoming a superforecaster wasn’t intelligence — it was “perpetual beta,” the habit of treating every forecast as provisional and revising it as facts changed, about three times as powerful a predictor as raw intelligence, its closest rival [3].
- The book names a real tension for managers: the hedging, self-doubt, and constant revision that make someone a good forecaster look like the opposite of what a confident, decisive leader is supposed to project [4].
- Tetlock and Gardner resolve that tension with Auftragstaktik, the nineteenth-century Prussian command doctrine of giving subordinates a clear objective and letting them improvise the execution [4], since “no plan survives contact with the enemy” [4].
Superforecasting: The Art and Science of Prediction
Tetlock and Gardner's account of the Good Judgment Project, profiling the traits of the best forecasters — actively open-minded, reflective, and quick to take responsibility for their misses.
- Author
- Philip E. TetlockDan Gardner
- Published
- 2015
- Purchase
- Amazon
- When you catch yourself estimating a hire’s odds of success as a rough “70/30” and stopping there: push for one more digit. Granularity itself predicted accuracy in the tournament data [2] — forcing “68%, not 70%” is not false precision, it’s the habit that separated the best forecasters from everyone else.
- When you’ve stated a forecast — a hiring bar, a revenue number, “this will work” — and new information arrives that contradicts it: the instinct to hold the line to look decisive is backwards. Commitment to revising beliefs was the single strongest predictor of forecasting skill in this research, about three times as powerful as intelligence [3].
- When you feel like the humility good judgment requires would undercut your authority as a leader: it doesn’t have to be one or the other. Set the objective and the constraints, and hedge and forecast on the path to get there — the “leader’s dilemma” the book poses is resolved by separating what you’re certain about (the goal) from what you should stay probabilistic about (the plan) [4].