▸ Concept also: probabilistic forecasting, calibrated forecasting, Brier score

Forecasting

The practice of assigning explicit probabilities to specific future events — and then measuring whether those probabilities were right.

Learn first

AI benchmarks

In a nutshell

Forecasting, in the rigorous sense, means giving a number — not "probably" or "likely" but 0.72 — before an event resolves, then scoring yourself on how close that number was to reality. The standard scoring rule is the Brier score: the mean squared error between your predicted probability and the outcome (0 or 1). A score of 0 is perfect; 0.25 is what you'd get by always guessing 50/50. The hard part is calibration: being right 70% of the time when you say 70%, across thousands of questions. Most people, and most institutions, are not well-calibrated.

Where it came from

Year1950

SourceGlenn Brier — 'Verification of Forecasts Expressed in Terms of Probability' (Monthly Weather Review, 1950)

Why it matteredBrier introduced the squared-error scoring rule for probabilistic weather forecasts; the general discipline of calibrated forecasting was later formalised through tournament work by Philip Tetlock.