ACCURACY, IN PUBLIC
Track record
We publish every question we ever opened — including the misses — and score ourselves against a naive forecaster who always predicts the historical base rate.
0questions resolved yet
2026-10-31first resolutions expected
Calibration curve
Appears automatically once 30 questions have resolved — a calibration curve on fewer would be noise dressed as rigor.
0 / 30 resolvedHow to read this. The Brier score is the squared error of a probability forecast — 0 is perfect, lower is better. A naive forecaster who always predicts each question’s historical base rate sets the bar; beating it is the entire test. Method and limits: how The Outlook works.