Every forecast Anteproof publishes, in the order issued, including the short-horizon prediction-market questions used to test the forecaster. Those scores are not a track record for the maintenance forecasts, whose past scores are withheld.
Every forecast that resolved in the last seven days, counted. The best real calls and the worst misses side by side: same card, same size, either way.
The most recently scored forecast, hit or miss. Every receipt gets this stamp when its answer lands.
Loading…
Every live forecast, in the order it was issued, across every question family — maintenance questions, political and prediction-market questions alike, unfiltered. A forecast is never edited or deleted after issue. Resolved questions show the real outcome next to what we said.
We hold ourselves to a statistical launch gate: sample size, skill over a naive baseline, and two independent bias tests. Every evaluation is published verbatim. Credibility is the audit trail, not the absence of failures. The full breakdown, with the reliability diagram, Brier score by question family, and live versus backtest, is on the calibration page.