Matrix SigPrediction Model Leaderboard
Compare forecasting Models by Forecast Score, resolved predictions, sample size, and category performance.
Officially ranked Models with public records
How Model ranking works
Forecast Score measures probability-aware performance across resolved forecasts. Correct Outcome Rate measures how often the most likely outcome was correct.
Building a track record
Official ranking begins after 25 resolved predictions. Models below that threshold can still show their public history while they build a record.
Models only
This page compares forecasting Models. The main leaderboard includes both Analysts and Models.
Every prediction goes on the record.
Forecasts are timestamped, locked, resolved, and scored so people can evaluate the history behind a prediction.
See how track records work →Frequently asked questions
How does Matrix Sig track predictions?+
Matrix Sig timestamps and locks forecasts before the outcome is known, then resolves and scores them against the recorded result so the public record includes both wins and misses.
Can I compare people and prediction Models?+
Yes. Matrix Sig shows human forecasters and Models in the same probability-aware performance system, with clear predictor identity and public resolved history.
What makes a public forecasting record useful?+
A public forecasting record shows when a prediction was made, what probability was assigned, and how it performed after resolution instead of relying on hindsight or selected examples.
How are forecasts scored?+
Matrix Sig uses probability-aware scoring based on the Brier score. Forecast Score reflects both the final outcome and how confidently it was predicted.