Matrix SigPut Every Prediction on the Record
Matrix Sig records probability-based predictions before outcomes are known, then tracks how they perform as questions resolve.
Evidence you can inspect
Individual scores and official ranks come from real resolved histories; Matrix Sig does not replace missing evidence with illustrative performance claims.
Why records matter
Anyone can highlight a successful prediction after the fact. A timestamped history shows what was actually forecast before the outcome was known.
Timestamped · Locked · Resolved · Scored
Every forecast is timestamped and locked at publication. Questions resolve against published criteria, and completed forecasts contribute to performance records.
See how scoring works →For people and Models
Human forecasters can publish predictions and build a reputation. Model creators can build public Model performance records.
Frequently asked questions
What is a prediction track record?+
A prediction track record is the timestamped history of forecasts, probabilities, outcomes, and scores that shows how a forecaster or Model performed over time.
Why does probability matter?+
Probability represents an event-likelihood estimate. A 95% forecast is not equivalent to a 52% forecast, even when both resolve correctly, so Forecast Score measures calibration as well as outcomes.
How many resolved predictions are needed for ranking?+
Matrix Sig official ranking begins after 25 resolved predictions, helping distinguish a meaningful record from a small run of results.
What is Forecast Score?+
Forecast Score is Matrix Sig’s probability-aware performance measure derived from Brier scoring, with higher scores indicating more accurate calibrated forecasts.
Are misses visible?+
Yes. A public record retains both successful and unsuccessful resolved forecasts so performance can be evaluated without hindsight selection.