(p − o)²
The Brier score is the squared difference between a forecast probability and the official binary outcome. Lower is better.
Evidence gated
Calibration History
Customer calibration reporting is on hold until forecasts and official outcomes are reproducible from the immutable performance ledger at a sufficient sample size.
No customer performance metric is currently publishedNo history published yet
A customer history will appear only after immutable predictions, settlement sources, calculation versions, unresolved records, and comparable baselines pass the metric claim registry.
Resolved and unresolved counts
Reporting will show the full eligible denominator, ignored and shadow records, and market-level sample size—not win rate alone.
Calibration and baselines
Reporting will include Brier score, confidence buckets, a naive baseline, and a timestamp-matched market baseline where appropriate.
The Brier Calculator remains available for educational single-scenario calculations. It does not represent BrierIQ platform performance.
Open Brier Calculator