📊 Public accuracy statistics
We publish our own performance metrics so you can judge whether to trust the system.
Public accuracy is not open yet
Startup seed values are not treated as track record. Forecasts exist, but their outcomes are not closed in predictions_log yet.
Public accuracy appears automatically: Brier/accuracy after the first closed outcomes, ROC-AUC after at least 10 closed forecasts with mixed true/false outcomes.
Technical stats seed from 6/20/2026, 6:15:09 PM; not a real accuracy recalculation.
Calibration is accumulating
Verified outcomes are still insufficient; the system will update automatically after forecast windows close.
Not enough data for calibration chart
Loading A/B…
How we measure accuracy
- Brier Score — mean squared error between predicted probability and actual outcome. Range 0–1; lower is better. A coin-flip baseline scores 0.25.
- ROC-AUC — area under the receiver-operating characteristic curve. 1.0 = perfect discrimination, 0.5 = random.
- Calibration — for predictions binned by probability, does the actual frequency match? E.g. 70% predictions should happen ~70% of the time.
The event feed updates online. Accuracy is computed only from forecasts with actual outcomes in predictions_log: hourly recalculation can pick up closed windows quickly, but it does not make the score more accurate without new outcomes. Failed predictions downgrade graph edge weights.
Metrics only cover scenarios with a recorded outcome in predictions_log. Not investment advice.