Every forecast, tracked.
VORENTH's Research Desk produces its own forecasts and scores every one against real-world outcomes. Brier scoring, calibration curves, contrarian calls, and recent misses — all public, all verifiable.
The Research Desk is currently accumulating forecasts. Accuracy metrics will appear as predictions reach their target dates and are resolved against real-world outcomes.
Only house forecasts appear here. User search queries are private and never included in the public record.
Loading forecast data...
How we score
Brier Score
Measures the accuracy of probabilistic forecasts. Computed as (predicted probability - actual outcome)². Lower is better. A perfect score is 0.0; random guessing scores 0.25.
Calibration Score
Measures how well predicted probabilities match actual resolution rates. If VORENTH says 70%, events should resolve ~70% of the time. Score of 100% = perfectly calibrated.
Market Calibration
Forecast probabilities are anchored against live prediction market prices (Polymarket). Deviations from market consensus require specific evidence justification.
Base Rate Anchoring
VORENTH uses its own historical accuracy to ground new forecasts. If past market predictions were overconfident by 10%, future estimates are automatically adjusted downward.
Only house forecasts (discrete, falsifiable, time-bounded claims produced by VORENTH's Research Desk) appear in the public record. User search queries are private and never included. No forecasts are retroactively edited or removed. Incorrect calls and unresolvable forecasts are displayed prominently — not hidden.