The 90th percentile of the oversubscription relative-error series, bounding the error of the worst 10% of predictions.
Methodology
The 90th percentile (P90) of per-stock relative errors.
Current Value
1018.9%
Sample Size
214
Methodology
This metric comes from a walk-forward backtest on stocks listed in 2025 and later — the anchors (market regime, index, sponsor heat, first-day P&L) use only data visible before each stock listed; the regression coefficients are fitted on the full history, matching production prediction.
This page documents the definition, methodology and current value of the prediction metric "Oversubscription P90 Relative Error". The value is aggregated from the prediction_history table and refreshed by the backfill script, for objectively assessing model accuracy — not investment advice.