LAB #4387 HARMFUL: LEDGER BENCH: Apply filter: USDCHF SHORT @ OFF_SESSION with 6 conf — improves win rate from 31.6% to — E8 Intelligence Research

IDEA: Adjustment made to the live model — verify against the backtests: Apply filter: USDCHF SHORT @ OFF_SESSION with 6 conf — improves win rate from 31.6% to 66.7% (+35.1pp, n=6) SAME-WINDOW EFFECT: Over the current live window, this filter would have blocked the USDCHF LONG (rizzy7m) — a winner — and would have had no direct effect on any other listed trade since none are USDCHF SHORTs at OFF_SESSION. The claimed +35.1pp win-rate improvement is based on n=6, which is statistically fragile; the actual PnL delta here would be negative (removing a confirmed win) with no compensating new entries. Net expected change: −1 win, no loss avoided, so win-rate would drop from 6/9 to 5/8 in this sample — the opposite of the backtest claim. OVER-TIME PROJECTION: Over 3–6 months, this filter would likely reduce trade frequency on USDCHF SHORTs during off-session hours, but the 66.7% win rate on n=6 is indistinguishable from noise (95% CI roughly 22–96%). The risk is overfitting to a tiny, possib

Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com