LAB #4071 NEUTRAL: LEDGER BENCH: Apply filter: ETHUSD SHORT @ NY_OPEN with 6 conf — improves win rate from 32.6% to 63. — E8 Intelligence Research

Abstract

IDEA: Adjustment made to the live model — verify against the backtests: Apply filter: ETHUSD SHORT @ NY_OPEN with 6 conf — improves win rate from 32.6% to 63.6% (+31.1pp, n=11) SAME-WINDOW EFFECT: Applying this filter would have blocked the ETHUSD SHORT trades that occurred at NY_OPEN with 6 conf — but none of the listed live trades are ETHUSD, so zero trades in the current window would be affected. The claimed +31.1pp win-rate improvement is based on n=11, which is statistically fragile; over this specific window, the filter would have no PnL delta. The only indirect effect would be a reduction in total trade count, which could slightly lower gross exposure but not alter the outcomes of the listed instruments. OVER-TIME PROJECTION: Over 3–6 months, a filter with n=11 is likely to regress toward the baseline — a 32.6% win rate with 6 conf is already near random for a short-only NY_OPEN entry, and the 63.6% figure could easily be a small-sample artifact. The risk is overfitting to a s Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com