LAB #4967 PROMISING: LEDGER BENCH: Apply filter: XAUUSD LONG — improves win rate from 31.7% to 54.1% (+22.4pp, n=85) — E8 Intelligence Research

IDEA: Adjustment made to the live model — verify against the backtests: Apply filter: XAUUSD LONG — improves win rate from 31.7% to 54.1% (+22.4pp, n=85) SAME-WINDOW EFFECT: The filter would have blocked the XAUUSD SHORT on 2026-09-10 (a closed loss), which is the only XAUUSD trade in the live window — so it removes one loser and leaves the other nine trades untouched. That's a +1 win-rate improvement in a 10-trade sample, but it's a single observation and tells you nothing statistically; the claimed +22.4pp on n=85 is not corroborated by anything in this window. Note the filter is direction-specific (LONG), so it does not touch the short-side bias that produced most of these losses. OVER-TIME PROJECTION: If the 54.1% vs 31.7% split holds out-of-sample, this is a large edge and worth real size — but n=85 with a 22pp swing is exactly the profile of a filter fitted to a handful of lucky XAUUSD long regimes (trending gold, specific volatility bands). Gold's behavior is regime-dependent; Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com