LAB #2318 NEUTRAL: LEDGER BENCH: Apply filter: GBPUSD LONG — improves win rate from 31.5% to 34.6% (+3.1pp, n=52) — E8 Intelligence Research

IDEA: Adjustment made to the live model — verify against the backtests: Apply filter: GBPUSD LONG — improves win rate from 31.5% to 34.6% (+3.1pp, n=52) SAME-WINDOW EFFECT: Over the current live window, the GBPUSD LONG filter would have blocked the single GBPUSD SHORT trade (entry 1.34508, closed LOSS). That trade was a loser, so removing it would improve the sample win rate from 4 wins/9 closed trades (44.4%) to 4 wins/8 closed trades (50.0%) — a +5.6pp boost, and a small PnL gain equal to that loss. However, the filter does not touch any other instrument, so the XAUUSD, AUDUSD, USDCAD, USDJPY, and EURJPY outcomes remain unchanged. The n=52 backtest improvement (+3.1pp) is directionally consistent but not directly comparable to this 9-trade window. OVER-TIME PROJECTION: Over 3–6 months, a GBPUSD LONG-only filter would reduce total trade count by roughly 10–15% (GBPUSD shorts are a meaningful fraction of the book). If the backtest edge is real, win rate improves modestly, but the big Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com