LAB #1930 NEUTRAL: LEDGER BENCH: Apply filter: GBPUSD LONG — improves win rate from 31.9% to 39.1% (+7.2pp, n=46) — E8 Intelligence Research

IDEA: Adjustment made to the live model — verify against the backtests: Apply filter: GBPUSD LONG — improves win rate from 31.9% to 39.1% (+7.2pp, n=46) SAME-WINDOW EFFECT: Over the current live window, applying a GBPUSD LONG-only filter would have blocked the GBPUSD SHORT (rizzy5m, loss) and the two GBPUSD LONGs (sqz win, wkjudas win) would remain. This removes one losing trade and keeps two winners, shifting the window from 4W/4L to 4W/3L — a modest win-rate improvement from 50% to 57% on this small sample. PnL delta is positive but small (~+0.5R) because the blocked short loss was minor relative to the two long wins already captured. The filter does not touch other pairs, so its immediate impact is narrow and directionally favorable. OVER-TIME PROJECTION: Over 3–6 months, a GBPUSD LONG-only filter assumes a structural long bias on that pair, which may hold in trending regimes but will fail in choppy or bearish phases. The backtest improvement (+7.2pp, n=46) is statistically thin — Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com