LAB #3168 HARMFUL: LEDGER BENCH: Apply filter: USDCHF LONG @ LONDON_OPEN with 5 conf — improves win rate from 31.4% to — E8 Intelligence Research

IDEA: Adjustment made to the live model — verify against the backtests: Apply filter: USDCHF LONG @ LONDON_OPEN with 5 conf — improves win rate from 31.4% to 80.0% (+48.6pp, n=5) SAME-WINDOW EFFECT: Over the current live window, the filter would have blocked the USDCHF SHORT (rejected) and likely flipped the bias to LONG at London open — but no USDCHF LONG entry exists in the window, so net effect is zero trades changed. The only USDCHF signal was a SHORT that was already rejected; applying a LONG-only filter would not have altered any other pair's outcomes. Win rate would remain identical (7 wins / 9 closed = 77.8%), PnL delta ≈ 0. The claimed +48.6pp improvement is based on n=5, which is statistically meaningless (95% CI spans roughly 30%–100%). OVER-TIME PROJECTION: Over 3–6 months, this filter would eliminate all USDCHF SHORTs and only take LONGs at London open with 5 conf. Given the GTX base loss rate (155 losses vs 80 wins in 30 days), a single-pair, single-session, single-dire Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com