LAB #4439 HARMFUL: LEDGER BENCH: Apply filter: NZDUSD LONG @ NY_OPEN with 6 conf — improves win rate from 31.5% to 85.7 — E8 Intelligence Research

IDEA: Adjustment made to the live model — verify against the backtests: Apply filter: NZDUSD LONG @ NY_OPEN with 6 conf — improves win rate from 31.5% to 85.7% (+54.2pp, n=28) SAME-WINDOW EFFECT: Over the current live window, the filter would have blocked the NZDUSD SHORT entries (rizzy5m at 0.58252 and wkjudas at 0.59214) — both were wins, so applying a LONG-only filter would have removed two profitable trades. It would also have blocked the AUDUSD LONG (rizzy7m, rejected/open) and possibly the GBPUSD LONG (wkjudas, loss), but no NZDUSD LONG existed in this window to benefit from the filter. Net effect: the filter would have removed 2 wins and 1 loss, reducing total PnL slightly (win rate would drop from ~5/8 to ~3/6 in this small sample), contradicting the backtest's direction. OVER-TIME PROJECTION: The claimed +54.2pp win-rate improvement on n=28 is statistically fragile — a 95% confidence interval around 85.7% spans roughly 67–96%, and the baseline of 31.5% likely comes from a di Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com