LAB #2108 NEUTRAL: LEDGER BENCH: Apply filter: NZDUSD SHORT — improves win rate from 32.0% to 34.3% (+2.3pp, n=102) — E8 Intelligence Research

IDEA: Adjustment made to the live model — verify against the backtests: Apply filter: NZDUSD SHORT — improves win rate from 32.0% to 34.3% (+2.3pp, n=102) SAME-WINDOW EFFECT: Applying a filter that blocks NZDUSD SHORT trades would have had zero direct impact on the current live window — none of the ten listed trades involve NZDUSD. The only indirect effect would be a marginal reduction in total trade count if any NZDUSD SHORT signals were generated and skipped during this period, but no such trades appear in the log. Therefore, the win-rate improvement of +2.3pp (n=102) cannot be validated or refuted by this window; it is statistically weak (95% CI roughly ±9pp) and likely noise. Expected PnL delta in this window: $0. OVER-TIME PROJECTION: Over 3-6 months, if the filter is genuinely selective, you might see a small win-rate lift, but the sample size (102) is far too small to distinguish a real edge from random variation. The risk is overfitting to a specific pair/direction that may b Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com