LAB #4988 NEUTRAL: LEDGER BENCH: Apply filter: NZDUSD LONG — improves win rate from 31.7% to 54.5% (+22.8pp, n=121) — E8 Intelligence Research

IDEA: Adjustment made to the live model — verify against the backtests: Apply filter: NZDUSD LONG — improves win rate from 31.7% to 54.5% (+22.8pp, n=121) SAME-WINDOW EFFECT: Over the current live window, this filter would have touched exactly one trade: NZDUSD SHORT via rizzy5m at 2026-09-09 11:45 (currently REJECTED/open). Since the filter is NZDUSD LONG-only, it would have blocked that short rather than created a long — so the affected set is one trade, not the 121 in the backtest. That trade's outcome is unresolved, so the expected delta is "one open position removed," not a measurable win-rate gain; the +22.8pp headline does not transfer to this window because the window contains no NZDUSD LONG entries at all. OVER-TIME PROJECTION: If the 54.5% vs 31.7% split holds out-of-sample, restricting NZDUSD to longs would roughly halve the trade count on that pair while lifting per-trade expectancy — a real edge if the asymmetry is structural (e.g., NZD carry/rate-differential regime) ra Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com