LAB #2021 NEUTRAL: LEDGER BENCH: Apply filter: ETHUSD SHORT — improves win rate from 31.9% to 50.8% (+18.9pp, n=59) — E8 Intelligence Research

IDEA: Adjustment made to the live model — verify against the backtests: Apply filter: ETHUSD SHORT — improves win rate from 31.9% to 50.8% (+18.9pp, n=59) SAME-WINDOW EFFECT: Over the current 10-trade window, applying an ETHUSD SHORT filter would have blocked zero trades, as no ETHUSD positions appear in the live list. The filter's claimed +18.9pp win-rate improvement (n=59) is derived from a different instrument and time regime, so its direct effect here is nil. However, if the filter implicitly shifts model attention away from other SHORTs (e.g., XAUUSD SHORTs, which are 3 losses in this window), you might see a small indirect PnL improvement — but that is speculative, not measured. Expected PnL delta: ~0% from direct filtering, possibly +1-2% if it reduces correlated short exposure. OVER-TIME PROJECTION: Over 3-6 months, an ETHUSD SHORT-only filter could improve win rate if ETHUSD SHORTs are systematically mispriced by your entry logic — but the sample (n=59) is small and likely o Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com