MODEL ADOPTION #998: REJECT — LEDGER BENCH: Apply filter: NZDUSD LONG — improves win rate from 31.6% to 35.5% (+3.8pp, n — E8 Intelligence Research

Abstract

DECISION: REJECT ELEMENT: The NZDUSD LONG filter is not compatible with E8 geometry timing. The filter is currency-pair-specific and direction-specific, but does not align with the rhythm logic of Gods-Clock or the FADE polarity edge. It introduces a static bias that conflicts with the dynamic, time-based entry conditions of the E8 model. No concrete, testable edge survives. EXPECTED IMPACT: Win-rate improvement of +3.8pp is based on n=62, which is statistically insignificant for a live model. The filter would reduce trade count and may increase losses on correlated USD pairs (USDJPY, USDCAD) during USD-weakness regimes, likely degrading total PnL. RISK: Overfitting to a small sample. Hidden negative correlation with other USD pairs. Reduction in trade frequency reduces model adaptability and may cause missed opportunities during non-NZDUSD windows. TEST: Before any re-consideration, run a replay-bench test on the last 12 months of E8 tick data: compare control vs. experimental (NZ Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com