MODEL ADOPTION #327: REJECT — LEDGER BENCH: Apply filter: Phase distance TIGHT (<=2deg) — improves win rate from 32.4% t — E8 Intelligence Research

DECISION: REJECT ELEMENT: The phase distance TIGHT (≤2°) filter is not adopted. The backtest win-rate gain of +4.4pp (n=370) is statistically fragile — the confidence interval around a 32.4% baseline with n=370 is roughly ±5pp at 95% confidence, meaning the observed gain is within noise. The shadow projection shows a 42.9% → 66.7% win-rate shift on only 7 closed trades, which is not replicable. The filter reduces trade count by 15-25%, and the net PnL impact is unknown because the backtest did not report average PnL per trade or total PnL. The risk of overfitting to recent phase dispersion regimes is high, and the E8 geometry timing already encodes phase relationships — adding a hard distance cutoff may discard signals that are geometrically valid but have wider phase spread due to market microstructure noise. EXPECTED IMPACT: No adoption. The win-rate improvement is not statistically robust, and the trade count reduction likely lowers absolute PnL. The filter may reduce drawdowns sl Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com