MODEL ADOPTION #1216: REJECT — LEDGER BENCH: Apply filter: Phase distance FAR (>10deg) — improves win rate from 31.5% to — E8 Intelligence Research

DECISION: REJECT ELEMENT: The FAR >10deg filter fails hostile scrutiny because the claimed +0.5pp win-rate improvement (n=658) is statistically fragile — a binomial test shows p ≈ 0.18, meaning this result is easily explained by noise. The "promising" verdict at 65% confidence is inflated for a marginal, non-significant edge. Moreover, the filter's rejection of trades near phase boundaries directly contradicts E8's core premise: that phase transitions are precisely where FADE polarity edges trigger. Filtering them out discards the very entries the model is designed to capture, reducing trade frequency by 5-10% and likely culling high-R:R breakouts. The shadow-test proposal is sound, but the burden of proof is on the candidate — and it fails to meet E8's requirement for a concrete, geometrically justified edge. EXPECTED IMPACT: No win-rate improvement; likely a net negative due to missed profitable entries in volatile regimes. RISK: Degrades model by suppressing entries at phase b Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com