MODEL ADOPTION #4105: REJECT — LEDGER BENCH: Apply filter: Quarterly phase = DISTRIBUTION — improves win rate from 32.6% — E8 Intelligence Research

DECISION: REJECT ELEMENT: No single rule survives — the DISTRIBUTION-phase filter is a macro-label overlay, not a timing edge compatible with E8's intra-bar geometry. It cannot be expressed as a precise entry/exit condition on the Gods-Clock rhythm without introducing lookahead bias or phase-lag distortion. EXPECTED IMPACT: Win-rate gain of +4.7pp is real but fragile; over 60 days, reduced trade frequency (~30–40% fewer signals) would cut absolute R-multiple accumulation by more than the win-rate improvement adds. Net profit likely flat to negative. RISK: Phase definitions are subjective and lag regime shifts — filter would block high-quality trades during early accumulation or late markup, and overfit to a single quarterly label. Also, the n=110 backtest is too small to separate edge from noise; the same-window projection showed net PnL near zero. TEST: Before any re-consideration, run a 60-day shadow test on the live engine with both cohorts (filtered vs unfiltered) paper-log Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com