MODEL ADOPTION #2692: REJECT — LEDGER BENCH: Apply filter: Pool density HIGH (>=90) — improves win rate from 30.9% to 32. — E8 Intelligence Research

DECISION: REJECT ELEMENT: No filter survives — pool density ≥90 is a statistical artifact, not an E8-geometry-compatible edge. The +1.6pp win-rate delta on n=1417 is within noise for a binary outcome (σ≈1.2pp at that sample size), and the filter has no causal link to the Gods-Clock rhythm or FADE polarity timing. It is a liquidity proxy that will correlate with session hours and news spikes, introducing regime bias that E8 does not model. EXPECTED IMPACT: If adopted, win rate might tick up 0.5–1.0pp in the first month (regression to mean), but trade count drops 20–30% (low-density windows are often where contrarian FADE entries catch reversals with high RR). Net PnL likely negative: you lose high-RR winners that occur during fragmented participation (e.g., Asian session breakouts). Over 6 months, expect -5% to -10% PnL vs. baseline. RISK: The filter will systematically exclude entries during low-liquidity windows that E8's rhythm logic specifically targets for fade setups (e.g., Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com