MODEL ADOPTION #5008: REJECT — TG AUTO: New edge proposed: kaufman_efficiency_ratio_regime_switch — E8 Intelligence Research

DECISION: REJECT ELEMENT: Nothing survives as a standalone adoptable rule. The only salvageable element is a *diagnostic tag*, not a gate: compute ER(20) at every historical entry bar and bucket trades by ER regime — but do NOT wire ER into live entry logic yet. Kaufman's 0.35/0.15 thresholds are calibrated for daily bars; on 15s-7m scalps the ER distribution is compressed and laggy (20 bars = 5 min of history on a 15s chart, which is longer than most of your holding windows). Gating on a laggy regime classifier inside a 15s-7m scalp with pre-set sacred-geometry targets is structurally incompatible: by the time ER confirms "trend," the geometry target is often already hit or invalidated. EXPECTED IMPACT: Zero-to-negative on live PnL if adopted now. Best case from Merlin's own projection is "4-6 fewer losses, 0-2 fewer wins" on a 10-trade sample — that is noise, not edge. Over 3-6 months expect 30-50% trade-count reduction with unknown sign on net PnL, because you have not yet establi Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com