MODEL ADOPTION #4921: REJECT — LEDGER BENCH: Apply filter: Min confirmations >= 6 — improves win rate from 31.9% to 36.2% — E8 Intelligence Research

DECISION: REJECT ELEMENT: Nothing survives as a standalone rule. "Min confirmations >= 6" is a threshold selected on the same n=944 window it was measured on, with no out-of-sample validation, no per-pair/session stratification, and no confirmation-count exposure in the live window to verify the filter would have actually fired on the losers. The +4.4pp lift is consistent with threshold-shopping on a single dataset — with ~30 plausible thresholds, a 4pp bump on n=944 is well within selection-bias range. It is not a concrete, testable edge yet; it is a hypothesis wearing a rule's clothing. EXPECTED IMPACT: If real, ~+4pp win rate on filtered cohort, but trade count drops (unknown magnitude — the lab didn't report it), so total PnL impact is ambiguous and could be negative if removed trades were net-positive in trending regimes. On the live n=10 window, Merlin's own projection is "small positive PnL delta" — statistically indistinguishable from zero. No credible profit claim can be mad Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com