Backtest Lab — locked production strategy · core pairs 2002–2026 · 2026 additions 2020–2026
This lab runs the RETIRED CORE6-era spec for research comparison. It does not test the current evolved system (agent worker + learning layer + HS60_75); the evolved stack's backtest is the track-record card's walk-forward study.
Select instruments, then run. Everything else is fixed.
Strategy locked to the retired CORE6-era spec — parameters not adjustable. Spec read from the retired CORE6 config on every run · 12h max hold · 0.20%/trade · 1.50% open-risk cap · 99%-of-peak ratchet floor · V3 drift trail (arm −0.40R, close-all −0.90R) · 1-loss 30-min cooldown · ML gate (R4 v2.1 model) · FCFS allocation.
running…
Results
Equity — fixed $100K sizing (1R = $200)
Drawdown — % underwater
FTMO challenge timeline — walk-forward, weekly starts
Evaluation = time to +10% (Phase 1). Verification = time to +5% (Phase 2). Full funding ≈ evaluation + verification medians. Breach = 10% trailing drawdown or 5% daily loss.
Monthly returns — $ at fixed sizing (hover for R)
Yearly profit estimates
Per-instrument breakdown
Exit mix
Baseline comparison — retired CORE11 backtest (engine basis)
History: the retired CORE11 official backtest — 2020–2026, 11 pairs, +0.0929R/trade, PF 1.80, maxDD 2.31%, 72,259 trades (engine basis; superseded by the 2026-10-14 live-true rebuild — operative baseline +0.0164R/trade, 10-pair roster (v2.2 yardstick re-issue 2026-10-16; v2.1-era figure +0.0150R)).
History: the CORE6 record this supersedes — 2006–2026, 6 pairs, +0.1266R/trade, PF 2.27, maxDD 1.68%, 45,243 trades (pre-roam-free spec).