The existing Z-Score EAs test profitable on M1 OHLC but fail under
Every Tick because they decide on individual ticks: noise fires the
Z-exit at the worst price, market-order exits pay spread twice, and
no minimum-edge check means quiet-market trades only ever pay spread.
XAU_Quant_Reversion_TickRobust.mq5 closes that gap:
- All decisions (entry, exits, breakeven, TP retarget) run exactly
once per bar on closed-bar data, so OHLC and Every Tick converge
- Cost gate: StdDev >= 3x and TP distance >= 4x the round-trip cost
(live spread + commission points) or the trade is skipped
- Server-side limit TP placed at the mean, retargeted per bar with a
cost-covering floor (no spread paid on exit)
- Turn-confirmation bar before entry (no knife-catching)
- Volatility-aware SL: max(800 pts, 2.5x ATR) against tick spikes
Docs: README section for the new EA, CLAUDE.md overview updated.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012knbN7oGPQ7K4TnRroZjWF
Adds full documentation for XAU_Quant_Reversion_Breakout.mq5 including
dual strategy overview, all input parameters, and design notes. Also
adds missing InpMaxPositions param to the primary EA's parameter table.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Old README described features that no longer exist (partial closes,
trailing stops, news filter, scaled entry, weekly/equity stops,
volatility filter). Updated to accurately document both gold and BTC
EAs with their current parameters and design rationale.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Covers all current functionality: RSI filter, 3-stage exit system,
stepped trailing, 3-layer capital protection, tightened entry filters,
and updated input parameter tables with current defaults.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Covers strategy logic, entry/exit flow, news filter tiers,
daily loss limit, all input parameters, and installation steps.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Neo Malesa