Strategies
Ranked by expectancy. IC = Pearson correlation between strategy signal (dir × confidence) and realised 5-min move; ≥+0.20 = ×1.5 boost, |IC|<0.05 or IC≤−0.10 = ×0.5 dampen. Needs ≥10 pairs. Sub-strategies that don't open trades on their own show italic "attr" cells — proxy stats from distributing each composite trade's PnL by signed contrib weight.
| Strategy | Trades | WR | Expectancy | Cum PnL | IC | Horizon | Pairs | IC effect | Weight | State |
|---|
By Timeframe
| TF | N | WR | Exp |
|---|
By Side
| Side | N | WR | Exp |
|---|
By Plan
| Plan | N | WR | Exp |
|---|
Loss patterns — market regimes that systematically lose
Cross-references closed PM trades with the regime snapshot at decision time. ≥8 trades for danger, ≥15 for sweet spot.
🔴 Danger zones (WR ≤ 35%, n ≥ 8)
| Feature | Bucket | Trades | WR | Mult |
|---|
🟢 Sweet spots (WR ≥ 60%, n ≥ 15)
| Feature | Bucket | Trades | WR | Mult |
|---|
All regime features & buckets
Confidence calibration
Engine's confidence vs actual outcome WR. Multiplier shown is what calibrator will apply.
| Bucket | Trades | Predicted | Actual WR | Δ | Multiplier |
|---|
Per-plan calibration Phase 10
Each plan's own confidence-bucket multipliers. Engine consults this first when
params.applyCalibration=true AND bucket health passes
(≥3 buckets with ≥5 trades). Otherwise falls back to the global table above.
| Plan | Trades | Buckets ≥5 | Reset cutoff | Decision | Multiplier range | Window |
|---|