Strategy Learning & Optimization
Uses automated optimization (regime-based + grid search) to tune strategy weights based on your trade history. Weights are used live by the trading engine. Apply only when improved.Strategy Performance
| Strategy | Trades | Wins | Losses | Win Rate | Avg R:R | Expectancy | Status |
|---|---|---|---|---|---|---|---|
| Trend Following | 10 | 6 | 4 | 60.0% | 0.03 | -0.382 | Weak (reducing weight) |
| Breakout | 12 | 8 | 4 | 66.7% | 0.04 | -0.306 | Weak (reducing weight) |
| Mean Reversion | 9 | 3 | 6 | 33.3% | 0.02 | -0.660 | Learning (9/10) |
| Momentum | 12 | 4 | 8 | 33.3% | 0.02 | -0.660 | Weak (reducing weight) |
| Scalping | 2 | 2 | 0 | 100.0% | 0.00 | 0.000 | Learning (2/10) |
| Swing | 11 | 4 | 7 | 36.4% | 0.04 | -0.621 | Weak (reducing weight) |
| Position | 6 | 2 | 4 | 33.3% | 0.01 | -0.664 | Learning (6/10) |
Current Scoring Weights
These weights determine how much each dimension contributes to a strategy's score. The learning engine adjusts them based on trade outcomes.
| Strategy | Trend | Momentum | Volume | Structure | Volatility | Risk Qual | Actions |
|---|---|---|---|---|---|---|---|
| Trend Following | 30 | 25 | 15 | 15 | 10 | 5 | |
| Breakout | 15 | 20 | 25 | 20 | 15 | 5 | |
| Mean Reversion | 10 | 25 | 20 | 15 | 20 | 10 | |
| Momentum | 20 | 30 | 20 | 10 | 10 | 10 | |
| Scalping | 5 | 20 | 20 | 15 | 25 | 15 | |
| Swing | 30 | 25 | 15 | 20 | 5 | 5 | |
| Position | 35 | 20 | 15 | 20 | 5 | 5 |
Performance by Regime
How each strategy performs in different market conditions.
| Strategy | Trending | Ranging | Volatile | Compression | Mixed |
|---|---|---|---|---|---|
| Trend Following | - | - | 4W / 1L (80%) | 2W / 3L (40%) | - |
| Breakout | 5W / 3L (63%) | 0W / 1L (0%) | 1W / 0L (100%) | 2W / 0L (100%) | - |
| Mean Reversion | 1W / 5L (17%) | 1W / 0L (100%) | 0W / 1L (0%) | - | 1W / 0L (100%) |
| Momentum | 3W / 6L (33%) | - | - | 1W / 1L (50%) | - |
| Scalping | - | - | 1W / 0L (100%) | 1W / 0L (100%) | - |
| Swing | 3W / 3L (50%) | - | - | 0W / 3L (0%) | 1W / 0L (100%) |
| Position | 2W / 4L (33%) | - | - | - | - |
How the Learning Engine Works
1. Record outcomes
Every closed trade saves its strategy type, market regime, P&L, and risk/reward ratio.
Every closed trade saves its strategy type, market regime, P&L, and risk/reward ratio.
2. Calculate expectancy
Expectancy = (Win Rate x Avg R:R) - (Loss Rate x 1). Positive = profitable strategy over time.
Expectancy = (Win Rate x Avg R:R) - (Loss Rate x 1). Positive = profitable strategy over time.
3. Adjust weights
After 10+ trades, strategies with negative expectancy get scoring weights reduced by 5%. Strategies with high expectancy (>0.5) get a 2% boost.
After 10+ trades, strategies with negative expectancy get scoring weights reduced by 5%. Strategies with high expectancy (>0.5) get a 2% boost.
4. Live use in signals
Strategy weights above are passed to the trading engine. Each strategy’s score uses its learned weights. Regime gating (e.g. no Mean Reversion in trending) and min 5 trades per strategy apply.
Strategy weights above are passed to the trading engine. Each strategy’s score uses its learned weights. Regime gating (e.g. no Mean Reversion in trending) and min 5 trades per strategy apply.
Note: Weight adjustments are gradual. A strategy needs 20+ trades with negative expectancy before weights start decreasing.
Optimize: Click "Optimize" next to a strategy (10+ trades) to run automated weight optimization. Only apply when improved.