ML Intelligence
4-model ensemble — vote confidence vs. realized win rate across all strategies
The 4-Model Ensemble
1.XGBoost (price action)
2.LightGBM (options flow)
3.Random Forest (regime)
4.Gradient Boost (volatility)
Vote Threshold → Position Size
4/4Full convictionFULL
3/4Strong majorityHALF
2/4Split — size downQUARTER
≤1/4Skip signalSKIP
Total Trades
…
Overall Win Rate
…
Avg Return
…
Bot Trades
…
Short + Long
Win Rate by ML Vote CountDoes higher consensus → better outcomes?
Loading…