Loading backtest results…
Performance Analysis
Walk-Forward Backtest Results
Out-of-sample validation — 70% training data, 30% unseen test data. Run
Methodology: Historical data split 70/30. ML models trained on the first 70%, tested on the remaining 30% — data never seen before. Proves real edge, not curve-fitting.
Risk/Reward: 2:1 (SL = 1.5×ATR, TP = 3.0×ATR). Trend alignment filter applied. 2% risk per trade.