Comparative analysis between traditional momentum and machine learning (random forest): evidence from the S&P 500 (2000-2024)

Introduction: This study examines the validity and persistence of the momentum effect in the S&P 500 Index (2000–2024), a developed stock market with high informational efficiency. It analyzes whether empirical evidence supports the persistence of momentum across different time horizons. Objective: To compare the performance of traditional momentum (TM) with a supervised learning model based on Random Forest (RF), evaluating predictive power, risk-adjusted returns, and out-of-sample stability. Methodology: Long–short TM strategies were implemented for 1-, 3-, 6-, and 12-month horizons, and the RF model was trained using equivalent cumulative returns. Out-of-sample validation was performed using an expanding window, homogeneous backtesting, and temporal stability tests. Results: MT showed limited performance over short horizons and greater consistency over long horizons. RF demonstrated greater predictive power and profitability, especially over long horizons, although it exhibited episodes of volatility and a risk of overfitting. Discussion: Machine learning models capture nonlinear patterns that cannot be identified by traditional methods, but they depend on market conditions