A machine learning model that predicts whether the S&P 500 index will go up or down the next trading day. Uses a Random Forest classifier trained on historical price data with rolling window features, evaluated through a walk-forward backtesting system to simulate real trading conditions.