An exploratory framework for rigorously backtesting algorithmic trading strategies. This project investigates how quantitative metrics can reliably separate robust models from curve-fitted anomalies, bridging the gap between historical simulation and live paper-trading validation.