A backtest is a simulation of a model-driven investment strategy's response to historical data. The purpose of performing experiments with backtests is to make discoveries about a process or system. Using historical data, you can save time in testing an investment strategy for the period forward. It helps you test an investment theory based on the movements of the tested period. It is also used to both evaluate and calibrate an investment model.

Creating a model is only the first step. The investment strategy will typically employ the model to help you drive simulated trading decisions and compute various factors related to either risk or return. These factors are typically used together to find a combination that is predictive of return.