The portion of historical data on which a strategy is developed or optimized. When backtesting, traders often split data into two sets: the in-sample (for designing and tuning the strategy rules and parameters) and the out-of-sample (for validation). The in-sample period is where the “learning” or fitting happens. For instance, one might use 2015–2020 data in-sample to optimize an expert advisor’s settings. (See also Out-of-Sample.)

Back to Robots Glossary Index