What are out of sample forecasts?

Out-of-sample forecasts are forecasts that are made using data that were not used in training the forecasting model. This means that the model has not been exposed to the out-of-sample data prior to making forecasts and so the accuracy of the forecasts can be assessed. Out-of-sample forecasting is often used in financial modeling, to determine if a model is accurate and reliable enough to be used for making actual decisions. The model's accuracy is then compared with the actual data to determine how well it performs at predicting future prices or other variables.