Adjusted R-squared and predicted R-squared can help prevent you from adding too many variables to your regression model.
