R² always increases or stays the same when features are added to a linear regression model. Why is this problematic for model selection?