Overfitting: Which of these two models would be a better fit to the data?