> In my opinion, any model that predicts any sort of outcome when doing so accurately is overfitted and lacks the proper contextual knowledge necessary For me,

> In my opinion, any model that predicts any sort of outcome when doing so accurately is overfitted and lacks the proper contextual knowledge necessary For me, models that do not make predictions are not very useful. Without predictions about the future, it's very hard to tell if the work is useful moving forward, or if it should be read as a historical analysis that doesn't matter in the current environment. Predictions are critical, they are a primary mechanism to find out if something is useful.