How do we know that our model is good, or bad?

TBD.