Comment by altcognito

Still have to trust that the model isn't doing something like reducing effectiveness in certain situations.