On the question of empirical corrections in ab initio model chemistries