Of course there's more than that. Predicting higher uncertainty than warranted would be a different failure in the model. But that didn't really happen in that election.