I was trying to get at Aumann being clever and statistics-savvy enough not to state that they exist for certain, but that he is almost as wrong to assign them any non-negligible probability (as he clearly does).