In science, the kind of evidence matters; all unlikelihoods are not created equal.