We should always imagine that there is some non-zero likelihood for absolutely any observation in absolutely any theory @SeanMCarroll