Comment by trencedamp

Did it confidently predict the same finalists consistently in every iteration? Or did it sometimes give different answers?