[EM] ICT, choosing differently among unbeaten candidates.

[Michael Ossipoff]

Though ICT, unless there's exactly 1 unbeaten candidate, chooses by
top-count, it seems to me that the top-count choice is only needed
when there are no unbeaten candidates.

If there hasn't been a defection that causes defeat for the candidate
defected against, then there isn't a defection problem. I don't think
that the top-count is needed for choosing among unbeaten candidates.

So, for that purpose, one could use any FBC-complying rank-count, such
as ER-Bucklin, MMPO, or SR.

MMPO has been criticized for failing the Plurality Criterion, but,
when MMPO is only used for choosing among the unbeaten candidates, it
won't make a Plurality Criterion failure.

Though MMPO requires more count labor than the other two above-listed
methods, ICT and Symmetrical ICT already are pairwise-count methods
with the count labor that goes with that.

Maybe MMPO would be the best choice, then, for choosing among the
unbeaten candidates. For FBC-compliance, it might be best to look at
_all_ of the pairwise oppositions to the unbeaten candidates, not just
those involving other unbeaten candidates.

I don't suppose that this change would bring any new
criterion-compliances to ICT and Symmetrical ICT, but there might be
instances when it would give a more aesthetically-appealing and more
easily justified result than the top-count, among the unbeaten
candidates.

I emphasize that, for choosing among unbeaten candidates, it doesn't
really matter what method is used (as long as it doesn't bring
FBC-failure), because all of the unbeaten candidates are
well-qualified to win.

Whether that aesthetic improvement justifies the complication of using
a different count rule for choosing among the unbeaten candidates, I
don't know. For simplicity, I'd be inclined to stick with the
top-count whenever there isn't exactly one unbeaten candidate.

I just wanted to mention that possibility.

Mike Ossipoff