[EM] Conclusion--MJ

[Michael Ossipoff]

MJ advocates claim this "advantage" for MJ:

MJ prevents extreme rating from having such extreme effect.

To say it differently, a non-extreme rating might (or might not) have
as great effect as an extreme rating.

(Of course, because it might not, and if you want to maximize the
effect of a rating, an extreme rating is better.)

This is perceived, by the MJ advocate, as being important, because it
helps to thwart the nefarious strategy of the extreme-rating voter,
against the voter who attempts to rate candidates proportional a
perception of their utility, regarded by the MJ advocate as more
honest, or even more ethical.

But extreme rating is widely understood to be, in many instances, the
best strategy, and the natural right strategy. And, because a
non-extreme rating might not have as much effect on a candidate's
final score, the fact that non-extreme rating _might_ be as effective
as extreme rating isn't enough to make non-extreme rating an adequate
substitute for extreme rating, where extreme rating is strategically
called-for.

So, then, what is the advantage of MJ over Score? Who knows.

What has been shown is the following:

1. MJ has a much longer and more elaborate definition, with its
tiebreaking bylaws, which require long wording in order to avoid
ambiguity.
......And the existence of several competing tiebreakers brings a
certain arbitrariness into those tiebreaking bylaws.

2. In MJ, SFR is more complicated, compared to that in Score or Approval.
.......In fact, with MJ's SFR requiring probabilistic voting, Score is
the only member of {Score, Approval, MJ} that has non-probabilistic
SFR.

3. Among {Score, Approval, MJ}, Approval has by far the simplest,
easiest, and therefore most fraud-secure count.

Mike Ossipoff