Kevin--
You wrote:
The problem was an egregious failure of Plurality: n a 1 a=c 1 b=c n b When
n>=1, C is elected, no matter how large n gets.
I reply:
So a thouand each of A and B voters would make each other lose, with the
help of those two other voters, who let C have only 1000 votes against it
instead of 1001.
As I probably said years ago: Even though I forgive Plurality failure, that
result is a little too much to forgive, if A is favorite to a million, B is
favorite to a million, and C shares first choice for two voters.
Thats regrettable, because MMPO was my favorite brief proposal, because of
its super-brief definition. (Results-wise, I liked MAMPO, MDDA, & MDDB
better, because I knew that MMPO had a problem, though I didnt remember
what it was. But, because of its briefest definition, and consequent good
winnability, MMPO was my favorite alternative simple rank method. But I have
to reluctantly abandon it because of that completely inexcusable behavior
that you pointed out.)
Mike Ossipoff
----
election-methods mailing list - see http://electorama.com/em for list info
