Paul Kislanko wrote:
A wins in the example ONLY because the method discards the C>A votes
True
because of the B>C>A set of ballots.
That is demonstrably false, as Steve has now pointed out twice.
If we're trying to find something better than plurality, it needs to be demonstrably better, and of course this example gives exactly the same results as plurality, which is why examples aren't proofs.
Indeed they are not, which begs the question of why you are asking about examples. Better to go to Steve's page where he proves properties of MAM in excruciating detail.
But, to use the terminology and techniques y'all do, let's examine the BALLOTS that result if B is not a candidate:
4: A>C 5: C>A
Adding B to the mix causes A to be elected, even though all voters who prefer B over anybody voted A third of the 3.
Congratulations! You've proven that MAM fails IIA. Big friggin' surprise there.
So I ask again, if A should win, why should I prefer any method over plurality?
It's up to you to answer that. But some methods satisfy properties that plurality fails, and I for one happen to value some of those properties.
-Adam
---- Election-methods mailing list - see http://electorama.com/em for list info
