As part of coding my simulator, I have implemented a Condorcet method
that goes like this:
A candidate X's victory score is equal to the sum of all victories of X
against other candidates. If the pairwise matrix is d, and d[A,B] is the
number of voters preferring A to B, then the victory of X against Y is:
If WV, 0 if d[X,Y] <= d[Y,X], otherwise d[X,Y]
If margins, max(0, d[X,Y] - d[Y,X]).
The candidate with the greatest victory score is the winner.
-
I've called this method "Offensive L-R", because it's like
least-reversal, only instead of awarding the first place to the one with
the least defeat sum ("defensive"), it awards it to the one with the
greatest victory sum ("offensive").
Does the method have a name already? It's so simple I can't but think
it's already been invented.
----
Election-Methods mailing list - see http://electorama.com/em for list info
