Re: [EM] Best Copeland Version ..

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view

Re: [EM] Best Copeland Version ..

Forest Simmons
A small change to make the method more symmetrical while making ties less likely:

Elect the alternative X with the greatest value of the difference D(X)  given by

diameter[BBOT(X)] - diameter[DOT(X)],

where the first bracketed set is the set of alteratives beaten by or tied with X, and the other one, those defeating or tying  X (pairwise in both cases).

This "Copeland score" should be a good one for making a social order with nice properties.

Election-Methods mailing list - see for list info