Dear Eric Gorr,
you wrote (13 Nov 2004):I am not talking about the IIA variant that you prefer, but Dr. Arrow's variant (which does not involve probabilities - unless I am mistaken), so my question to you stands.
So when you want to use the resolute model for IIA, then of course you also have to use the resolute model for ICC when you want to compare IIA and ICC. How do you define ICC for the resolute model?
I would suggest the one proposed by Dr. Tideman which I believe can be found here:
http://condorcet.org/emr/criteria.shtml
---- Election-methods mailing list - see http://electorama.com/em for list info
