Kevin Venzke stepjak at yahoo.fr
Wed Sep 7 20:17:28 PDT 2005


--- Warren Smith <wds at math.temple.edu> a écrit :
> is ICA the same thing as Smith//Approval?  And if not, what is the matter with
> ICA?

It's not the same. Improved Condorcet Approval is Condorcet//Approval with a
tweak to eliminate favorite betrayal incentive. There's no effort to be
Smith-efficient. (I don't think such an effort could be successful.)


What's the matter with ICA? Let's see:

1. It is not 100% Condorcet-efficient. This could be a deal-breaker if you
insist on electing a Condorcet winner no matter how iffy some of his pairwise
wins are.
2. It makes no effort to be Smith-efficient, relying on Approval instead.
3. Cloning the winner, such that a cycle is created, can move the win to 
someone outside the clone set, if none of those clones are the approval
4. I don't think it works well with an explicitly-placed approval cutoff.
I think it makes burying strategy too unlikely to backfire, if a voter is
not obligated to approve a candidate he potentially uses solely to prevent
another candidate from being the decisive (Condorcet) winner.
5. The FBC tweak, and why it works, is not the easiest thing to explain.

Kevin Venzke


Appel audio GRATUIT partout dans le monde avec le nouveau Yahoo! Messenger 
Téléchargez cette version sur http://fr.messenger.yahoo.com

More information about the Election-Methods mailing list