[Condorcet] RE:[EM] MinMax(ao)

Kevin Venzke stepjak at yahoo.fr
Sat Sep 10 09:45:27 PDT 2005


--- "Simmons, Forest " <simmonfo at up.edu> a écrit :
>>> Yes, this is the measure of defeat strength used in AWP, but here's the question:  Is AWP
>>> on River the same as AWP based on MinMax?  
>>I think the answer for AWP is no. All these methods using WA are the same probably
>>because any given candidate's wins are all of the same strength. That isn't true
>>with AWP.
>I'm more interested in whether or not MinMax(AO) is clone proof like MinMax(WA) or fails clone
>winner like the better know MinMax versions MinMax(wv) and MinMax(margins).

Hmm... I couldn't come up with a failure example, but I did find something
rather fishy about AWP/AO:

55 A>B|
45 B>A|

The A>B win is of strength 0. So unless the AWP method being used either
1. treats "no losses" as better than "no losses of positive strength" or
2. explicitly finds the Smith or Schwartz set,
This election is a tie!

Solution #1 doesn't work for Minmax without sacrificing Clone-Winner, though. 
Using #1, A does win. But if A is replaced by a set of clones in a cycle, no 
candidate has "no losses," so that B becomes a potential winner.

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