[EM] Re: Empirical data on cycles

[Kevin Venzke]

Rob,

--- Rob LeGrand <honky1998 at yahoo.com> a écrit :
> Kevin, could you post an example that gives a favorite-betrayal
> incentive under margins but not winning-votes?

I believe this works:

49 A>C
24 B
27 C>B

Margins elects A. If the 27 C>B voters reverse this to B>C, then B
is the Condorcet winner. It's inadequate for them to vote C=B.

Under WV, B wins in the first place.

If we suppose that the A>C voters want to at least elect C under WV,
it is sufficient for them to vote A=C.

The specifics of my simulation were: One randomly-sized faction first
voted A=B. Then the election was run again with these ballots changed to
A>B. If this changed the winner from someone outside {a,b} to A or (in
a non-monotonic method) B, this trial counted as a failure.

Kevin Venzke



	

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