[EM] Trying to protect even weak Condorcet winners

[Stephane Rouillon]

The last message I sent you about three month ago if I remember well
when you pop out back of retirement summarized all this.

I agree about winning-votes property for any strong Condorcet winner.
I just added that treated with an extended graph {G,Z} where Z represents
a virtual candidate for the acceptable candidates limit (we suppose unranked
candidates are disapproved), then Ranked Pair with winning-votes seemed
to protect EVEN A WEAK Condorcet winner.

You said you would try to find a counter-example.
Have you?

Stephane

MIKE OSSIPOFF a écrit :

> The winning-votes Condorcet versions don't have truncation as an effective
> offensive strategy. With wv, voters can't steal the election, for their
> candidate, by truncation, from a CW who is voted over that candidate by a
> majority. The criterion SFC is a statement of this property, though SFC
> doesn't mention truncation, and SFC's applicability isn't limited to methods
> subject to truncation. For instance, IRV fails SFC, though truncation isn't
> an offensive strategy in IRV.
>
> But all the wv versions meet SFC: RP, SSD, CSSD, BeatpathWinner, PC, SD.
>
> All of those except for PC also have been shown to meet GSFC.  I'd be
> interested if anyone has a proof that PC passes or fails GSFC.
>
> Do you advocate the wv Condorcet versions, as opposed to the non-wv
> Condorcet versions?
>
> Anything that you can show about the truncation-resistance of the wv methods
> would be of interest. If you sent something about that before, when I was
> away from voting systems, would you send it again?
>
> Mike
>
> _________________________________________________________________
> Groove on the latest from the hot new rock groups!  Get downloads, videos,
> and more here.  http://special.msn.com/entertainment/wiredformusic.armx