[EM] DMC the greatest Condorcet? I wish, but doubt it...

[Jobst Heitzig]

Hello Warren again.

You wrote:
> **First, I hear Jobst Heitzig himself (DMC's inventor) 

Sorry for the misunderstanding. I understand my support of DMC led you to think I discovered that method, but that's not at all true. 
To my knowledge Forest, Russ and Ted had the largest share of its development.

> claims his favorite method
> is something else (DFC, whatever that is).  !!??

Yes, of course. The reason I discussed DMC recently is simply that currently it is a big issue which *Condorcet-efficient* method will be suggested for Washington state's electoral reform. So I suggested to use my favourite *Condorcet-efficient* method DMC.

Still, my favourite method is not a deterministic Condorcet method but a non-deterministic Approval/Condorcet-hybrid called DFC, which is however strongly related to DMC but it not Condorcet-efficient. Feel free to search the archives for DFC. In short, DFC is simply this: Find the set of candidates which are not doubly defeated (in the same sense as in DMC), then let a randomly chosen voter select one of them (instead of electing the Condorcet-Winner among them).

> Fine.  But:  DMC, the supposed "unifying new champ", still appears to lead
> to 2-party domination.  I want you to modify DMC to get it out of this
> trap, or admit you cannot.  Otherwise, it seems to me, DMC has failed to
> live up to your hype.  (Also, the "DH3 pathology" - I want you to get DMC out of
> that one too, please:   http://math.temple.edu/~wds/crv/DH3.html .)

I will try to understand why DMC should lead to 2-party domination.
As for the DH3-example: Could you explain why the A voters should start to vote A>D>B>C in the first place?

Yours, Jobst

______________________________________________________________
Verschicken Sie romantische, coole und witzige Bilder per SMS!
Jetzt bei WEB.DE FreeMail: http://f.web.de/?mc=021193