[EM] Re: Combating the Approval Burr dilemma

[Kevin Venzke]

Hi,

--- Abd ul-Rahman Lomax <abd at lomaxdesign.com> a écrit :
> At 05:10 PM 11/16/2006, Kevin Venzke wrote:
> >If the method
> >offers voters incentive to behave in such a way that when their faction
> >is represented by multiple candidates, this faction is penalized, then
> >as a result, political parties will not want to run multiple candidates
> >appealing to the same portion of the electorate.
> 
> Probably true under just about any political system that is 
> party-organized. Running multiple candidates is *expensive*. And 
> you'd better know what you are doing.

Probably, but incentives can exist within the mechanics of the method,
also.

> The problem here is not the election system itself, in the example 
> chosen, but that the party is not actually united on the two 
> candidates. It is as if there are two parties.

The concern is that they could be united (looking at sincere preferences),
but some voters perceive that they will get a worse result if they
vote for both candidates, than if they just vote for one.

So what the designer of an election method wonders is whether things can
be set up so that this perception (that defection will get a better
result) is never true.

> Approval will not solve *this* problem, if the members of the 
> factions refuse to sincerely approve the other party member, perhaps 
> trying for the election of their own and perceiving that other party 
> candidate as a threat to that.
> 
> But this is voluntary behavior on the part of the voters, 
> individually, in the voting booth, and simply demonstrates that the 
> party failed to nominate consensus candidates.
> 
> If I'm correct, the desired property is Independence from the 
> Insertion of Clones, or something like that. Right?
> 
> A party which is two factions deeply divided is not a single party, 
> no matter what it calls itself. It's a coalition.
> 
> IIC must require that voters vote sincerely. But in the example 
> given, they don't vote sincerely.

Clone independence isn't enough. For Approval and Schulze the relevant
property is Later-no-harm. For example

49 A
24 B>C
27 C>B

Elects C under Schulze, and B or C under Approval. But

49 A
24 B
27 C>B

Elects B under both methods, so that there is some incentive (that has
to be weighed, of course) for a faction to not support the second
preference, hoping the second preference's supporters will still support
the first preference.

This could be important to the result if C is considered an unlikely
winner, so that B voters expect that C voters see they have to vote for
B. Under this assumption the B voters can only throw the election away
by giving a second preference to C.

Kevin Venzke


	

	
		
___________________________________________________________________________ 
Découvrez une nouvelle façon d'obtenir des réponses à toutes vos questions ! 
Profitez des connaissances, des opinions et des expériences des internautes sur Yahoo! Questions/Réponses 
http://fr.answers.yahoo.com