[EM] RE: Election-methods Digest, Vol 15, Issue 37

[Abd ulRahman Lomax]

At 04:45 PM 9/21/2005, Kevin Venzke wrote:

> > It is unclear why Mr. Venzke's presence would alter the result
> > unfavorably, unless he opens his mouth and nastiness comes out in
> > such a way as to alter people's perception of the otherwise-winner.
>
>Let's say I'm candidate B and I receive 33 votes, so that at the convention
>these preferences are represented:
>
>33 A>B>C
>33 B>C>A
>33 C>A>B
>
>In the absence of additional information, we have to suppose that the odds
>of each candidate winning are equal. But perhaps my supporters have a
>sentiment something like: B>C>>>>>>>>>>>A so that they'd rather not risk
>electing A. In that case, they are better off giving their 33 votes to
>the C>A>B candidate.

Ah, I thought he meant that the result would be unfavorable to *him*. 
My understanding of the role of the proxy is that the proxy is not 
charged with making the decisions the voter would make, necessarily, 
but rather the decisions that the *proxy* makes based on his or her 
own best judgement. That's the task.

I would assume, however, that if all the supporters of B truly hated 
A, then B would not be giving his votes to A! Instead, B could give 
his votes to C and C would win handily. Or A could give his votes to 
B, would be unlikely to give them to C. The preference orders are 
pretty strange, by the way. I would not expect to see orders like 
this, unless the preferences are about pure personality, not issues.

A voters second choice is B, B voters last choice is A
B voters second choice is C, C voters last choice is B
C voters second choice is A, A voters last choice is C

There is no top-2 approval winner in this election, it's a tie.

Condorcet analysis:
A:B 66:33
A:C 33:66
B:C 66:33

There is a Condorcet cycle, there is no Condorcet winner. Who would 
Mr. Venzke have win this election? I'd say that to get a good result, 
deliberative process is necessary, the minimum process being a rerun 
of the election, but Asset Voting would be a better process, I'd 
suggest, for the candidates could actually negotiate with each other. 
Maybe they would find a way to deal with the weird polarization....

Election methods generally presume fixed opinions on the part of 
voters. Asset does not, except for an opinion as to whom to trust in 
determining the election outcome. Presumably B could communicate with 
his voters for advice, as well as negotiating with the other 
candidates. B could cause the election of A or C,   we know that B 
would very likely not want to vote for A. However, B might also hold 
out for his own election, saying to C, look, twice as many voters 
preferred me to you.

In the end, what is important in Asset is whom to trust, not so much 
whom to elect. If B is a good choice, B will presumably make a good 
decision! And if B decides to elect A, he will have some explaining 
to do! Perhaps he had a good reason!

> > I find people making the weirdest objections to proxy systems....
>
>Did I make an objection to proxy systems? I was only writing about FBC
>potential of proxy systems.

Okay. Proxy systems, because they incorporate deliberative process, 
actually step outside the realm of "election methods," which 
generally presume a fixed vote by voters. In a proxy system, the 
voter is deciding whom to *trust*, not necessarily whom to *elect*. A 
very important distinction. So election criteria don't strictly apply 
in the same way.