Ken,
You wrote to Chris:
>Here's a more explicit illustration of the scenario. There are 10
>Candidates A ... J with the following CP's:
> A: +0.9, B: +0.7, C: +0.5, ... J: -0.9
> (<-- liberal ... conservative -->)
>There are 100 voters with the following sincere CR profiles:
> 51 voters (liberal): A(+0.9), B(+0.7), C(+0.5) ... J(-0.9) (rank: A
> > B > ... > J)
> 49 voters (conservative): A(-0.9), B(-0.7), C(-0.5), ... J(+0.9)
>(rank: J > I > ... > A)
> avg CR: A(+0.018), B(+0.014), C(+0.010), ... J(-0.018) (A wins CR)
> Approval: A(51), B(51), ... E(51), F(49), ... J(49) (A...E tie
>Approval)
> Plurality: A(51), B(0), C(0), ... I(0), J(49) (A wins Plurality)
This model is lunatic. Anybody with a positive weight for the issue is going to
have the "sincere CR profiles" (+0.9), B(+0.7), C(+0.5) ... J(-0.9). The
"sincere CR winner" is always going to be either the furthest left or furthest
right candidate.
This is not realistic unless you think voters are all extremists and only the
candidates show moderation.
I would suggest that you abolish voter "weights" for issues, and implement
voter positions on issues as you have for candidates.
Kevin Venzke
[EMAIL PROTECTED]
__________________________________
Do you Yahoo!?
Yahoo! Domains � Claim yours for only $14.70/year
http://smallbusiness.promotions.yahoo.com/offer
----
Election-methods mailing list - see http://electorama.com/em for list info
