[EM] 3 or more choices - Condorcet

[Kevin Venzke]

Hi Ted,

>Hi Chris,

>
>You discuss Winning Votes vs. Margins below.
>
>What do you think about using the Cardinal-Weighted Pairwise array in
>conjunction with the traditional Condorcet array?

I'm not Chris of course, but in my own simulations I found CWP (namely
the two-slot "AWP") to be one of the very best methods in terms of 

minimizing insincerity and electing sincere Condorcet winners when they
exist.

Some methods similar in format were also good. My own personal method
in this category is "Single Contest" 

http://wiki.electorama.com/wiki/Single_Contest

As I recall, "SC" was usually better than AWP, but in some settings the
compromise incentive was noticeable. (Neither of these satisfy FBC.)


Arguably the "low insincerity" aspect of such methods is a magic trick,
since the strategy is being relocated from the rankings to the approvals,
where they won't be called "insincere." But if having an approval 

component means voters succeed more often in electing sincere CWs then
I'm probably for it.


Kevin