[EM] Sports and 'The Condorcet Mindset'

[Michael Rouse]

----- Original Message -----
From: "Alex Small" <asmall at physics.ucsb.edu>
To: <election-methods-list at eskimo.com>
Sent: Sunday, November 17, 2002 11:59 PM
Subject: Re: [EM] Sports and 'The Condorcet Mindset'
> However, whether or not the things that Saari points to are sufficient for
> picking a method is another story.  Let us assume for a moment that all of
> the geometric properties favored by Saari are socially desirable.  If
> voters are honest then Borda is a socially desirable method.
>
> However, because Borda gives huge incentives to vote insincerely in almost
> every race, Borda is unable to achieve its stated advantages.  Granted,
> all methods are prone to strategic voting, but some are more prone than
> others.

I agree. If we were using several machine inputs and needed to come up with
a decisive output -- like in OCR or pattern recognition -- Borda would be
great because machines wouldn't have any reason to lie. Once you throw
humans into the mix and give them various punishments and rewards depending
on who wins, then you get people with a sincere top choice and then a list
of candidates in reverse order of competitiveness. This would be even worse
if parties got into the act and started putting out voter sheets -- if a
first choice did not reach 50% of the vote, the election could be chaotic,
with a "winner" that would lose every 1 on 1 contest.

That's one of the reasons I like the mean Kemeny order. Sure, it's hard
(NP-hard, actually) to figure out the winner this way, but it's even harder
to successfully manipulate it by insincere voting. It obeys the extended
Condorcet criterion without getting caught in circular ties, and has been
used to fight against spam websites where the owner tries to manipulate page
rank "votes" in order to move his wesite up search engine "candidates" list
(see http://www10.org/cdrom/papers/577/ ).

Michael Rouse

----
For more information about this list (subscribe, unsubscribe, FAQ, etc), 
please see http://www.eskimo.com/~robla/em