Re: electing multiple people
On Mon, 8 Oct 2007 23:41:41 +0200, Josip Rodin <joy@entuzijast.net> said:
> On Mon, Oct 08, 2007 at 02:54:18PM -0500, Manoj Srivastava wrote:
>> I think you misunderstand. The graph is a perfectly clear
>> representation of the pairwise defeats as it corresponds to the
>> single winner outcome; and it shows the relative strengths of the
>> defeats. There is nothing wrong with it -- as long as you do not try
>> to project and use that data for multi-winner predictions. All the
>> candidates are rpesented, and the relative pairwise defeats are
>> presented -- quite correctly. Assuming that the pairwise defeats
>> have something to do with global ordering is a leap, however, and not
>> one which is promoted by the web page.
> Uh, yes, it is. The picture clearly orders candidates from top to
> bottom, so it's a fair assumption. But never mind.
You do have a point. Based on the pairwise defeats, one can get
a sense of how candidates might have ranked --- and the ranking so
uncovered is likely to be correct in absence of circular defeats (in
other words, if the second place race did not have a clear [condorcet]
winner).
The picture gives, in other words, a ranking that conveys
information, is often not far from the truth, and thus, useful
interpretation of the beat matrix, and yet is still unsuitable for more
formal use cases -- like deciding how to order N winners, where N > 1.
Anyway, I'll try and place the winner first on the next set of
updates I do.
manoj
--
What matters is not the length of the wand, but the magic in the stick.
Manoj Srivastava <srivasta@debian.org> <http://www.debian.org/~srivasta/>
1024D/BF24424C print 4966 F272 D093 B493 410B 924B 21BA DABB BF24 424C
Reply to: