[EM] Proportional Representation beyond STV

James Green-Armytage jarmyta at antioch-college.edu
Mon Jul 7 12:56:02 PDT 2003

Dear Rob Speer,

	The sexiest proportional representation formula I know of is Nic
Tideman's CPO-STV. That is, Comparison of Pairs of Outcomes by Single
Transferable Vote. This system is a real STV-Condorcet blend. It works on
the same principle as Condorcet (and you can use whichever cycle-breaking
method you prefer), but instead of just comparing candidates to each other
in pairwise competitions, it compares complete outcomes with each other,
that is, possible outcomes for the entire set of seats. The computational
cost of this is somewhat forbidding, but if it is in fact as good of a
system as it sounds, I think that it should be worth it for some purposes.
Also, I imagine that there are various shortcuts that can be used to
significantly reduce computational cost. 
	Anyway, for a full description of the method, I encourage you to read
Tideman's paper "Better Voting Methods Through Technology", which can be
found at http://www.econ.vt.edu/tideman/rmt.pdf
among other places. It has a reasonably short description of the method
along with some informative background about the history of STV theory.
	In general, I understand that this list is dedicated to single-winner
systems. However, I am quite interested in discussing multiple-winner
systems as well. So I ask the more long-standing list members: what do you
think is the best venue to discuss PR? Is there a PR sister list for this
one? Should there be? Is there an existing PR list of the same calibre, or
with some overlapping members?
	I hope that you will find CPO-STV interesting, Rob. I sure do. For me, it
kind of seems like CPO-STV is to STV roughly as Condorcet is to IRV. (So,
of course if you don't like Condorcet, then I doubt that CPO-STV will
appeal to you.) Of course, I would be interested to know if any Condorcet
fans can poke holes into CPO-STV; show examples where it produces a weird
outcome. The little tests I have done have turned out well, but it is of
course very difficult to do serious tests on a method like that using pen
and paper. 

all the best,

P.S. I post my messages to 
election-methods-electorama.com at electorama.com
	Did you sign up to this list via electorama.com, or from some other
direction? I don't really know how it all works, but if I were you I'd try
looking at the electorama.com web site to see if that offers any clues as
to how to post.

More information about the Election-Methods mailing list