[EM] Request for Multiwinner Methods

[Greg Nisbet]

For the Intel Science Talent Search, Warren Smith and I are working on a
system to measure how representative the groups created by various
multiwinner electoral methods are. This shall be done by having the public
have opinions of various binary social issues and the candidates will have
stance on these. Then the group of winners will get together and vote (using
the only reasonable voting method with two options) and this will be the
basis for seeing how well the winners have "emulated" society.

Several things to note,
a) This represents a departure from the tradition concept of utility
somewhat. The candidates themselves won't have utility (they may have
something resembling it, but that comes much MUCH later) instead the net
opinions of parliament will.
b) In order to simulate strategic voting, either the Vote By Result system I
described earlier will be used or one based on assigning "victory
probabilities" to candidates in a massive feedback loop that will eventually
approach equilibrium. More on this later

Anyway, onto the main point.

If you have a multiwinner method of some sort, that would be great.
Party-related methods are great, but as we aren't exactly sure how to
emulate parties yet.. they are less useful now than they would be, say in a
month, that's no reason not to suggest a nice party method, but just sort of
keep that in mind.

Both ranked and rated ballot methods are appreciated. If you have some
variation of an existing method that is great too.

Single winner methods are also appreciated. We are going to attempt
districted single winner methods as well for the purposes of this thing. The
number of single winner methods tested won't be quite as generous as the
other Bayesian regret tests, but whatever.

The current multiwinner methods I can think of off the top of my head are:
CPO-STV
STV with various transfer rules (I don't anticipate TOO great a difference
here)
RRV
PAV (proportional approval voting)
PRV (  "         "      range    "    " )
SNTV
MMP
Cumulative Voting
Sainte-Lague
Largest Rem
D'Hondt
Limited Vote
Block Vote
Sortition (Random Winner)

And, a special place of (dis)honor is reserved for distrticted FPTP. It is
the norm for Anglophone countries and a cause of political misery the world
over.

I know many of those methods listed are crappy, but I say some kind of
yardstick is called for.

Oh yeah, if your method is sufficiently obscure, please maybe give a brief
description of how it works or a hyperlink to one.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20081012/187c2037/attachment-0002.htm>