[EM] Simulation of Duverger's Law

[Greg Nisbet]

I had an idea to test how much particular methods enforce two party
domination.

Start out with some voters, each of them has some utility for a certain
candidate. Utility = reproduction in this case. So voters with a positive
utility of the outcome will reproduce and voters with a negative utility of
the outcome will die off. The candidate generator will also be fairly cloney
so that clone positive methods don't take over. Each voter has a tag that is
heritable. You can measure the trend towards two party domination by
counting the number of voters with a particular tag and graphing it.

There are some things to be sorted out such as how to prevent population
explosion after two party domination and how to balance utilities so that
voters who like everything don't show up.

Complete non-sequitur but still a point I don't entirely understand:
IRV, FPTP and Contingent Vote all lead to two party domination according to
Duverger's law.

Why doesn't France's Two Round System lead to the same result? Pretend you
have a ballot consisting of rank ordering and a separate FPTP checkbox,
would this similarly avoid two party domination?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20081015/920ff73b/attachment-0002.htm>