[EM] Simulation of voter preference of imethods' results, for at least 10 methods?

[Michael Ossipoff]

...ideal majoritarian methods, mostly., though I'd like to also
include Benham and Woodall.

As I've mentoned, Steve Eppley did a simulation in which he compared
Beatpath and MAM. He compared them in a few ways, but, of most
interest, he compared them for which method's results were preferred
by the voter,  in the most instances in which MAM and Beatpath gave
different winners. MAM won overwhelmingly in that regard.

It would be of interest to find out how MAM-River does against MAM, in
such a comparison.

Also, how CIVS-RP and MMT, in its two vesions, do.

It would be good to compare the following methods:

A. Ideal Majoritarian Methods:

   1. Ranked-Pairs
......a. MAM
......b. CIVS-RP
......c. MMV, Unconditional
......d. MMV, Condiional

   2. The same RP methods, but
.....modified for River

   3. Beatpath

   4. Maybe SSD

B. Green Scenario Methods

...1. Benham

...2. Maybe Woodall

---------------------

That's 12 methods. SSD doesn't really have advocacy, but, just to be
sure, it would be good try it. Benham is a lot more proposable for
Green Scenario public elections than Woodall is. It's barely diferent
from the IRV that's proposed by at least 5 U.S. political parties'
platforms. Benahm has an excellent chance of being the Green senario's
voting system, if the Green scenario (with Greens, Justice Party, or
any other progressive party) ever comes to pass.

Would anyone be interested in doing such a simulation? Probably best
to use randomly normally-distributed voters and candidates, or maybe
make the candidates' positions have some resemblence to those of
Republocrats, far from the voter median, and lots of progressive
parties closer to it.

Because more candidates increases the computation time more than more
dimensions does, then maybe have only about 5 or 10 candidates, but at
least 10 or 15 dimensions--to get lots of cycles, without increasing
the computation time as much?

For the same purpose, no correlation between dimensions, to get as
many cycles as possible, and as many diveregences of outcome as
possible.

Then, maybe, later, spending more time, maybe, more realistically,
just 3 issue-dimensions, and more elections.

Is anyone else curious about how various methods would do in that
regard? Because voting systems are used for polling and organizations,
at least the ideal majoritarian methods are surely of great interest
at EM.

Michael Ossipoff