[EM] Warren Smith writes big "best voting systems" paper

[Warren Smith]

The paper in its current form (it will change some, but it is pretty much done)
is available  in the form of an HTML web page:
   http://rangevoting.org/BestVrange.html

I warn you it is a math paper about 50 pages long.  But it has a
concise summary table and picture-plot, and you
can pretty much understand the results and ideas just by reading them
+ sections 0 and 1.

ABSTRACT
The only correct objective mathematical way to measure the "quality"
of a voting system is "Bayesian Regret" (BR). Contrary to common
mythology about Arrow's Impossibility Theorem, "best" voting systems
can exist under this metric. This paper

1. Evaluates the BRs of several voting systems (including Borda,
plurality, antiplurality, approval, and range voting) under a
mildly-realistic probabilistic setting, the "random normal elections
model" (RNEM), in closed form.
2. Identifies the best rank-order voting system. In the 3-candidate
case it is Borda voting. With ≥4 candidates it is new.
3. Proves that both "approval voting" and "range voting" equal or
better the best rank-order voting system, for both honest and
strategic voters (or any mixture) in 3-candidate RNEM elections. For
range voting this superiority is strict.
4. Identifies the best voting system based on ratings-type ballots in
3-candidate RNEM elections with honest voters. This new system has the
same vote-scores as range voting to within 7.5% but differs from hence
is even better than range voting. However, it only is better if over
≈91% of the voters are honest. For that and other reasons we do not
recommend it over plain range voting.

Range voting is: each voter awards a real score in [-1, 1] to each
candidate, and the one with greatest average score wins. Although the
paper is long, there is a concise summary table and plot in sections 7
and 8 respectively.
(END ABSTRACT.)

The paper is a breakthrough in the sense that no voting theory paper
ever accomplished 1,2,3, or 4 before, far as I know.

If you really are a masochist, there also is an extra "Discussion and
Open Problems" section which
is not currently included in the paper but probably will be.   It is here:
    http://rangevoting.org/BVR10.html
(Section 10.1 in this add-on is the thing that currently annoys me the
most about this paper and I'd really like
to be able to get rid of it.)

Comments welcomed... email them to me (says email address at start of paper)

-- 
Warren D. Smith
http://RangeVoting.org  <-- add your endorsement (by clicking
"endorse" as 1st step)
and
math.temple.edu/~wds/homepage/works.html