[EM] Introduction (cont.)

[Bart Ingles]

I skimmed Warren D. Smith's paper, but didn't have time to go into it
too closely.  A couple of initial reactions:

1)  I have a problem with referring to strategic or tactical voting as
"dishonest voting."  It leads to the interesting situation in which the
author pits "honest voters" against "rational voters" -- implying that
honest voters are irrational, and that rational voters are dishonest.

2)  Range voting generates ballot information which can be interpreted
in a number of ways, which may tend to undermine the legitimacy of
results.  It's not hard to imagine a post election scenario in which
candidate A wins with the highest ratings total, but supporters of B run
around saying, "yes, but the majority of voters chose B over A."

3)  Range voting and approval voting are logically equivalent.  This
equivalence has already been pointed out w/r strategic voters, since the
rational (strategic) voter would always cast votes at either the minimum
or maximum of the allowed range (unless absolutely indifferent).  What
hasn't been pointed out recently is that the systems are also equivalent
regarding so-called honest voters.

Under approval voting, a voter who wishes to give 1/2 vote to a
candidate can do so in two ways:  (1) find a like-minded voter, and
arrange for one voter to give a full vote to the candidate, while the
other voter gives no vote, or (2) toss a coin, and vote to approve or
not approve based on the outcome of the coin toss.  Assuming other
voters use the same strategy, this should be approximately equivalent to
all such voters giving 1/2 vote each.  For vote strengths other than
1/2, you can replace the coin toss with a roll of a Parcheesi die, but
keep in mind that any voter who wants to give more than 1/2 vote should
probably give a full vote (and less, none).

While this may seem cumbersome, I would expect honest/irrational voting
to be the exception rather than the rule (or at least I would prefer it
so).  I think it would be preferable to encourage indifferent voters to
do the additional research required to form an opinion for or against a
given candidate, than to give them the "easy out" of casting a partial
vote.  

Bart



Douglas Greene wrote:
> 
> Warren D. Smith of the NEC Research Institute modeled 120 or so voting
> systems with both sincere and strategic voters.  He found that range
> voting produced the least Bayesian regret as compared to all the other
> systems.  His work is available at:
> 
> http://www.neci.nec.com/homepages/wds/works.html  Scroll down to
> number 56, and it's all there.