[EM] Bouricius reply, BR, Tideman, recent USA elections

[Warren Smith]

>>2. Bouricius forgot to mention, same way he usually forgets to
> mention, that Tideman also found IRV to be "unsupportable."
>
> *** 2. Warren Smith is wrong. He either hasn't read Tideman or is
> intentionally miss-representing Tideman here.

--I wrote a review of Tideman's book, remember?  And I cited one page
that said this in that review, remember?

Tideman said IRV was unsupportable if it is feasible to compute
pairwise matrix.  That was
because Tideman had other voting methods he considered clearly
superior to IRV and these methods used the pairwise matrix.    By
"clearly superior" I mean, so superior in every respect, that Tideman
felt there was no conceivable use for IRV, ever (in situations where
it was feasible to compute pariwise matrix) where that use could be
"supported."
That is what "unsupportable" means.

I'm not necessarily agreeing with Tideman -- in fact if you read my
review, I criticize some of what he says:
http://www.rangevoting.org/TidemanRev.html -- but that is what he
said.

> ***3. Warren Smith's conviction that Bayesian Regret is the gold standard
> for evaluating voting methods is not universally, nor even very widely
> held. It is a unique philosophical view held by those who subscribe to the
> Utilitarian philosphy, and is at least arguable. Many (most) people
> believe that when electing a single seat, the will of the majority should
> win out over the minority. This is necessarily rejected by the  believers
> in Bayesian Regret and advocates of Range voting.

--I cut & paste below, part of the puzzle-answer re Armytage, which
compares his measure
of "vulnerability to strategy", with BR:

k: Criticism of the "vulnerability to strategy" probability as a
yardstick for comparing voting systems: What matters to civilization
is not the probability that some voter-subset could strategize; what
matters is the expected utility-decrease caused by the net effect of
all strategic behaviors undertaken by all classes of voters. Some of
these strategic moves will reinforce, whereas others will "cancel
out." Certainly it is utterly unrealistic to pretend the pessimal
voter subset colludes based on perfect information about the
noncolluders – while all other voters (not in that subset) forswear
all strategy! And it is also utterly ludicrous to pretend all
strategy-caused winner-changes are "equally bad!"

(In short, the right yardstick for measuring vulnerability to
strategy, as well as lots of other things too, instead is Bayesian
Regret.)

[Also] If strategic voters have to act on imperfect information, the
situation also changes. For one thing, they might try a strategy which
fails to work out for them, but still causes damage – and such damage
ought to be counted, not ignored.

--second: Bouricius misleads when he says believing in BR contradicts
believing in the "will of the majority."   Well.   It is certainly
true one can set up situations where the least-regret choice and the
majoritarian choice differ.   But I hardly think the two are
diametrically opposed.  It's more like "they usually agree in
situations where majoritarian is defined, but when they do not, the
least-regret choice would be the better way to go (and it is always
defined)."

Incidentally, IRV also can contradict "the will of the majority" and
probably more often than BR does.  In particular, the IRV mayoral
election in Bouricius's hometown, recently did that, see
http://www.rangevoting.org/Burlington.html
and of the three most-watched USA elections this year, there were two
cases where IRV
(would have) drastically contradicted the will of the majority, i.e.
eliminated the apparent Condorcet winner.   See
http://www.rangevoting.org/Races2009.html

In all these elections it appears likely range and approval voting
would have behaved
reasonably, unlike what unfortunately actually happened,and also
unlike how IRV would have behaved.   It looks plausible Condorcet
would also have behaved reasonably
(although less evidence for that) in all these cases.


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