[EM] Approval vs ??
MIKE OSSIPOFF
nkklrp at hotmail.com
Thu Mar 25 15:29:06 PST 2004
Much of the criticism of Approval is done in a vacuum. I always want to
reply:
"...So then, Approval is worse than....what?"
IRV?
James cited Mutual Majority (MMC) as Approval's failing. But the mutual
majority situations where MMC applies are situations where IRV demonstrates
its failure of WDSC & FBC, two criteria that Approval and CR pass.
Sure, when there's a mutual majority, the members of that mutual majority
benefit from it in IRV. But, for a particular example, are we interested in
how well off the most fortunate voters are, or aren't we more interested in
how bad a strategy problem some voters can have? The fact that IRV, when
there's a mutual majorilty, makes some voters better off, and some worse
off, than they'd be in Approval, doesn't mean that IRV is better than
Approval.
Some argue that the voters who aren't in that mutual majority have nothing
to complain about if their candidate can't win. Of course, but they have
strategic incentive to bury their favorite. Because their candidate can't
possibly win, their best strategy will often be to bury their favorite by
insincerely voting one of the mutual majority candidates to 1st place, to
avoid the election of a worse one.
The problem with that, of course, is that we don't have perfect information.
Sometimes people will mistakenly bury their favorite, believing that it's a
situation like that described in the previous paragraph. And isn't that the
old problem of Plurality? Isn't that what we want to get away from, with new
voting systems?
With Approval, no one ever has any incentive to bury their favorite by
voting someone else over him/her.
Then is Condorcet better than Approval? I prefer Condorcet to Approval. But
now you're holding Approval up to an unreasonably hilgh standard. What do
you want from the simplest, most easily proposed, most easily-implemented
better voting system? It's unrealistic and unfair to demand that Approval be
as good as Condorcet.
I prefer Condorcet to Approval because Condorcet receives and counts all of
the pairwise preferences that anyone wants to vote. That's why Condorcet is
better at electing the CW, and meets SFC, GSFC & SDSC. And it's why
Condorcet does better at SU (Though Approval also does very well at SU,
significantly better than IRV).
Approval has its particular advantages though. By its limiting of the voting
of preferences, Approval lets the voter make a statement that a ranking of
all the candidates doesn't make. Some candidates are good enough to vote
for, and every Approval voter is making a strong and unmistakable statement
about the difference between the ones s/he votes for or doesn't vote for. Of
course you could make a similar statement by refusing to rank some
candidates in Condorcet. But it's easy to rank a candidate in Condorcet, it
doesn't make a very strong statement about someone's merit, that you include
him/her in your ranking. Approval makes stronger statements. Though
Condorcet allows better expression in one sense, then, there's also a sense
in which Approval allows, and requiers, stronger expression.
Though Condorcet better maximizes SU, Approval has its own social
optimization: It's been shown that, when people vote strategically in
Approval (and Approval very much encourages strategic voting), Approval
maximizes the number of voters who are pleasantly surprised by the outcome.
The number of votrers for whom the result is better than what they expected
from the election. I'm not saying that that's better than SU maximization,
only that Approval still has a social optimization.
Approval isn't Condorcet, and it isn't a fair comparison. But Approval is a
strong 2nd. And Approval is remarkably good for such a simple method, such a
minimal change from Plurality, such an easily proposed and implemented
method.
Mike Ossipoff
_________________________________________________________________
Get reliable access on MSN 9 Dial-up. 3 months for the price of 1!
(Limited-time offer)
http://join.msn.com/?page=dept/dialup&pgmarket=en-us&ST=1/go/onm00200361ave/direct/01/
More information about the Election-Methods
mailing list