At 12:40 PM 11/26/2008, Kristofer Munsterhjelm wrote:
I want to add to this by saying that if Approval is about approval,
well, then discussions about frontrunner plus strategies won't
capture the intent or point of the method. If the statement for
Approval voting is "vote for those you like", or "vote for those of
which you approve", then one should expect voters to do that, absent
strategic incentive. Say there's a certain group of people that a
voter approves of. If he has to plan beyond that point, then that's
strategy. On the other hand, if Approval really is "pick those
candidates you like more than or equal to the frontrunner you like
the most", then there's not much approval-ish about the method, in
the ordinary sense. It asks the voter to optimally configure his
ballot. If we're going to do that, we should leave the task to a
computer and use DSV instead.
Tell me, how important is the name of the method? If a good method
has a bad name, how much damage will that do, long-term? If a bad
method has a good name, or a misleading name, what will happen in the long run?
"Approve" is undefined, except in relation to a candidate set and
probabilities. Whether I approve of some possible outcome, in a
situation, depends on what I think is reasonably possible. I'll
"accept" $10 for something if I don't think I can get $100. But
surely I approve of both, in that context, whereas if I think I can
get more than $10, I won't approve it.
The seriously bad assumption being made, over and over, is that there
is some absolute relationship between candidates, in themselves,
without any consideration of alternatives and probabilities, which is
"Approval."
A very good candidate, I might not approve, if I think I can get
better and there is little risk of getting worse!
Now, the strategy suggested above is fairly simple. Identify the
frontrunners. *Of that set," which one do you prefer to accept?
That's the real choice being presented, and a voter who does not vote
that way is very, very likely to waste his or her vote. It is the
same with Plurality! The same choice.
Approval was thought of as some new method, and enthusiastic
supporters liked the idea of imagining electing candidates with wide
"approval." In fact, it's a very old method, and it is very similar
to Plurality, but with a very small, simple, tweak which fixes a
major problem with plurality, the same problem that makes many long
for IRV, the spoiler effect.
Most voters, by definition, don't need to worry about additional
approvals. They simply vote for their favorite frontrunner. If you
are voting for a minor candidate, and you don't know that your
candidate is a minor candidate, you are one uninformed voter. So
those who might need additional approvals know they need them, and
they decide, with Approval, whether or not the care enough about the
difference between the frontrunners to vote in that pairwise contest
by approving one of them.
And with Bucklin, of course, they get to vote preferentially for
their favorite *and* add additional Approvals if needed. This isn't
some new idea, it was extremely popular in the U.S. for a time early
in the last century. It was never rejected by the public, to my
knowledge. Rather, the usual sorry story....
I'd like to dump the name "Approval Voting." "Open Voting."
The system is "open" to voters voting freely, with the vote-for-one
restriction. Voters still have one effective vote to cast, never, in
this system, can they contribute more than one effective vote. (If
they happen to do it, i.e., they vote for real frontrunners, both of
them, they have, as it were, abstained though they do increase the
vote percentage for the winner by one vote. They simply did not
participate in that particular choice, except to help create a
majority if that is needed, for either. It only matters, though, with
the winner. All other votes are moot, they could be striken from the
ballots and the outcome would not change.
Open Voting is like the Alternative Vote (another name for IRV)
except that the alternatives are equated and simultaneously
considered. In the end, only one vote is effective.
----
Election-Methods mailing list - see http://electorama.com/em for list info
