[EM] equal rankings not defensively needed in Condorcet

[MIKE OSSIPOFF]

I've just realized that yesterday, when I acknowledged that
equal rankings could be the best strategy in Condorcet with
0-info, and when I said that I hadn't really taken a good look
at equal rankings--that might seem to discredit what I've
already said about Condorcet meeting SDSC.

Actually there are at least 3 reasons why that isn't so:

1. Blake's discussion was about a 0-info situation. Never happens
in public political elections. And even committees should have
discussion before voting, and then there won't be 0-info.

2. Blake's argument used Plain Condorcet. With just 3 candidates
, all the Condorcet versions give the same answer, includiing
Plain Condorcet. But with more candidates, that argument would
no longer be valid.

3. Because all the Cycle Condorcet methods meet SDSC, a majority
who (in the example, which I believe was from Bart) prefer
B to C can ensure that C can't win merely by not ranking C.

Of course in that 0-info example, we don't know where there's
a majority. No matter. Supppose that it's true that it's in
your best interest to vote in 1st place position all the candidates
with utility over the mean, as Blake suggested. I don't think
anyone knows, but suppose that's true with 0-info.

The goal in the discussion so far about that example is to make
C lose. Likewise with the defensive strategy criteria.

Ok, say you use that strategy of ranking in 1st place every
above-mean candidate. If the count will by by a Cycle Condorcet
method, then the only reason why you're doing that would be
in case there isn't a majority ranking some same candidate
over the one you want to defeat. In other words, you aren't
doing it as defensive strategy.

Defensive strategy is what you have to do to ensure that a
majority gets something it wants. With any method except Borda,
a majority can easily make a candidate win just by ranking
him 1st. But the more difficult thing is making a candidate lose,
something that many voters here are obsessed with doing.

So the real defensive strategy problem is what a majority has
to do in order to make someone lose. Since you aren't using
Blake's stratgegy to defeat someone who has a majority ranking
some same candidate over him, because you don't need to--
since you aren't doing it to enforce a majority wish--then
you aren't doing it as a defensive strategy. No need to.

I don't know if "offensive strategy" has meaning with 0-info,
so let's just say that you're doing non-defensive strategy.
You're trying to maximize a candidate's chance of losing
(or the chance of all the sub-mean candidates of losing),
in case they don't have a majority pairwise defeat, in case
a majority isn't trying to get that defeat. Fine. I don't have
a problem with what we have incentive to do non-defensively.
Maybe I'd vote the strategy Blake suggests, in a 0-info
election with Condorcet, but I don't regard that incentive as
a problem for Condorcet.

Mike Ossipoff

______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com