[EM] MCA strategy. MCA vs MTA?
MIKE OSSIPOFF
nkklrp at hotmail.com
Tue Nov 15 14:21:14 PST 2011
Someone said that, in MCA, you aren't significantly helping your top-rated candidates against
your middle-rated ones, because your top-rated ones might not get a majority.
I disagree. MCA and MTA provide the strong majority rule guarantees described in the definition of 3P.
If one of your top-rated candidates has majority support (as described in my 3P definition) against those you don't top-rate,
then no one you and that majority don't top rate can win. That's significant help.
In the earlier discussion of MCA strategy, it was correctly said that a mathematical strategy for MCA would
be difficult to find. That's true. For any method more complicated than Plurality or Approval, mathematical
strategy becomes hugely complicated, and requires many probability estimates.
But, when there are, in the election, some completely unacceptable candidates who could win, that greatly
simplifies strategy.
In that case, the best strategy for MCA is to top rate all of the acceptable candidates, and to not rate any
of the unacceptable candidates.
Would middle rating ever be used? Maybe sometimes there's a borderline candidate for whom it's difficult to
say whether s/he's barely acceptable. Someone who is good enough to use against the completely unacceptable, but
isn't good enough to help against the completely acceptable.
There was discussed an example like this (the numbers are utilities):
A 100 B 15 C 0
For whom do you vote in Approval or MCA?
Of course, if it's an Approval election, with zero-information, then you vote
only for A. With 0-info, you should vote only for the above-mean candidates.
The mean in that lineup is 115/3 = 38 + 1/3 Only A is above mean.
In MCA? Of course one can't really say, since we don't have a mathematical strategy.
The example was intended as zero-information, wasn't it?
All I can say is that I wouldn't regard a 15-utility candidate as much of a lesser-evil. If C
were completely unacceptable, and if b weren't, then B's utility would be more
than 15, to C's 0. So then I'd say that B is probably completely unacceptable,
along with C.
So most likely I'd top rate A, and bottom-rate B and C.
By the way, what strategy protection difference is there between MCA and MTA?
They differ only in what they do when two or more candidates are majority top rated.
MCA then elects whichever of those is top rated by the most people.
MTA elects whichever of those is top or middle rated by the most people.
Which of those is better?
Mike Ossipoff
More information about the Election-Methods
mailing list