[EM] New methods wrap-up
nkklrp at hotmail.com
Tue Jan 6 22:18:02 PST 2004
Yesterday I said that I hadn't checked out automatic candidate withdrawal
with wv Condorcet. Now I've checked it out and it doesn't improve on wv. It
just shifts the winner one position around the cycle, losing the special
advantages of wv. So: Re-abandoning automatic candidate withdrawal.
I gave up trying to find out of NES, (Nash Equilibrium Selection), first
suggested by Alex, it seems to me, meets the majority defensive strategy
criteria. Finding that out, establishing the properties of the new methods
such as NES, BBB, DSV, etc. would probably be quite a job (but if you find
out, let me know).
Most methods don't meet the majorilty defensive strategy criteria, so maybe
those methods don't.
When the premise conditions of SFC are met, its less-favored candidate can't
win at Nash equilibrium. If someone else can, that means that the less
favored candidate can't win in NES. But if there's no Nash equilibrium, the
method goes to its tiebreaker, and the less favored candidate could win the
tiebreaker. So showing that the less-favored candidate can't win means
showing that someone else can win at Nash equilibrium, or else showing that
the less-favored candidate can't win the tiebreaker.
Of course all the new methods have their own justification. For instance
there's certainly virtue in an outcome that no one can improve on. But I
can't very well propose something that I don't know the properties of. And I
like the majority defensive strategy criteria, which wv has been shown to
Check your PC for viruses with the FREE McAfee online computer scan.
More information about the Election-Methods