Reply to Bruce's article

[Mike Ossipoff]

First, I've already commented about the academic criteria, and
especially about how they relate to Condorcet's method. But let
me briefly repeat that Smith//Condorcet meets those criteria
that Bruce lists, while meeting the lesser-of-2-evils standard
in a way that Copeland & Regular-Champion do not. For that reason
I've said, several times, that Smith//Condorcet has Condorcet
& all of its versions, including Regular-Champion completely
dominated. 

***

The "Generalized Condorcet Criterion" that Bruce mentions
is better known to us, in this list, as the Smith Criterion.

***

Bruce's "Generalized Majority Criterion" isn't very general.
It only applies to a very limited & special class of majorities.
I call it the Mutual Majority Criterion"  (MM). 

I have posted a Generalized Majority Criterion to this list,
which is much more general in its application than is the
Mutual Majority Criterion.

***

Monotonicity: The only methods that we're considering that
violate monotonicity are MPV, Shugart's modified Runoff, and
probably York's WB (point-assignment) elimination method that
Steve mentioned.

By the way, of course, since Steve more or less "nominated"
York's method, we could consider it one of the candidates.

I agree that violation of monotonicity is a serious problem,
but the 3 methods in our election that violate it also
have big violations of the lesser-of-2-evils standard.

***

Mike


--