Re: [EM] IRV inconsistency

[Richard Moore]

Markus Schulze wrote:

[EMAIL PROTECTED]">

It doesn't make much sense to criticize IRV for violating the
consistency criterion and then to promote Condorcet methods and
to say that a violation of the consistency criterion is not a
problem when the used election method meets the Condorcet criterion.

I think Forrest is right; consistency violations in IRV are going to be
more severe than in Condorcet. This is because the consistency violations
in IRV are the result of elimination of candidates who have a majority
preference over the actual winner. In Condorcet, they simply result
from non-transitive majority wins, which are fact of life in ranked
elections. IRV is actually creating inconsistencies where none need
be; Condorcet is merely getting trapped by pre-existing ambiguities.

[EMAIL PROTECTED]">

And it doesn't make much sense to criticize IRV for violating
reversal symmetry and then to promote a Condorcet method that
violates reversal symmetry and to say that a violation of reversal
symmetry is not a problem when the used election method meets the
Condorcet criterion. As Blake Cretney wrote to Mike Ossipoff
(16 Apr 2001):

You must admit that if you rail against IRV's non-monotonicity,
and then propose a method that is non-monotonic, you're going to
run into some rhetorical problems.

I should point out that rhetoric and logic are quite different things. Of
course, rhetoric seems to stick in the minds of most people more than
logic does. But that doesn't mean that a persuasive rhetorical argument
is always valid.

Richard

[EMAIL PROTECTED]">