Brief criteria list. Not Conjecture.

[Mike Ossipoff]

Bruce says that he's heard nothing but conjecture about ways in
which Condorcet is better than Copeland. He said that he hasn't heard
any precisely defined criterion relating to strategy (presumably
he includes, in that, freedom frem need for defensive strategy).

On the contrary, six precisely-defined criteria have now been
posted to this list, relating to such things as the lesser-of-2-evils
problem & the need for defensive stratgegy. Let me list them here:

LO2E1 (Lesser-of-2-Evils Criterion #1)
LO2E2 (Lesser-of-2-Evils Criterion #2)
IME (Invulnerability to Mis-Estimate)
GMC (Generalized Majority Criterion)
ITA (Invulnerability to Twin Alternatives)

Two of these, LO2E1 & GMC, have been defined in 3 versions,
because different versions could seem more important to
different people.

***

It's been demonstrated on this list that Condorcet's method
meets all of these criteria. Examples posted to this list
show that Regular-Champion fail every one of these criteria,
as would any Copeland version.

***

I repeat that here because Bruce has denied that there have
been precise definitions of such criteria, or demonstrations
that Condorcet meets them.

As regards LO2E2, I showed that plain Condorcet & Smith//Condorcet
strictly meet it if the subcycle rule is used. And I specified
the conditions that would be needed for Condorcet to fail
LO2E2. Especially for Smith//Condorcet, these conditions for
failure of LO2E2 are so demanding & improbable (cheaters engineering
a circular tie, and also engineering a subcycle on that cycle,
among the non-S2 alternatives) that it can be said that 
Condorcet meets LO2E2, even without the subcycle rule, for
all practical purposes.

As I said, plain Condorcet & Smith//Condorcet have been
demonstrated to strictly meet all of the rest of the
6 criteria listed in this letter.

***

It isn't conjecture, Bruce.

***

Mike




--