Hi--
The Smith Criterion doesn't imply GMC. There can be
a 1-way _majority_ beatpath in the Smith set.
I've long felt that a really good method has to meet
Smith & GMC. Or at least Condorcet & GMC.
That result where ranking someone lower in IRO helps him
is a violation of what's called the "Monotonicity Criterion".
That's probably the thing that the academic authors dislike
most about IRO. IRO is quite unpopular with people who write
on single-winner methods. CVD seems quite oblivious to that.
For me, IRO's violation of Condorcet & GMC is worse.
Smith compliance implies Condorcet compliance.
Condorcet compliance implies Majority Criterion compliance.
Mike
