MTAOC conditionality doesn't need the change that I suggested in a post a few
minutes ago.
I'd believed that there'd be a problem if (in the ABE) the number of A voters +
B voters
was odd. But that isn't so. If middle(x,y) is even one greater than
middle(y,x), then saying
par(x,y) = "no" is consistent with the method's intent.
So I'm leaving MTAOC conditionality as it was, and retracting the "fix" that I
posted
a few minutes ago.
By the way, someone objected to conditionality by mutuality, but any FBC/ABE
method
that elects C in the Approval bad-example is effectively doing the same as
that. That
includes the FBC/ABE proposal posted by the poster who objected to
conditionality by mutuality.
Mike Ossipoff
----
Election-Methods mailing list - see http://electorama.com/em for list info
