[EM] Cyclic method comparisons

MIKE OSSIPOFF nkklrp at hotmail.com
Sat May 13 15:50:56 PDT 2000

Markus said [referring to the comparison of methods based on
one method's winner usually beating another method's winner]:

>This argument is problematic because it can be cyclic.
>
>Suppose that somebody else would propose a new election method
>and would demonstrate that the winner of his election method
>beats pairwise the MTM winner more often than vice versa and
>that the Schulze winner beats the winner of his election method
>more often than vice versa. What would you conclude?

But that hasn't happened. Saying that something (maybe) could
be cyclic isn't the same as finding a cycle. If the DCD winner
and the MTM winner usually  beat the winner by your method,
and you can't show us a method that your method beats in that regard
and which beats both DCD & MTM in that regard, then the mere
talk of the possibility of cycles isn't enough to take the meaning
from the fact that those 2 methods' winners usually beat your
method's winner.

And even if you had another method, to make the relation cyclic,
the significance of the cycle would still be questionable if
that other method isn't something that is being proposed at least
as much as the other methods in the cycle.

Mike Ossipoff

________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com