[EM] New Condorcet/RP variant

[Paul Crowley]

On Fri, 5 Nov 2004 10:39:57 +0000, Paul Crowley <ciphergoth at gmail.com> wrote:
> On Thu, 04 Nov 2004 21:33:19 +0100, Markus Schulze
> <markus.schulze at alumni.tu-berlin.de> wrote:
> > Dear Paul,
> >
> > your Condorcet/RP variant sounds like Steve Eppley's
> > "minimize thwarted majorities" (MTM) method.

> For each of the n! possible
> orderings of the candidates, MTM generates a "sorted thwarted list" by
> sorting the lower triangle least first, then picks the ordering with
> the least sorted thwarting list; my method generates a "sorted
> affirmed list" by sorting the upper triangle greatest first, then
> picks the ordering with the greatest sorted affirmed list.

I am mistaken; MTM sorts the thwarted list greatest first just like my
method, and it would be crazy to do otherwise.  With this correction,
MTM is exactly equivalent to my method.

It may not be a problem for MTM that it's not cloneproof in the case
of ties; in fact, it would seem to me positively desirable to
sacrifice cloneproofness in face of deterministic resolution in this
instance.  However, it leaves a problem of defining a cloneproofness
property that MTM does satisfy - perhaps that the result set from
which the result is chosen fairly with clones is a subset of the one
generated when some proper subset of the clones are deleted?

Does Eppley still read this list?  I'd be interested to know why he
now favours MAM over MTM.
-- 
  __
\/ o\ Paul Crowley
/\__/ www.ciphergoth.org