[EM] Ranked-Pairs, conclusion

[MIKE OSSIPOFF]

Yesterday I said that it seemed to me that, in my example, the
difference between RP & SSD seems to come down to the question of
how imporant a beatpath is, in disagreeing with a defeat, meaning
that I felt that RP recognizes beatpaths as able to nullify defeats
that they oppose.

But after writing it occurred to me that CSSD is equivalent to
BeatpathWinner, and BeatpathWinner of course is about opposing beatpaths.

So they both recognize beatpaths and their ability to nullify
defeats or other beatpaths, and it's interesting to look at the
difference in how they judge.

At first, BeatpathWinner might appear to come out on top, because
RP's winner's defeat of BeatpathWinner's winner is nullified by
a stronger beatpath in the opposite direction.

But BeatpathWinner only looks at beatpaths between 2 candidates in
isolation. That beatpath that would nullify A's defeat of B contains
a defeat that is itself nullified by a still stronger beatpath.

If we allow that a beatpath can nullify another one, then CA10
isn't in a position to nullify anything, since it itself is nullified
by AD20 & DC30.

So then it seems that the difference between BeatpathWinner and
RP is that BeatpathWinner looks at beatpaths between 2 candidates in
isolation, while RP considers the whole picture, the consistent whole.

Either is right, depending on what looks better to you. Personally,
I agree with RP that if BC3 & CA10 would nullify AB2, then AD20 &
DC30 nullify CA10, and so it makes sense to not let CA10 nullify anything. 
So RP seems more right to me, where the 2 methods differ.

But they both meet the criteria important to me, and the considerations
discussed in this letter don't change my choice of what to advocate.

I'll continue to recommend CSSD to committees & organizations. My
choice of whether to propose RP, CSSD, SSD, or PC publicly will depend
entirely on what polling says about which is more acceptable.

...if I propose a rank-method. Of course Approval is the best public
proposal, and is what I've actually been proposing whenever the
subject comes up. Of course we must start bringing the subject up
more.

The person whom I asked said that she prefers SSD to RP, probably because 
the Schwartz set is very convincing.


Mike Ossipoff



_________________________________________________________________
Chat with friends online, try MSN Messenger: http://messenger.msn.com