[EM] Discover Magazine article

[MIKE OSSIPOFF]

you wrote (25 Oct 2000):
 > Ok, Markus, but what would you say to a suggestion that we
 > count the ballots by Tideman(wv) and by BeatpathWinner, and then
 > hold a 2nd balloting so people can vote which of those 2 winners
 > they prefer? That sounds democratic to me.
 >
 > ...
 >
 > If EM ever needs to vote on something, a good system might be
 > to count the ballots by BeatpathWinner and by Tideman(wv), and
 > choose the winner that pairwise-beats the other.

I would say that this method violates independence from clones
and monotonicity.


I reply:

Fine because we don't vote on EM anyway. Of course saying that
that would violate monotonicity and demonstrating it aren't quite
the same thing, are they. As for independence from Clones, that
doesn't seem very relevant, because, for reasons that I've already
described, the methods that _do_ pass ICC aren't accomplishing anything
meaningful, because of ICC's unrealisticness. In any case, I'm not
really suggesting that the above combination method be used for EM
votes, because there aren't any EM votes.

But about that 2nd balloting between the BeatpathWinner winner
and the Tideman(wv) winner, would that violate monotonicity too?
If so, could you demonstrate that it would?

In any case, that's irrelevant, because I didn't mention it as an
actual proposal. I mentioned it as a way of leading up to the
idea of using the method, among those 2, whose winner usually
beats the other method's winner.

You said:

You wrote (25 Oct 2000):
 > But say, for reasons of simplicity, we aren't permitted to count
 > the ballots by 2 methods. Then, the best we can do toward the
 > goals in the previous paragraphs would be to use the count rule
 > that usually picks someone who pairwise-beats the other count
 > rule's winner.

Due to Steve's opinion, the Tideman method is better than the
Schulze method because in those cases where the Tideman method
and the Schulze method choose different candidates the Tideman
winner usually pairwise beats the Schulze winner.

I reply:

Yes, that's what I said too--that it's therefore more democratic to
use Tideman(wv). It's what the people would choose if they were allowed
to vote between those 2 methods' winners.


You continued:

To my opinion,
the Schulze method is better than the Tideman method because
in those cases where the Schulze method and the Tideman method
choose different candidates the worst pairwise defeat of the
Schulze winner is usually weaker than the worst pairwise
defeat of the Tideman winner.

I reply:

Can you demonstrate the accuracy of that statement?

In any case, it sounds as if you'd  be happier with PC, which always
chooses the candidate whose greatest pairwise loss is the least.

You continue:

You wrote (25 Oct 2000):
 > As for Copeland, of course Copeland would likely do very well
 > in winner comparisons. But, as Steve pointed out, the fact
 > that we have more important criteria doesn't mean that, when
 > those criteria are met by 2 methods, doesn't mean that we're
 > forbidden to then consider winner comparisons. None of us
 > advocate Copeland, of course.

There is a difference between Steve and you. Steve considers
independence from clones to be very important. You consider
independence from clones to be "too unrealistic to be
important." Therefore election methods like highest Copeland
ranked BC winner are better than Tideman due to Steve's
argumentation when one doesn't consider independence from
clones to be useful.

I reply:

I have to admit that I'm not quite sure what method you mean by
"highest Copeland ranked BC winner" Do you mean that we would choose
among the winners of the BC complying methods by using Copeland?

Would that be monotonic?

Remember, though, that I'm talking about which single method, by
itself, would be more democratic. Any mention of combination methods
is only to show why it's more democratic, when considering 2
methods, to use the one of those 2 methods whose winner usually
pair-beats the other method's winner.

If we counted the ballots in a Presidential election by Tideman(wv)
and by BeatpathWinner, and then held a 2nd balloting, letting the
people choose which of those 2 winners they wanted, then the voters,
in that 2nd balloting would usually choose the Tideman(wv) winner.
That's why Tideman(wv) is more democratic than BeatpathWinner, when
those are the 2 methods being considered.

Mike Ossipoff









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

Share information about yourself, create your own public profile at 
http://profiles.msn.com.