[EM] Condorcet definitions

[MIKE OSSIPOFF]

Alex--

You wrote:

We just wrapped up a long thread over "What is one person/one vote?" and
in the end decided to apply the term to the most common interpretation,
while inventing a new term for the (arguably ill-defined) "voting power"
interpretation.

I reply--

Anyone wanting to apply 1p1v to methods other than Plurality would
have to define and justify it anew.

You continued:

So, to avoid another "dueling dictionaries" thread, how about the following:

Condorcet's Method:  Any method reasonably close to Condorcet's original
proposal(s) for resolving cycles.

I reply:

I don't want to be difficult, but Condorcet & Copeland have always
been treated as different circular tie solutions, on EM and even
elsewhere.  So it seems more in keeping with EM usage if "Condorcet's
method", or "Condorcet" refers only to Condorcet's circular tie solutions 
based on the magnitude of defeats--methods that drop weakest
defeats or keep strongest defeats. Markus says that Condorcet suggested
Copeland, but we shouldn't call Copeland Condorcet's method, if we're
going to stick with current & past usage.

You continued:

A Condorcet method:  Any method that satisfies the Condorcet criterion.

or, maybe you prefer:

Pairwise method:  Any method that satisfies the Condorcet Criterion.

I reply:

Why not just call a methods a "Condorcet Criterion method" if it
meets Condorcet's Criterion.

A "Condorcet method" seems to imply a version of Condorcet's method.

Pairwise-count methods, or pairwise methods, are the family of methods
that look at pairwise defeats and elect a candidate if he beats everyone
pairwise. So they meet Condorcet's Criterion. But so does Nanson's
method, which isn't defined in terms of a pairwise count. So Nanson
is a Condorcet Criterion method without being a pairwise-count
method.

Nanson has more than one method named after him, but I'm referring to
the one that repeatedly drops the candidate with lowest Borda score.
Nanson was used in Wisconsin for a while. So far as I know, the only
Condorcet Criterion method ever used in public political elections.
I don't like Nanson because, like most methods, inlcuding lots of
Condorcet Criterion methods, Nanson doesn't meet any of our defensive
strategy criteria.


I'm just saying how these terms have been used so far.

Mike Ossipoff


_________________________________________________________________
MSN 8 helps eliminate e-mail viruses. Get 2 months FREE* 
http://join.msn.com/?page=features/virus

----
For more information about this list (subscribe, unsubscribe, FAQ, etc), 
please see http://www.eskimo.com/~robla/em