[EM] Re: Black box election method tests--Is it possible?

[MIKE OSSIPOFF]

>Which of the
>criteria, if any, require knowledge of the internals of the counting
>method used?

Criteria should mention only a voting systems results, and not its rules.

Nearly all of the criteria that we use are purely results-criteria, not 
rules criteria.

But sometimes we hear of a rules-criterion. For instance, sometimes someone 
criticizes Approval because it doesn´t meet their requirement for "one 
person one vote" (opov). opov is a rules criterion that says that a voter 
should only be able to vote for one candidate. Presumably all rank methods 
likewise violate opov. Or maybe opov advocates would only disallow methods 
that let people vote several candidates equal. That determination would need 
a precise statement of opov.

I asked the opov advocates to justify opov in terms of other criteria, or in 
terms of a widely-accepted fundamental standard. They never did.

Some other criteria are defined as rules criteria. For instance, someone, at 
their website, defines Condorcet´s Criterion in such a way that it applies 
only to rank-balloting methods. That makes his Condorcet´s Criterion a rules 
criterion. My definition of Condorcet´s Criterion doesn´t mention methods´ 
rules.

I claim that there´s no place for rules criteria. Criteria should mention 
only a method´s results.

Mike Ossipoff

_________________________________________________________________
Don’t just search. Find. Check out the new MSN Search! 
http://search.msn.click-url.com/go/onm00200636ave/direct/01/