[EM] Expert wrap-up. Criteria inclusion. Objectivity.

[MIKE OSSIPOFF]

When I asked Dave who had said that their critreria had been selected by 
experts, Dave quoted Russ's statement that I'm an expert on voting systems, 
and said that that means that I have said that our criteria have been 
selected by experts. I suppose, indirectly, he could say that since I helped 
"select" criteria, and since I was said to be an expert, that means that we 
were indirectly implying that our criteria were selected by experts, if you 
excuse the incorrect noun-plural.

But Dave lists that as a promotional tactic. If a website owner says 
anything complimentary about a contributor, then, to Dave, that website is 
doing so as a promotional tactic to gain authority for the website's 
statements. I suggest that Dave is being ridiculously oversuspicious about 
shrewd promotional tactics.

What Dave was saying was that we were justifying out criteria on the 
authority of an expert named by the website. But we give plenty of 
justification for our criteria. The introductory page discusses he 
lesser-of-2-evils problem and the need to use methods that avoid it. After 
each criterion's definition, we have an explanation for why the criterion 
means something to the person interested in majority rule and getting rid of 
the lesser-of-2-evils problem.

So, for Dave to suggest that  we rely on expert authority to justify our 
criteria suggests that Dave is either dishonest or astoundingly stupid.

If Dave wants to find fault with out criteria, he needs to say something 
about the criteria themselves. What in particular does he not like about 
them? In what way does he feel that a different criterion says something 
more important about voting systems by some widely-valued standard?

But Dave is the familiar type who can only criticize the motives & honesty 
of those with whom he disagrees.  This isn't unfamiliar. People unqualified 
on an issue tend to fall back on statements about the character, motives, or 
honesty of others.

As I said, I've never said that I was an expert--only that I'm an expert as 
far as Dave is concerned.

Russ said that I'm an expert on voting systems because he agrees that my 
approach to method merit is useful and consistent with voters' strategic 
concerns and those of voting system reform advocates. And that my 
definitions of criteria and methods are better than the sloppiness that is 
so often encountered.

I'm sorry that it disturbed Dave so much that I was called an expert, at the 
website.

It's worth repeating that website owners are not obligated to publish things 
that they don't considser important. In particular, any idiot can come up 
with a criterion and say "This criterion is important". Sorry, Dave, but 
we're not obligated to list them all.

Some criteria don't deserve being listed. Others are meaningful and their 
compliance desirable, but that doesn't mean that we can list them all 
without confusingly cluttering our criteria list and compliance table. For 
clarity & simplicity we chose to limit the number of criteria that we list.

Dave, again, seems to think that it's a lie when we say that we leave a 
criterion out because it isn't as important. He believes that our real 
reason for not listing  criteria is because they don't make our methods look 
good.

But I've always said that FBC is important and desirable. We list it even 
though it seems to me that someone showed us a Condorcet FBC failure 
example. No method meets every desirable criterion, but Condorcet wv meets 
enough important criteria to make Condorcet wv a good method.

Our website advocates Approval, and Approval passes Participation, and yet 
we don't list Participation. If we add it, Dave will say that we're trying 
to favor Approval. If we don't add it, Dave will say that we're trying to 
avoid disfavoriing Condorcet. Participation compliance is desirable, and 
Participation is a simple criterion. I'm for adding it. I'll tell Russ that 
someone has requrested it.

I'm glad that Dave agrees that Participation is important. It helps show how 
Approval is better than IRV. As I said, Condorcet wv has other criteria 
compliances that outweigh its Participation failure. IRV has nothing any 
good that outweighs IRV's Participation failure.

I emphasize that Participation failure is more of an embarrassment than a 
strategy problem, and that's one reason why I  haven't so far really pushed 
for listing it at the website.

WV & Approval offer different  criteria compliances:

Approval: FBC, WDSC, Participation

SSD, MAM, SD, BeatpathWinner/CSSD: SFC, GSFC, WDSC, SDSC, Smith Criterion

PC: SFC, WDSC, Condorcet Criterion.

All of these meet Summability.

IRV meets none of the abovementioned criteria.

Does IRV meet anything? It meets the Clone Criterion and Mutual Majority. 
Those are its advantages over Approval. But the situations in which that 
advantage shows are also situations in which IRV demonstrates its 
inclination to give incentive for favorite-burial.

If you want ICC & MMC, the better wv methods meet those criteria, in 
addition to CC and the majority defensive strategy criteria, not met by IRV.

Again, we don't list ICC & MMC because we don't have room for everything, 
and so we list the criteria that we consider most important.

Dave complained that we say that methods should be objectively judged by 
criteria, saying that we were un-objective in leaving out some criteria. As 
I said, the decision on what criteria to include is unavoidably a subjective 
decision, weighing completeness against clutter.

One should determine what guarantees one wants from a voting system, and 
then write or adopt criteria that best measure for that desired standard. 
Evaluating methods by such critreria is the objectivity that our website was 
referring to.

But objecitivity doesn't require us to list  criteria other than the ones 
that we consisder most important.

Mike Ossipoff

_________________________________________________________________
Check out the coupons and bargains on MSN Offers! 
http://shopping.msn.com/softcontent/softcontent.aspx?scmId=1418