[EM] Margins

[Kevin Venzke]

Dave,

--- Dave Ketchum <davek at clarityconnect.com> a écrit : 
> >>>> If the voter leaves two candidates unranked, 
> >>>> don't count that as a half vote for each, count it 
> >>>> as zero for each.  But if the voter explicitly ranks 
> >>>> two candidates as equal, count that as a half vote 
> >>>> for each.
> > Since you argue for it, I want to name it after you.  Unless someone complains,
> > I am going to start calling it the "Ketchum" measurement of defeat strength.
> > 
> > Since Ketchum is a compromise between WV and Margins, it doesn't make sense to say 
> > you advocate it "for both wv and margins."  It only makes sense to advocate it 
> > *instead* of WV or Margins.
> 
> NO SALE!!!  Our debate has been as to how to count votes.  HOWEVER, WV and 
> Margins differ as to how to interpret the counts after they have been 
> done.

The interpretation of the counts is exactly the same, when Margins is defined
in terms of giving 1/2 votes.  Giving half-votes and then using the WV
interpretation of the counts == Margins.

>  I have NOT SUGGESTED ANY change in the interpretation, so they 
> remain different and could not survive a shared name.

Let's say we apply your preferred rule to WV, and let's call the resulting 
method "WV-Ketchum."

Also, let's say we apply your preferred rule to Margins, and call the
resulting method "Margins-Ketchum."

Then WV-Ketchum and Margins-Ketchum are exactly the same thing.

That is why I say that it makes no sense to argue for your count rule "for both 
WV and Margins," because your count rule is a complete replacement for WV and
Margins.

> > I think you doubt this, so I'll give some brief definitions to make this
> > clearer:
> > 
> > WV: Equal-ranking and truncation result in no vote for either candidate.
> 
> For WV it would be destructive to include truncated candidates as if the 
> voters had expressed an interest in them, so leave them out.
> 
> Where the voter has explicitly done equal ranking, I continue my claim 
> that these should be given half counts, consistent with the voters likely 
> doing about half A>B and half B>A if A=B was not permitted - and ending up 
> with comparable counts.

Ok, but when this is done, the method is more like Margins than WV.

> > Margins: Equal-ranking and truncation result in 1/2 vote for each candidate.
> 
> Assuming I understand the rules, margins CARES NOT about such counts, so 
> doing them unlike WV puzzles.
>
> I would do the counts here the same as in WV, so that the two methods 
> differ only where there is a reason for difference.

I see the problem: You don't realize that using Margins gives you exactly
the same results as when you use WV but give half-votes in cases of equal 
ranking.

I will show this with some examples if you want.

> > measure     ranked equal      truncated
> > WV           0 votes           0 votes
> > Margins      1/2 vote          1/2 vote
> > Ketchum      1/2 vote          0 votes

Kevin Venzke
stepjak at yahoo.fr


	

	
		
Créez gratuitement votre Yahoo! Mail avec 100 Mo de stockage !
Créez votre Yahoo! Mail sur http://fr.benefits.yahoo.com/

Dialoguez en direct avec vos amis grâce à Yahoo! Messenger !Téléchargez Yahoo! Messenger sur http://fr.messenger.yahoo.com