[EM] Summed Ranks (SR)

[Kevin Venzke]

Mike,

--- MIKE OSSIPOFF <nkklrp at hotmail.com> a écrit :
> Summed Ranks (SR):
> 
> The winner is the candidate for whom the number of candidates ranked over 
> him/her, summed over all the ballots, is the least.

This seems to have a clone-loser problem. If SR doesn't satisfy SDSC then
this would be very bad.

> SR meets FBC. SR is a simple rank method that meets FBC. I haven't studied 
> it yet, and I don't know what other properties it has. It may have 
> disadvantages that I don't know of. I like it because it's a simiple and 
> completely ordinal rank method that meets FBC.

I guess you suppose that SR might be better than MMPO.

As far as rank methods, you could also use ER-Bucklin(whole) or one version
of Borda. Suppose the ballot is A=B>C=D=E>F>G.

ER-Bucklin(whole) counts this ballot towards A and B in round 1, CDE in round
3, F in round 6, G in round 7.

The FBC-friendly Borda variant counts 6 points to both A and B, 4 points to 
each of CDE, 1 point to F, and zero to G.

ER-Bucklin(whole) also satisfies SDSC, but not SFC.

Kevin Venzke



	

	
		
___________________________________________________________________________ 
Appel audio GRATUIT partout dans le monde avec le nouveau Yahoo! Messenger 
Téléchargez cette version sur http://fr.messenger.yahoo.com