[EM] FBC + majority defensive strategy criteria--2nd generation best methods

Kevin Venzke stepjak at yahoo.fr
Mon May 23 14:08:25 PDT 2005

Hi Mike,

--- MIKE OSSIPOFF <nkklrp at hotmail.com> a écrit:
> WV and its enhancements could be called 1st generation best methods. 
> Approval combines FBC and WDSC, but when FBC is added to wv's better 
> criteria, that counts as a new kind of merit, a 2nd generation of best 
> methods.

I'm happy my attempts stand up to a first glance! I hope I haven't made
any bad mistakes.

> As you pointed out earlier, MMPO meets SFC. And, with AERLO, it meets SDSC 
> too. In fact, with AERLO, methods that meet FBC meet Strong FBC--at least in 
> regards to 1st choice ranking. As you said, that's the important aspect of 
> FBC compliance.
> So: FBC, SFC, and SDSC, all with one method.

Could be. (Which I say, because I can't quite wrap my head around the 
implications of AERLO.) On the other hand, I don't want to forget, MMPO
is rather indecisive.

> with IRV. And you or someone showed that wv fails FBC. That means that it 
> isn't possible to reassure people that there's never a need to bury their 
> favorite.

Yes, that was me.

> You didn't name the 2nd method you defined yesterday. But it's more 
> complicated than MMPO. AERLO could be added to MMPO in a subsequent public 
> proposal, so that the 1st proposal could be plain MMPO.
> If I refer to it here, I'll just call it the 2nd method. It's true that a 
> majority who prefer X to Y can keep Y out of set S by just ranking X over Y, 
> and Y over no one. But what if X is in a viscious majorilty cycle, added-to 
> by voters other than the X>Y voters? Might S not be empty?  So that method's 
> SDSC compliance could come at the cost of indecisiveness. I don't know. That 
> could be incorrect. It's just my first impression.

In that method (as opposed to the three-slot one), S can't be empty. The
worst case scenario is that there is a pairwise tie among all candidates, so
that S contains everyone. (The condition for inclusion of candidate A is
maxfor[a]>=maxagainst[a].  ...And I just now notice that my original post
contained a typo on that line, reading "maxagainst[b]." Hopefully that
didn't confuse anyone.)

On these ballots:
40 A>B>C
35 B>C>A
25 C>A>B

The contests are A:B 65:35, A:C 40:60, B:C 75:25.
Maxfor:Maxagainst for each candidate is A 65:60, B 75:65, C 60:75.

So S contains {a,b}. Then B has the greatest number of non-last rankings
and is elected.

> Anyway, I like the simplicity of MMPO.

Me too, although I miss compliance with Clone-Winner and SDSC.

Kevin Venzke


Découvrez le nouveau Yahoo! Mail : 1 Go d'espace de stockage pour vos mails, photos et vidéos ! 
Créez votre Yahoo! Mail sur http://fr.mail.yahoo.com

More information about the Election-Methods mailing list