[EM] "Unmanipulable Majority" strategy criterion 2

[Kevin Venzke]

Hello,

--- En date de : Ven 28.11.08, Raph Frank <raphfrk at gmail.com> a écrit :
> By requiring that at least 3 levels are possible, you are
> effectively
> forcing lots of methods to fail.  

I'm sure that's the intention, though I'm not sure why. The criterion
wants to limit "cheating" in a sense, and methods like FPP are hard to
cheat under.

I have no problem with the approach... I use a somewhat convoluted 
framework to explain why strictly-ranked methods (with no equal ranking
permitted) don't satisfy my take on votes-only FBC, but if I wanted to
be understood quickly I could also just say that a method must allow
equal ranking to satisfy the criterion.

> Also, just because most
> methods
> would meet the criterion in the 2 candidate case isn't
> a reason to
> exclude that case.

I'm puzzled on this point also because satisfying the criterion in one
case isn't enough to satisfy it overall, so why state it. I think the
intention must be to just say that three levels are not required, in
cases where there are not three candidates.

Kevin Venzke