[EM] Chicken Dilemma--To whom is it a problem?

[Kristofer Munsterhjelm]

On 10/20/2013 10:34 PM, Dick Burkhart wrote:
> The problem with Condorcet, SDSC, and the like is that they ignore crucial
> information - namely, the strengths of the voters preferences of one
> candidate over another. That is how mathematician Donald Saari can argue for
> the superiority of Borda, and why others argue for cardinal voting.

On the other hand, preference strength is not well defined - at least 
not unless one provides some kind of utility standard. A rating could be 
an opinion (going from bad to good) or an expected value (taking 
dynamics of the "game" of the election plus polls into account). The 
more it is akin to the latter, the more it moves the calculations from 
the method itself to the minds of the voters.

(I wrote about this once before. See 
http://www.mail-archive.com/election-methods@lists.electorama.com/msg11343.html 
, where I probably explained it better than I am now.)

So if ranking has a problem in that it reveals to little, then rating 
has a problem that exactly what a rating means is ambiguous. It can be 
solved objectively, by stating a standard scale or saying the people 
define such a standard themselves. It can also be solved subjectively, 
by saying a vote means what the voter wants it to mean and thus its 
instrumental purpose is what is important. But neither case is without 
problems of its own: the former in that one has to show why exactly this 
objective standard is right, and the latter in that it can hollow out 
the method, requiring the voters to do calculation that the method 
should properly do.