[EM] math 103 website - Arrow & Saari

[Bart Ingles]

Forest Simmons wrote:
> 
> On Mon, 7 Jan 2002, Steve Barney wrote:
> 
> > Bart:
> >
> > What is the definition of a "deterministic" voting system, as Saari apparently
> > uses the term?
> 
> I think in this context it refers to how well you can predict the ballots
> from the utilities.  As Bart says, given these extreme utilities it is
> easier to predict the Approval ballots than to predict the Borda ballots.

Actually I don't recall Saari making reference to utilities.  That was
one thing that struck me as odd in his approach -- that he seemed to
assume all preference levels were equidistant.  Although he does seem to
attach significants to number of "hops" between preference levels.

Bart