[EM] Re: Participation criterion: a thought

[Eric Gorr]

At 7:23 PM -0400 9/22/03, Dgamble997 at aol.com wrote:
>Eric Gorr wrote:
>
>"What are these essential standards you speak of?
>
>In my mind there is only a single essential standard:
>
>
>Can the method reliably find the group preference?
>
>
>For any method that is not clone-proof, the answer is no.
>
>For any method that is not monotonic...I would guess the answer is no.
>
>For any method that does not pass Participation...the answer is can be yes.
>
>As such, while Participation may be nice to have, it does not appear
>to be necessary."
>
>
>Eric do you really mean that? Group preference is a term open to 
>very wide interpretation.

True, but that does not imply my interpretation is very wide open.

A method that, at a minimum can select a winner not in the Smith set, 
could not possibly reliably find the group preference.

>If your sole criteria is "does the method find the group 
>preference?" Why do you prefer Condorcet to Borda counts?"

   51 A>B>C
   35 B>C>A
   14 C>B>A

By the only definition I would consider reasonable, an alternate 
which received > 50% of first place, sincere, votes is the group 
preference.