Paul K wrote:
> Steve Eppley wrote:
-snip-
>> Aren't all the voting methods we've been promoting
>> both anonymous and neutral? Doesn't that mean
>> none of them are entirely non-random?
>
> No.
>
> If "none of them are non-random" then all of them are random.
> That's definitely not true. And the there is no connection
> between "anonymous"
and "neutral" (whatever that means).
Anonymity means each voter is treated alike. Neutrality
means each candidate is treated alike. (You can find
clearer definitions in the em archive and in the
social choice literature.)
These are two criteria that the public is unlikely
to be willing to discard.
> There's only a few of the ones discussed on the list that
> have a random
component. The suggestion that "none are
> non-random" is the
same as "all are random". It would
> suffice to demonstrate the "randomness" of any Condorcet
> method, or of
IRV, to support the claim that "all methods
> we discuss are random."
Okay, I'll repeat the example I wrote that Paul neglected:
50%: A > B
50%: B > A
Pick any of the single-winner methods we've been
considering. How does it decide which _one_ candidate,
A or B, to elect? IRV and Borda and the Condorcetian
methods must appeal to chance (as a tiebreaker)
if they are to be anonymous and neutral.
-- Steve
---- Election-methods mailing list - see http://electorama.com/em for list info