On Tue, Nov 28, 2000 at 03:12:51PM -0500, Raul Miller wrote:
> On Wed, Nov 29, 2000 at 01:44:33AM +1000, Anthony Towns wrote:
> > The reason I'm not accepting your interpretation, or considering it
> > at all reasonable, is that I'm still not seeing any basis for your
> > interpretation than that it comes up with the right answer.
> I'd say "plausible answer" instead of right answer -- your interpretation
> does not come up with a plausible answer.
That depends what you consider "plausible". I'm willing to believe the
constitution has bugs, and that in some circumstances it may very well
come up with nonsensical results for a vote. So I'm not willing to rule
such an answer implausible.
> Single Transferable Vote biases the selection in favor of first
> preferences at the expense of other preferences. Can you think of a
> better kind of criteria for making the selection?
Other methods can be found at the URLs I cited at the start of the
thread. "Reversing the fewest and weakest pairwise defeats" (ie,
using the smallest possible casting vote) is probably another reasonable
alternative. But, as I said, I don't profess to know enough about cycle
breakers to really say.
> Or, can you construct
> an example where you feel that Single Transferable Vote seems unfair?
Well, trivially, if you have:
60 ABC
50 BCA
40 CAB
you end up with a cycle, and STV goes with:
A: 60 B: 50 C: 40
A: 100 B: 50
so A wins, in spite of a 90:50 majority who explicitly prefer C to A,
and, if they'd wanted and known that a cycle was likely to happen,
could have made C win by insincerely changing their votes to:
50 CBA, 40 CAB
(which coincidently would've removed the cycle)
This sort of situation happens no matter how you resolve a cyclic tie,
though. You pretty much have to be "unfair" in some sense to choose a
winner. As I said, I'm inclined to suspect that there other means are
likely to be more optimal, although I'm not clear exactly how.
Cheers,
aj
--
Anthony Towns <aj@humbug.org.au> <http://azure.humbug.org.au/~aj/>
I don't speak for anyone save myself. GPG signed mail preferred.
``Thanks to all avid pokers out there''
-- linux.conf.au, 17-20 January 2001
Attachment:
pgpXu91mdlCfs.pgp
Description: PGP signature