Anton Sherwood <bronto@pobox.com> wrote:
> How about fractional approval? While not addressing the strategic
> question mentioned by Brams & Fishburn
> (http://bcn.boulder.co.us/government/approvalvote/altvote.html), it does
> allow a distinction, lost in other methods: does a ballot saying "Smith
> > Jones > Brown" mean approval ratings of .9 .8 .1, or .9 .2 .0? Does a ballot saying "Smith or Jones but not Brown" mean .9 .8 .1 or .9 .5 .1?
No time to check myself, but isn't "Single Transferable Vote" or something
like this? I'm pretty sure I've read (in my urls) of a system where you have
a bunch of votes and distribute them among the candidates. Same idea,
except using small numbers rather than reals, so a bit less granularity.
Saari argued that symmetry pushes for equal weighting of the ranks, but maybe
that was at a global level: the system shouldn't assume people have 4:3:0 or
4:1:0 style preferences. But I don't know if that has any relevance here.
> I like it too, though (like all the other methods mentioned here) it
> fails to capture *degrees* of preference
Well, ranking lets you express more degrees of preference than simple "I
approve/don't approve". Also, I wonder if Borda counting lets you skip ranks,
or if that degrades the system's performance.
I've always liked the idea of "None of the Above". So you could rank the
liberals, then NOTA, then the conservatives, if you wished.
-xx- Damien X-)
Received on Mon May 8 10:52:21 2000
This archive was generated by hypermail 2.1.8 : Tue Mar 07 2006 - 14:45:30 PST