On Mar 4, 2010, at 1:04 AM, robert bristow-johnson wrote:
On Mar 3, 2010, at 9:13 PM, Dave Ketchum wrote:
Actually, a chance to move ahead, if we grab it.
IRV lost, and good riddance.
well, i believe the chance to move ahead in Burlington is lost for
at least a generation.
...
so, i'm for Condorcet too. i am sorta agnostic about what to do
about a cycle (because i really doubt it will happen at all often in
reality) as long as it's a sensible resolution (Shulze would be okay
if it was easy for a layman to understand, so probably Ranked Pairs
is the simplest, but i might just say give it to the Plurality
winner in the Smith set to toss the IRV haters a bone).
Let me try making it reasonably simple - perhaps better to not get too
fancy - make it understandable to the public but leave gory details
for formal definition.
Ranked Pairs is reasonable as best base. Every pair of candidates,
including those that involve write-ins, gets counted (and, unlike for
IRV, equal ranking is permitted).
If one candidate wins every one of its pairs, it is the winner - the
CW. These are common for, in most elections, the top one or two
candidates are far ahead of the rest and the one, or best of two, wins.
Else we have a cycle of three or more members, such as A>B>C>A, due to
differing opinions:
Each cycle member, if the only such running, would be CW over
those outside.
The pairs that define the cycle have varying margins between
their stronger and weaker member. By canceling the weakest of these
the stronger remain to define the winner.
Bringing Plurality in would be a distraction, since we have no need to
go near this method and risk a worse answer. Further, our calculating
has not necessarily identified which cycle member would win this
(though my method of doing the N*N matrix does provide this).
BTW - we should not discourage bullet voting - we should NOT encourage
voters to go beyond their desires, doing what is really nonsense
ranking.
----
Election-Methods mailing list - see http://electorama.com/em for list info
