[EM] Stephane's poll--let's go with it.

[MIKE OSSIPOFF]

I'd like to participate in Stephane's poll, in the single-winner
category.

When voting beteen voting systems, when different people advocate
different ones, there's always the question of which voting system
to use.

When we did our poll last year, our solution was to use Voter's Choice.
A base method is needed for Voter's Choice. Plurality is the natural
one for public use. But many here like Approval, and Approval serves
the purpose well too.

Balloting was done by rankings, Approval ballots, and Cardinal Ratings
ballots. Anyone could designate a method (a balloting & a count rule)
for the purpose of having that method determine the placement of his
Approval votes in the final Approval count. When one chooses that option, 
one's Approval set is extended down to the winner by the
method that one has designated. In that way, one chooses the
method that he considers best at choosing his best compromise, to
determine his Approval cutoff point.

But the use of Voter's Choice was optional--By designating "manual",
the voter could have his own Approval ballot place his final Approval
votes.

But since this is Stephane's poll, and he has called for this poll,
he of course has the right to say how it will be conducted. His
way of choosing methods, by repeated polling as more people participate,
will be interesting.

Stephane gives a good reason for using FPTP as the initial method:
It's fair because everyone dislikes it, and so it doesn't favor any
of us. However, it's in the spirit of Stephane's procedure to start
with Approval, because Approval is the method that most recently
won one of our polls. It seems unnecessary to start from scratch,
with FPTP, when we already know that EM has chosen Approval.

So I suggest that change in the rules: Approval should be the
initial method.

Do we need nominations? We all know what the method proposals are,
but maybe our voting strategy would be affected by knowing just ones
are in the poll. So maybe we should do nominations.

I nominate Approval, Bucklin, Cloneproof SSD, BeatpathWinner(wv), &
Ranked-Pairs(wv).

I don't know how recently Bucklin has been defined here.

Bucklin:

1. Each ranking gives a vote to its top choice.
2. If any candidate(s) have a vote total exceeding half the number
   of voters, or if every ranking has given to all of its candidates,
   then the count ends and the winner is the candidate with the most
   votes.
3. Otherwise, each ranking gives a vote to its next candidate. Go to 2.

[end of definition]

Bucklin is an excellent method when people want rank balloting, but
want to do a handcounted election with lots of candidates.

My Approval votes:

If there are fewer than 5 candidates, or fewer than 5 voters, or
if someone offers to count the rankings by computer, then I vote
for all of the methods that I've nominated except Bucklin.

Otherwise, I vote for Approval and Bucklin.

Strategy in Bucklin is to rank down to the expected CW, the best
compromise that you can get, but no farther.

Mike Ossipoff



_________________________________________________________________
Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp.