SW Committee?

[Mike Ossipoff]

I don't object to answering criticisms of my proposal, but I
find that I'm answering them again & again. The debate in this
list is going over & over the same ground. That's a good sign
that everyone has said what they have to say. And that means
that it's time to move on to a vote.

The original goals of the Single-Winner Committee were to
recommend to ER about what single-winner method(s) would
be the best to propose to the public. In fact, members of ER
have said that they'd value a recommendation by us. That
was said by ER members who did _not_ want discussion by us.
Recommendation yes, discussion no.

So I suggest that a recommendation is the important thing for
us to send to ER. 

Sure, a FAQ would be good too, if we can get one prepared.
We've all agreed that we'd like to send ER a FAQ as well
as a recommendation, along with optional individual 
conclusions & reasons for voting as we did.

Rob L. suggested that my www article on standards & criteria
for single-winner methods would be a good initial working
document for such a FAQ. I agreed, and made 2 minor corrections
in it.

No one, so far as I'm aware, has suggested any modifications
to that initial FAQ. Either we could send it to ER as-is, or
we could abandon the FAQ idea. I prefer the former, but of
course I'm not an unbiased judge of my www article.

But those seem to 2 options for the FAQ, unless the 3rd option
is to just wait longer, while nothing is happening.

So I propose that we take 2 votes:

1. Vote on whether to send the initial FAQ to ER  (y/n)

2. Vote on which sw method is the best to propose to
the public. We could also report the results gotten by
applying the count method 2 or 3 times, to pick our 2nd
& 3rd collective winners.


I suggested voting on sw methods a month or 2 ago, and the proposal
was seconded. Since nothing is happening now, wouldn't you agree that
we should move on to voting?

Anyway, I move that we conduct the 2 votes named above.

***

As I was saying before, there's a problem when voting among
several alternatives, when no sw method is official for doing
such a vote. 

But, there _is_ a method that has some official status: Condorcet's
method. We  voted, earlier, on what sw method to use in an election
about splitting ER, a while back. We voted to use Condorcet's
method (True, only 2 people voted, but should we use something
that _no one_ voted for? No choice can be made without some
method).

Other methods have been proposed here since then, such as Copeland,
and Demorep's votes-for count, (which sounds to me like MPV without
the elimination, but with the transfers & 1st choice count).

But though more methods have been proposed here, it makes the most
sense to use the most recently official method, the method we
chose for ER splitting election.

And it simplifies things to have a method that's official.

***

My motion, therefore, is that the election between sw methods
be counted by Smith//Condorcet. I propose Smith//Condorcet
instead of plain Condorcet because some people have said
they like it better, and because its longer definition won't
be a problem to the people on this list.

***

Maybe, instead of voting on whether to do those 2 votes (on
sending the initial FAQ to ER, and on voting among the sw
methods), people can "vote" on whether to vote, merely by
voting on those 2 elections or not voting on them. Not voting
indicates opposition to holding the vote. Voting indicates
approval for holding the vote.

True, you don't have to vote in these elections, but I suggest
that if you don't, then you're thereby voting for the SW Committee
to not return any results to ER. That's what the result will be,
isn't it, if we don't consider what recommendation to send them?

***

I believe that anyone can propose a vote, so my doing so doesn't
mean that I claim to be "chair". We were always a chair-less committee,
and that seems a good structure.

***

So if you have any opinions on the relative merits of the methods
that have been proposed to this list (I'll list them later in
this letter), or the relative merits of a few of them
(remember, you can rank as few as you want to), then please
post a ranking to this list, a ranking of (at least 1 or more
of) the sw methods that have been proposed here.

***

Though we have no structure for deciding these things, wouldn't
you agree that if 1 or more people strenuously object to Condorcet's
method being used, then we should do the count in a way that favors
no method?

In that case, I'd suggest what I've called the Mixed Method:

1. We collect voters' rankings, on this list.
2. We count the rankings by all of the methods that have been
   proposed to this list.

3. We hold a 2nd balloting between all of the winners byk th

methods. Would Approval be ok for the 2nd balloting? It's an
obvious extension of Plurality. It wouldn't favor any of
the rank-balloting  count proponents.




 


--