[EM] {Process} Collective ordering of SW standards

[Mike Ossipoff]

Steve Eppley writes:
> 
> [I've inserted {Process} into the subject line.  I'll try to be
> disciplined enough to always do this for messages about the process
> to be used in election-methods, and I encourage others to do so.]


1st, let me skip ahead, to say that I agree with the voluntary
inclusion, in our report to ER, of the rankings collected whwen we
vote on which method to recommend, and people's explanation (voluntary)
for why they voted as they did. As you said, this will greatly help ER
members judge our recommendation.

It probably wouldn't be necessary for us to rate the standards--the
ER members can do that for themselves. But we should report our
individual & collective conclusions about which methods meet which
standards. And of course our final recommendation of a method, along
with the rankings for that vote, & voluntary explanations for why
we voted as we did. And any other comments &/or conclusions that
EM members care to add.


> 
> Mike O. wrote:
> 
> >I emphatically urge that we just cover single-winner methods in
> >this FAQ & recommendation project, which is a SWC project.
> 
> I'm willing to narrow the scope to SW.  If we succeed, we can try MW
> afterward.  I think we'd be doing a valuable service to the movement
> if we can show CVD and elections-reform, to name a few, how
> important SW reform is.

Quite so.

> 
> Does anyone who intends to participate in this document production
> process disagree about SW first?
> 
> >PR is too different from SW methods & issues. They can't be dealt
> >with in one project. PR is more straightforward & uncontroversial.
> 
> Funny, I have the opposite view-  SW reform seems less controversial
> since it's about improving a method, not restructuring the govern-

I agree. I just meant that SW reform is more controversial among
electoral reformers. I know how contentious some people are about
PR methods, but the it's SW that has genuine cause for controversy,
due the very radical merit differences exist between SW methods, and
which _don't_ exist between PR methods.

Yes, it's true that, while PR is a new kind of legislature, SW reform
is merely a better way of doing what we already do: electing 1 person
to 1 office. And I completely agree that SW reform will be easier to
get for that reason. I've never heard any opposition to it, except for
one clown on the street who kept objecting to everything on the grounds
that it would be too expensive. He even objected to Approval on those
grounds, though Approval would cost exactly zero dollars to implement.

In countries where there's much PR debate, PR opponents very frequently
advocate preferential SW elections instead of PR. Unfortunately the only
preferential SW method they've heard of is the abysmally inadequate MPV,
but it's our job to change that.

[Speaking of MPV, I promised Jim that I'd relay his arguments about it
to the committee, and I will do that. I haven't yet, because we've
been away from the MPV subject for a while]


> ment.  PR algorithms (thresholds, fractional transfers, hybrid
> systems, etc.) are more complex than MPV, Condorcet, Approval, etc.

Yes, fractional STV is incomparably more complicated than any
single-winner method. And the system being used right now in Cambridge
is more complicated than fractional STV. So when people say that
Condorcet's method isn't simple enough (but fractional STV is), they're
showing an outrageous double-standard. Besides, Condorcet's method is
defined as simply as MPV.

> And there are energetic advocates for each of a variety of PR systems 
> whose emergent properties are richly complex, affecting the nature of 
> the government.
> 
> >Yes, I like the idea of voting among the SW standards, for the
> >purpose of creating a collective ordering of them, and then
> >using the standards in that order, to evaluate & compare methods
> >by each standard.
> 
> That goes beyond what I suggested.  We don't need to vote on each
> standard's importance, or how well each method serves each standard,
> until the end.  If earlier votes are needed, they're only about
> what subtopic to write about next.  These agenda votes don't need to


I agree. It would be only to determine the order in which we discuss them,
evaluate the methods by them. ER doesn't need our ratings of the standards,
since that's something they can judge for themselves.

> be very accurate or well-reasoned since any subtopic which loses is
> really just being postponed awhile.
> 
> >The alternative would be to just have a free-form discussion where
> >each person could name standards & tell why a method does or doesn't
> >meet that standard. But this would mean everyone talking on different
> >subjects, not dealing with other standards. The procedure of dealing
> >with 1 standard at a time puts each standard before everyone, gives
> >everyone an opportunity to deny that their method fails a particular
> >standard. It would always be possible to say "Then why didn't you
> >speak up when we were discussing that standard & it was said
> >that your method failed it?"
> 
> Those aren't the only two alternatives.
> 
> This appears to call for overly restricted discussion; I don't think
> my brain is that disciplined.  My answer to that last question would
> be "I just now thought of such-and-such".  Someone else might answer
> "I just subscribed" and these should be legit answers--the document
> needs to be alive for amendment and expansion.

True, but the fact that each standard was held in front of all of our
faces, and we each had a chance to defend our method with respect to
each standard, will inspire more confidence in our report.

> 
> I'm also not comfortable with this proposal of advocacy.  I expect
> advocates of a method to fully criticize the method, not depend on
> others to do so.  Those who rank a method low should still point out
> its merits.

I will, reluctantly, list the (erroneous &/or fallacious) "con" arguments
against Condorcet's method, even though that will require bringing in
additional (academic) standards. I don't really want to do that, but
I will if others feel it should be done.

I usually reserve these possible "con" arguments, based on academic
standards (which are academic in both senses of the word, since they
are only about what the academics like, and ignore the concerns &
interest of actual voters) for when someone is considering actually
doing a Condorcet initiative, or other form of public proposal of
Condorcet's method. At that point, the person has a right to know 
everything, however fallacious, that can be thrown at Condorcet, and
what the answers are.

But, at the merit-discussing stage, as oppposed to the actual
initiative-organizing stage, I don't know that standards that none
of us have proposed are at all relevant to our discussion. Standards
that are important neither to the public nor to electoral reformers.

However I'll bring those academic standards into the discussion if
anyone wishes.

But, based on experience, I expect that proponents of methods aren't
going to offer the arguments against them, though it would be good
if they did, if those "con" arguments involve standards that are 
important to the public or to electoral reformers.
> 
> Though it makes sense to try to focus on one standard at a time, it
> won't derail us if we deviate some.  Messages can (should!) include
> links to the standards and methods that they address, such as:
>    {{Majority Rule {{Plurality {{oxymoron
> so we can easily find them later when discussion eventually focusses
> there by searching for "{{Majority Rule {{Plurality".
> 
> The "oxymoron" keyword (or keyphrase) named the argument being
> discussed.  We're likely to want to search old messages for
> arguments too.


I'll try to start including those keywords. I don't know if my
e-mail software is able to search whole messages, as opposed to
just subject lines. The keywords will certainly be helpful in
searches.

> 
> I'm using "{{" to denote "keyword:", but one of you can probably 
> suggest a smarter syntax.
> 
> We'll have to maintain a list of accepted keyphrases to name the
> Standards and the Methods, and maybe also the Arguments.  This
> should be easy to maintain.  We can rotate this secretarial duty
> among volunteers.  (Don't everyone stand up at once...  :-)
> 
> Rob has written about hypermail.  Maybe he can suggest how we can 
> structure our messages to take advantage of it.
> 
> >I prefer Condorcet's method for our voting method.
> [snip]
> 
> We have two distinct voting needs.  If our rankings of the Standards 
> and how well the Methods serve each is performed at the end, as I 
> think it must be, then we don't need to address this voting need 
> until late.
> 
> The other voting need, as I mentioned above, is to schedule the
> subtopics for discussion.  I don't see a need for a complex method
> for this, since discussion of the 2nd and 3rd place subtopics would
> follow soon after the winning subtopic anyway.  How about if we
> occasionally hold an approval vote on the remaining subtopics
> (approval means "ok to focus on this subtopic next"), and discuss
> the top few in order?  If there's a tie we can use luck, or the 
> chronological order the subtopic was appended to our outline.

As I was saying in another message today, I wouldn't object to Approval,
though it would probably be better to order all the standards in 1
Approval vote. I propose Condorcet's method. Collect rankings of the
standards & do a Condorcet count for 1st standard to discuss. Then,
among the remaining standards, do a Condorcet count for next standard
to discuss. etc. This uses only the 1 set of rankins that we collect
in that 1 balloting.

But I'll go with whatever method someone else suggests.

> 
> We should also allow messages about previously discussed subtopics to
> be injected at any time, to keep our rules simple and our process
> flexible.

Fair enough.

> 
> One other thing I'd like to add to the project: we should include
> our raw votes on the rankings of Standards and Methods, and each try
> to explain them.  This would be voluntary, of course.  Our reasoning
> could be helpful to the larger community--they can discount some of
> our votes if our reasons don't make sense to them (or aren't
> provided), and we can help to inform their own ranking process. 

Most definitely.

> 
> --Steve
> .-
> 


--