Well I nominate Kristofer to administer the poll.  Michael, if you are plonking people or vise versa, it might not be as good for you to administer the poll.There seems to be a whole shitload of alternatives.  But a good method can handle that.Powered by Cricket Wireless------ Original message------From: Kristofer MunsterhjelmDate: Sat, Apr 6, 2024 12:40To: robert bristow-johnson;Toby Pereira;Filip Ejlak;Michael Ossipoff;Cc: EM;Subject:Re: [EM] Poll on voting-systems, to inform voters in upcoming enactment-electionsOn 2024-04-06 16:33, robert bristow-johnson wrote:

For the ranked ballots, I'd just use one of the defeat-droppers. I
imagine that they would all lead to the same result up to ties. I like
Ext-Minmax[1] myself because it has a very low tie rate, but again, I
think anything would do.

If there's a cycle, I'd report the Smith set.

Rob LeGrand's ranked-ballot voting calculator, listed at
https://electowiki.org/wiki/Online_poll, can both do Minmax and identify
the Smith set. It can't do Ext-Minmax, but it should be relatively easy
to verify the order by eyeballing the Condorcet matrix if that becomes
an issue.

As I understood it, Mike also wanted to do an Approval poll by asking us
to submit Approval ballots alongside the ranked ones. For these, the
method would be Approval itself. I probably wouldn't ask for an Approval
poll if I were running it, but it doesn't bother me if one's included
either.

-km

[1] Ext-Minmax is just: the winner is the candidate with the least worst
pairwise defeat. Break remaining ties by second worst defeats, third
worst defeats, etc.

On 2024-04-06 18:55, robert bristow-johnson wrote:

Since Michael Garman also said MO shouldn't be running this poll, if
there are no other suggestions, I'd like to administer it.

I've unblocked the two people I got plonked. But someone should probably
tell MO about the change because I think he's got me plonked, since the
lists I've seen since I unblocked him don't contain my suggestions.

I would also suggest merging the Condorcet-IRV hybrids into a Smith-IRV
category. While Condorcet methods can handle lots of alternatives, it's
easier to fill out a poll with fewer alternatives. But if you'd prefer
having the full unabridged list, that's okay by me too. Alternatively we
could remove just Schwartz-Woodall.

If anybody else has merge ideas, just lemme know. The most obvious
candidates would be combining Copeland//Borda and Black.

And of course, if MO also wants to count the votes too (based on what
posts he get), I'm not going to stop him - the mailing list posts are
public, after all.

The current methods are:

Approval
RP(wv)
Schulze
IRV
Plurality
Majority Judgement (as a category; includes usual judgement etc.)
Approval with manual runoff
Copeland//Borda (also called Ranked Robin)
MinMax(wv)
Black
STAR
Smith//Score
Baldwin
Benham
Woodall
Schwartz-Woodall
Smith//Approval (implicit - of all ranked)
Smith//Approval (explicit - specified approval cutoff)
Margins-Sorted Approval
Smith//DAC
Margins-Sorted Minimum Losing Votes (equal-rated whole)

The nomination deadline is 2024-04-11 05:14:59 UTC, inclusive. (Not like
we're going to need to be accurate down to the second, but there we go.)

-km

Some comments about this voting methods poll seem to be in order.

A referendum on voting methods is paradoxical, because you have to know
good method, in order to find it. It's standard scientific method,
explained by HG Wells over a century ago, that voting method is not a
matter of opinion but a matter of demonstration.

This opinion poll would access the views of list members but it also
shows the disconnect with voters of the world, judging by the list of
options, mentioned by Kristofer. The caution by, or attributed to,
Werner Heisenberg is worth mentioning again: if you cannot at last come
back and explain wat it is you have found, all your finding has been
worthless.

Social choice theory has had 70 years to do this. They held a 50 year
celebration, at turn of 2000. Where is their standard model of election
method? Or is it an "Impossibility" in their fallible opinion? Their
insisting on the Anglo-American single member system is a case of: what
is may not be what's right (as David Hume said, and half the world seems
not to think so). Single member elections automatically halve
representation, its most inefficient form: "only half a democracy"
(Roert Newland). Compared to that, all the tweaks to single member
systems are insignificant, and not calculated to promote democratic
reform or progress. Only a Dr Pangloss would suggest this is the best of
all possible democratic worlds.

Regards,

Ricchard Lung.

On 07/04/2024 01:02, Kristofer Munsterhjelm wrote:

On 2024-04-07 09:39, Richard Lung wrote:

There's something about the way you write that's hard for me to
understand. For instance, I don't know what you mean by "demonstration"
here.

But on the general point: I doubt the results will be different between
Condorcet methods, except perhaps up to ties - some methods will return
more ties than others. That's why I suggested a method that has a low
tie rate.

We do know what we would find, though: the accumulated opinions of the
people who participate, about what methods are more suitable than others
for political use.

There are so many methods because yes, there is an impossibility. We
can't have every desirable criterion at once, so we have to choose. Namely:

Arrow's says that there exists no ranked single-winner method (apart
from certain random exceptions not listed here) that can't be made to
change its mind by removing candidates who didn't win.

Gibbard says that you can't be strategy-proof: whatever your method, it
will sometimes pay to vote tactically.

Duggan-Schwartz says that the impossibility of being strategy-proof
remains even for multi-winner, unless you elect every candidate who is
someone's first preference.

I would hardly dismiss them.

I wouldn't say single-member district is the best of all possible
worlds. I might even say that better democracy reduces the use of
elections, by using selection by lot combined with deliberation instead.
But for the situations where you have to use single-winner, it's a good
idea to know what to use. Or to have a selection of good methods to
choose from.

-km

Krisofer, thank you for reply, the reference is to HG Wells, who, used
the term "demonstration" in The Elements of Reconstruction, 1916, with
accompanying elucidation. This can be downloaded from the collection of
his books at Project Gutenberg or from my edition of collected writings
by Wells: The Angels Weep. HG Wells on Electoral Reform:

https://www.smashwords.com/books/view/804800

 There does not have to be different methods for single member and
multi-member election systems.. In fact, consistency demands that they
use the same method. A general voting method can be done.

And majority counts were dismissed by John Stuart Mill as maiorocracy,
the tyranny of the majority - taken-up by Lani Guinier. (Successful
Tammany Hall-type Machine assaults on PR did not always achieve even
majorities.) The majoritarian approach dismissed itself, with the
limitation to its idea of democracy.

In my experience, voting does not always have to be "strategic" or
deceitful. That is a sort of belief in an electoral Realpolitik, as the
inevitability of systemic electoral fraud. Rather (as rationally
invented) I have found that the consistent use of keep values always
responds rationally, in a re-count of shuffled votes.

Regards,

Richard Lung.

On 07/04/2024 11:12, Kristofer Munsterhjelm wrote:

On Sat, Apr 6, 2024 at 17:02 Kristofer Munsterhjelm km_elmet@t-online.de
wrote:

Not sure what KM means by “administer”. Make rules & decisions by decree?
Or maybe just be chairperson or facilitator.

This is a poll, not an organization. It neither has nor needs a charter or
officers, including chairperson.

I proposed a poll because there wasn’t one. I suggested a voting method &
nomination & voting periods because those are needed for a poll. But those
were suggestions, & were obviously open to discussion.

I didn’t perceive myself as “administering” the poll, because it didn’t
occur to me that any participant should be special, or have unique right to
make suggestions or ask questions.

KM evidently believes otherwise.

“unless someone has another suggestion”?

Yes, I do: There’s no need or reason for one participant to have unique
authority or privilege to make decisions or rules, or even have a
chairperson’s procedural authority. …or the unique privilege of making
suggestions/motions or asking questions, or any amount or degree of such
privilege.

I don’t care who posts nomination-updates.

With the understanding that it’s unclear what “run” & “administer” mean, &
that I don’t believe in any degree of unequal authority or privilege in
poll-decisions (which should always be made collectively, with identical
equal voice for everyone):

KM said that he’d like to “administer” the poll because two people said
that I shouldn’t.

What they said was that I shouldn’t administer if I had people blocked.
Not the same thing. KM changed the meaning of what they said, by
conveniently leaving out a clause in both utterances.

Yes I’d blocked people, as had KM. Because of the poll, I unblocked, as did
KM. I’d been posting nomination-lists, & then KM posted one.

BTW, I was reading the poll-related posts of blocked members, in the
spam-folder. My latest list, which I hadn’t posted yet, had the same 23
nominations as KM’s posted list.

Incidentally, the two people KM quoted saying that I shouldn’t “run” or
“administer” the poll were two people I’d recently blocked for continuing
repetition of some quite nonsensical claims including the most contorted &
silly defenses of FairVote’s lying, & an attempt to use the argument,
familiar from IRVists, that one-person-one-vote means Plurality-Voting or
IRV (With Robert’s convenient unexplained exception for Condorcet).

Pretty much no one here agrees with those claims. I blocked those two
because, given what they’d just said, I was
no longer interested in what they say, & had a right to not display their
posts or emails.

Given those circumstances & background I’d hardly recognize those two as
impartial evaluators of me.

There remains the matter of the poll’s voting-method. There’s obviously a
very strong collective preference for rank-methods, & a near-unanimous
preference for Condorcet-complying methods. …suggesting a choice among only
such methods.

Of course those methods only rarely choose differently.

In fact, is it even necessary to specify a rank-count for the poll, which
anyone can count & report as they choose, according to what they consider
best tells how people feel?

But I initially wanted to report EM’s method-finishing-order, & felt that
it would have stronger meaning if it as counted as EM preferred.

I suggested Schulze as the primary method because of its great long
popularity, & because of the abundance of count-software, including at
CIVS, which I suggested as the most natural count process.

There were a suggestions for other methods, & so, given that collective
choice should take precedence over one person’s initial suggestion, I
brought up the matter of voting on the method.

Only 3 possibilities, unless someone can suggest another:

1. The initial Schulze suggestion
2. Collectively-choose a voting-method
3. Voter’s choice

I invited people to express any wish for a vote on the method, bro only if
there’s desire to do so instead of just using the popular Schulze.

Evidently there’s no sentiment that we should find a substitute for
Schulze, as the nominal EM choice (to whatever extent one is even
needed)That’s good, because it simplifies things.

BTW, of course, because not all nominees are Condorcet, & because even
Condorcet methods can choose differently, then, with so very many
preferred methods, many methods might have to be counted in Voter’s choice.
Maybe Voter’s Choice isn’t extremely practical for such a big field.

Then, if it were necessary to choose a new method, it would be necessary to
first choose how to vote in that choice.

If that were necessary, I’d suggest Voter’s Choice, among a few nominees.

But I agree with you that a new choice isn’t necessary.

I nominate dictatorship by KM.

On Tue, Apr 9, 2024 at 4:57 PM Michael Ossipoff email9648742@gmail.com
wrote:

Of course, in the poll’s final choice, it’s very unlikely that several
Condorcet versions will choose differently. But, for thoroughness, since we
have a few days before voting,why not choose a version, just in case
there’s a circular tie.

Maybe it’s premature to assume that you prefer to leave it at Schulze.

So:

Part 1) (not primary): Why not, for starters,  post a few suggestions.
Let’s state a few of our favorites—ones that we like & would agree with.

For example, these are my most favorite, that I’d agree with—I’d agree with
any one of them!

Condorcet-IRV
RP(wv)
Schulze
MinMax(wv)

So, I suggest & ask (as a member, not an “administrator”, whatever that
is), what are a few of your most favorite, that you’d agree with?

Part 2) (primary): Since you most prefer rank methods, then let’s rank a
few of our most favorite methods.

My preference-order ranking is the one in which my favorites are listed
above.

Of course, as before, there’s probably a CW, & so it doesn’t matter which
Condorcet counts those rankings.

But what if they disagree? Then, for that eventuality, I suggest Voter’s
Choice

It’s simply this:

You automatically approve what your designated method chooses, & everything
you like better. Simple & well-justified, & doesn’t need any agreement on a
method.

My designated method is RP(wv), because it’s usually socially preferred to
the Schulze winner, it meets LIIAC, & it’s simple, natural, intuitive

On Tue, Apr 9, 2024 at 16:54 Michael Ossipoff email9648742@gmail.com
wrote:

Thanks for your honesty. :-)

On Tue, Apr 9, 2024 at 18:13 Closed Limelike Curves <
closed.limelike.curves@gmail.com> wrote: