Election Results

[clay shentrup]

11/20/2009 2:00:43 Clay Shentrup
11/20/2009 2:33:40 Raph Frank
11/20/2009 5:13:19 Abd ul-Rahman Lomax
11/20/2009 5:23:00 Sean Walker
11/20/2009 6:15:51 Clay Shentrup
11/20/2009 7:10:56 Clay Shentrup
11/20/2009 8:06:18 Abd ul-Rahman Lomax
11/20/2009 8:12:41 Abd ul-Rahman Lomax
11/20/2009 8:29:49 Sean Walker
11/20/2009 11:16:00 Warren Smith
11/20/2009 19:25:29 Abd ul-Rahman Lomax
11/20/2009 22:38:51 Warren Smith
11/21/2009 6:27:56 Abd ul-Rahman Lomax
11/21/2009 9:34:57 Warren Smith
11/22/2009 9:33:04 Abd ul-Rahman Lomax
11/23/2009 6:29:15 Abd ul-Rahman Lomax
11/23/2009 17:03:22 Clay Shentrup

so it's
Clay = 4
Raph = 1
Abd = 7
Sean = 2
Warren = 3

[clay shentrup]

Remember from the "parliament" web page:

- The richest 3 candidates at the stroke of midnight on Monday morning
(Pacific Time) will be the winners.

This thread can be used for any public negotiation of asset, including
any final orders to redistribute.

[Raph Frank]

Well, guess my campaign didn't go so well :).

With 17 votes, the quotas are:

Droop: 5 (4.25)
Hare: 6 (5.67)

Anyone with more than 4.25 votes is guaranteed a seat.

If we allow fractional transfers, then Clay just needs 0.25 (+delta)
of a vote to be guaranteed a seat. In practice, 4 votes is likely
enough.

[Sean Walker]

Interesting to see that I managed to garner a vote from someone other
than myself. Although, even Frank + myself couldn't change the result
of the election without one of the 3 top people giving us votes.And
even then, only Abd has any to spare.

[clay shentrup]

Why don't you guys let me take you out for martinis, and we can talk
about how best to redistribute your asset? :)

[Just testing something]

On Nov 27, 8:43 pm, clay shentrup <thebrokenlad...@gmail.com> wrote:
> Why don't you guys let me take you out for martinis, and we can talk
> about how best to redistribute your asset? :)

Hmm, this thread is pretty quiet. Maybe I should transfer to Sean so
as to cause a crisis, as we didn't define a tie breaking rule :).

Also, I note that Warren and Adb haven't said anything yet.

[Raph Frank]

On Sat, Nov 28, 2009 at 9:11 PM, Just testing something
<esf_m...@electionsciencefoundation.com> wrote:
> Hmm, this thread is pretty quiet.  Maybe I should transfer to Sean so
> as to cause a crisis, as we didn't define a tie breaking rule :).
>
> Also, I note that Warren and Adb haven't said anything yet.

Opps, that was me. I had the wrong account logged in.

[Sean Walker]

That would work, until Adb decided to transfer one of his spares to
someone of his choice, likely being Warren.Then he, Clay, and Warren
win.

On Nov 28, 2:11 pm, Just testing something

[clay shentrup]

My favorite tie-breaking rule is using randomness to choose between
the tie-ers.

[Warren D. Smith (CRV cofounder, http://RangeVoting.org)]

OK, look, you've made the deadline be midnight monday morning, and I
can't
do any asset-negotiating worth a damn because
I'll be on a plane and other travel crap.

Clay = 4
Raph = 1
Abd = 7
Sean = 2
Warren = 3

anybody with >4.25 assets wins.

Assuming Abd wishes to win, he has 2.75-epsilon assets
to give away, while everybody else needs assets.
Abd could singlehandedly make Clay+me win, or Clay+Sean,
but to make it be Sean+me he'd need the cooperation of Raph, whose
actual name
is Ivan by the way. (Also with cooperation he could make Ivan win.)

Sean is a high school student.
Ivan is an Irish engineer, don't know much about him,
he seems to be pretty smart and knowledgeable and
he wrote the districting software.
Clay is a web designer
in his early 20s, formerly aspiring to be a rock musician.
He founded the ESF, more than anybody else anyhow.
I'm a mid-40s PhD mathematician, currently unemployed.

Well, why me? Well, on the downside, I'm kind of getting burned out,
and
I've kind of had too much influence already; fresh blood is desirable.
On the upside, I've thought about this stuff more than anybody except
perhaps Abd & Ryan, and I know a lot, and I know something about the
grants world, not that that's helped so far (but it might if+when we
get incorporated as a nonprofit). Far as I understand, the purpose of
the elected board, is
to make decisions. I think I have a lot of what it takes, to make
good ones, though
I'm certainly not infallible.

I think the rule on the parliament page
http://groups.google.com/group/electionsciencefoundation/web/parliament
for elections every 2 months is silly, I advocate a longer
term like, oh 6 months, say. I also think the asset-weighting rule is
too dangerous as it stands,
a dictator could happen, and then keep it that way. Also this might
be illegal; I mean,
if state law calls for a board with >=X members, and some nonprofit
says "we've got that board, it's
just that all the other members have low voting weight, so it's
effectively a 1-member board"
then that could be illegal under state law. Also more complicated to
weight.

Be nice if Ivan, Sean, Clay etc were to post reasons why we should
give them assets, stances on
whatever issues they can think of etc. (or did you already and I
missed it?)
And it better be ASAP, if you want my assets, since I'm largely out of
e-contact.

As it stands right now I'm elected but not clinching, i.e., I could
become unelected.
In the event I cannot achieve election, I donate my all assets to
Raph=Ivan.
This offer may be rescinded by me, and I'm only saying it because
right now I have no information,
and I'm saying what happens if I do not get any more info in time.

-WDS

[Raph Frank]

On Sun, Nov 29, 2009 at 6:38 PM, Warren D. Smith (CRV cofounder,
http://RangeVoting.org) <warre...@gmail.com> wrote:
> As it stands right now I'm elected but not clinching, i.e., I could
> become unelected.
> In the event I cannot achieve election, I donate my all assets to
> Raph=Ivan.
> This offer may be rescinded by me, and I'm only saying it because
> right now I have no information,
> and I'm saying what happens if I do not get any more info in time.

Thanks for the vote of confidence. I think you probably have enough
to take the last seat.

If Abd doesn't transfer, then the Droop quota drops to 3.334. If I
also don't transfer, then it drops to 3, which means you exactly meet
it (so worst case, you tie for the last seat).


So,

<<<<<<<<<<<< NOTICE OF TRANSFER >>>>>>>>>>>>>>>>

I transfer 0.0000001 votes to Warren :).

(Though it comes back if he ends up assigning his votes to me)

<<<<<<<<<<<<<<<< END >>>>>>>>>>>>>>>>>>>

This means that unless Abd does some transferring, the winners will be
Abd, Clay and Warren. I think this is a reasonable result.

If Abd transfers more than 1 vote to Sean, it will activate Warren's
offer and I will take the 3rd seat (ahem, such monotonicity). If he
transfer more than 2 votes to Sean, then Sean will rise above Clay
(and still eliminate Warren).

Effectively, Abd can pick between

(do nothing)
"Abd, Clay, Warren"

(give me 2.1 votes or Sean 1.1 votes)
"Abd, Clay, me".

(Given Sean more than 2 votes)
"Abd, Sean and 50/50 me/Clay"

Also, I think given Abd received lots of votes, I think people support
the DP system, so the exact winners are perhaps, less important.

[Sean Walker]

Well, I don't have very many reasons for why I should necessarily be
elected, given that I am indeed just a high school student. Yeah, not
very good at selling myself, am I? Stance wise, I support using RRO as
a starting point for creating our own government setup, and any
general policy support for SV, splitline, etc. (I think splitline has
the advantage that it can't be gamed at all, since there aren't any
real modifications that could be proposed that would change it in
favor of one group)

Also, if needed, I have enough points to force the election of Clay
and Warren if I want to end things immediately. (Assuming fractional
votes are allowed, and I don't see why not)

On Nov 29, 11:38 am, "Warren D. Smith (CRV cofounder, http://RangeVoting.org)"

> I think the rule on the parliament pagehttp://groups.google.com/group/electionsciencefoundation/web/parliament

[Raph Frank]

On Sun, Nov 29, 2009 at 6:38 PM, Warren D. Smith (CRV cofounder,
http://RangeVoting.org) <warre...@gmail.com> wrote:
> I also think the asset-weighting rule is
> too dangerous as it stands,
> a dictator could happen, and then keep it that way.

One option would be to limit any one candidate from having more than
than 50% of the votes.

Ofc, with 3 people, as long as 1 isn't a dictator, any 2 would have a
majority of the weight.

The only effect would be if the weights were used to work out ratings.
The candidate with lower weight would have less of an effect on the
ratings sum/averages.

[clay shentrup]

On Nov 29, 10:38 am, "Warren D. Smith (CRV cofounder, http://RangeVoting.org)"

<warren....@gmail.com> wrote:
> OK, look, you've made the deadline be midnight monday morning, and I
> can't
> do any asset-negotiating worth a damn because
> I'll be on a plane and other travel crap.

We could extend the deadline.

> Assuming Abd wishes to win, he has 2.75-epsilon assets
> to give away

But then he would lose power.

> Clay is a web designer
> in his early 20s, formerly aspiring to be a rock musician.

1) I'm not a web "designer". That usually implies a graphic design
bent. I do computer programming. I've been at LiveNation.com but I'm
starting a new gig in a week, using the Ruby on Rails framework.

2) I'm not in my early 20's. I'll be 31 on Wednesday.

3) I *AM* a rock musician (with aspirations of doing it for a
living). I'm in the middle of an album right now. I just recorded
the drum and bass tracks to the first 5 songs in Hyde Street Studios
last month. I'll continue with the album when my producer gets back
from touring with the band Train (don't laugh - it's a paying gig;
he's played on the Tonight Show with them).

> He founded the ESF, more than anybody else anyhow.

Well, it's really just an outgrowth of your organization. Sort of a
rebranding and generalizing and attempt to get more organized.

> I think the rule on the parliament pagehttp://groups.google.com/group/electionsciencefoundation/web/parliament

> for elections every 2 months is silly, I advocate a longer
> term like, oh 6 months, say.

As I explained, that term was only meant to last until we are
incorporated -- which I thought could happen in 2-4 months. So it was
just a way to have sort of an "insurance policy" in case our initial
structure turned out to be less ideal and/or more game-able than
assumed; in which case people might want to usher in different
leadership who would fix it.

I'm sure once our initial "parliament" or "board" or whatever is
elected, it can be persuaded to change that term. It doesn't sound at
all an unreasonable proposition.

> I also think the asset-weighting rule is
> too dangerous as it stands,
> a dictator could happen, and then keep it that way.

That also sounds reasonable. That would nullify my above comment
about Abd's extra asset then.

> Also this might
> be illegal; I mean,
> if state law calls for a board with >=X members, and some nonprofit
> says "we've got that board, it's
> just that all the other members have low voting weight, so it's
> effectively a 1-member board"
> then that could be illegal under state law.  Also more complicated to
> weight.

Sure. But I already think we're going to have to invoke some clever
loopholes for getting around laws that might prevent some of our more
advanced processes. Not so much by intention, but just because the
authors were not social choice experts. For instance, say the law
says that decisions have to be ratified by a unanimous vote or a
majority vote of the board. Maybe we want some system (because the
evidence says it's more utilitarian) that just uses the best scored
variant. We could simply have the board agree to uphold whatever the
internal process selects, according to internal process. Of course
that's tricky. There are clever loopholes, and then there are
outright violations of the intent of the law. I/we really need to
read through some guidebook on corporate elections and pertinent law.

> Be nice if Ivan, Sean, Clay etc were to post reasons why we should
> give them assets, stances on
> whatever issues they can think of etc.

I had considered that, but I figured that my "legacy" of posts already
describes myself pretty well, and I'd somewhat prefer that people
judge me on the basis of who I've been rather than the nice-sounding
sales pitch that I might make. If anyone has specific questions to
ask us about our "positions", that might be a better system.

[clay shentrup]

To respond to this, AND talk about some of my stances...

I'm now more convinced that the candidates should have equal weight.
It makes things simpler, and it creates more of an incentive for
negotiation of asset. That is, Abd now has a reason to use his extra
asset to affect the other winners, and that probably creates more of
an incentive for candidates to negotiate and reveal more about
themselves.

As for Robert's Rules of Order, I see it as a convenient tool for
making some "quick and dirty" initial decisions. But I'm a strong
supporter of using something superior, that is actually based on
Bayesian Regret analysis. Since we're supposed to be experts in this
field, it seems a bit disappointing to use a system that is so...
primitive. Of course, using the Score Variants scheme does require
some software (manually is possible, but a little too demanding of
people's time), which I'll hopefully have time to write in the not-too-
distant future.

Probably my biggest problem with RRO is the same as my problem with
Abd's ratification notion. The basic problem is that whatever number
you choose is arbitrary. If you say "2 of 3 representatives must
approve this decision for it to be ratified", it's not as if you have
some Bayesian Regret metric that says 2 of 3 is better than 3 of 3, or
vice versa. That sort of process is fine for "quick and dirty" -- but
as an ultimate formal well-designed process, it's just reckless to
pick such arbitrary values, especially when you have no real evidence
that ANY of those values is ideal (because the underlying process of
using a binary vote as opposed to accepting the results of the Score
Voting election plausibly is less optimal).

It seems to me almost like Abd has this sort of metaphysical view of
human choice. That somehow through enough deliberation, there will
ultimately be some "consensus" that is achieved (which is the "right
choice"). But deliberation only has so much potential to inform and
allow reps time to re-evaluate the logic behind their views (aka
"utility assessments"). Eventually, at the end of all that
deliberation, people still have some set of utility evaluations for
all the options. Score Voting is an excellent method for revealing
those preferences. Whereas asking people whether they want to accept
the SV winner or hold a re-election, and then acting on the basis of
quantity rather than intensity, may actually diminish the revealed
preference ability of Score Voting. Abd's dismissive attitude about
"forcing utilities out of people" (which IS possible using revealed
preference techniques), and his deference to the "humanistic" view as
opposed to the materialistic/utilitarian view, is concerning to me.
It suggests to me that Abd would work to enact policies which serve
his intuitions more than any scientific indicators of effectiveness.
Call me left-brained, but I do not hold intuition in very high
regard. I want numbers and facts.

That being said, I now revisit the topic of redistributing asset. I
believe that Sean and Warren and Ivan all have demonstrated the kind
of analytical/engineer characteristics that give me confidence in
their overall judgment. Warren certainly has a logical-mathematical
ability that seems unrivaled in the group. Indeed, his Bayesian
Regret calculations and various theorems have been the formative
impulse for this group. We would be nowhere in this mission without
him. But I often feel he lacks the sort of practical knowledge that
is generally encompassed by the term "common sense", and I believe
that may be part of the reason that he hasn't historically made better
marketing and (ironically) organizational decisions. That is an arena
where I think Ivan and Sean have demonstrated more natural ability. I
say that with the utmost respect and even admiration for Warren.
Between Ivan and Sean, it is clear that Ivan has more developed
mathematical and programming skills, which have been quite valuable to
the group (e.g. his map-drawing software). Sean is clearly
exceptionally bright and has already demonstrated tact, diplomacy, and
insight that are beyond his years. If I had to choose someone to give
my assets to, it would have to be Ivan, based on the _combination_ of
his experience, judgment, and logical acumen.

But why bring that up? Well, like I said, I am an aspiring musician,
and I have for too long put off my musical ambitions. I have used too
much time doing voting reform discussion and advocacy (which has not
been particularly effective except for just getting the SV meme out
there on the Internet). Some days I have spent literally hours at
work putting off important tasks because I just HAD to write a bunch
of responses to some blog where people like Rob Richie were trashing
Score Voting. I've somehow justified that seemingly futile sort of
act, because I felt the stakes were so incredibly high that anything I
could do was worthwhile. Getting Score Voting used in government just
means so much that I've felt compelled to help in any way possible,
even for the 1 in a million chance that I could make a difference.
It's kind of like the irrationality of voting -- what are the odds
that my effort will make a difference?

But I'm eating into my 30's now, and if I really care about various
other ambitions, I need to focus on them, and just help with voting
reform as time allows. I can always come back to it in a few months,
or years. Or when I retire I can put 100% of my free time into it (if
Earth is still inhabitable by then of course). I recently saw a
taping of Demetri Martin's show (on Comedy Central), and even got a
chance to ask him who is comic idols were (Steven Wright, Steve
Martin, and Woody Allen, in case you're interested.) This guy dropped
out of law school (at Harvard I believe) with a year left, to pursue
his dream of being a comedian. It was so inspiring to sit there in
the studio and see him and his co-writers having a great time creating
a form of art. Music is my artistic passion, and it moves me in a way
that no scientific or political discipline can. You only live once,
so I've got to try to make these dreams of mine happen.

So at this time I think the best thing I can do is give my asset to
the best person, and then donate some time and/or money to the
incorporation effort. As a friend of mine (who has similar varied and
politically oriented interests) just told me, you can drive yourself
crazy trying to solve global warming, or you can enjoy planting one
tree or reducing your energy usage, and know you've made at least some
small difference.

Hence this decision...

<<<<<<<<<<<< NOTICE OF TRANSFER >>>>>>>>>>>>>>>>

I transfer all 4 of my votes to Ivan/Raph

<<<<<<<<<<<<<<<< END >>>>>>>>>>>>>>>>>>>

In so doing, I ask that he keep my Score Variants proposal as a
serious consideration, perhaps suggesting improvements to it or even
writing some prototype code to implement it. And I hope that he will
keep Abd's interminable philosophical rants in check with the rigor of
hard science and mathematics. :)

I also pledge to help financially to the greatest extend I reasonably
can when it comes to incorporating. Literally, I will contribute
hundreds of dollars to this effort if others can get the paperwork in
order. I can also handle some basic tasks like filing paperwork if we
do decide to incorporate in California. I am happy to be reached by
email or phone to handle these kinds of things.

This group has my passionate conviction. I just need to devote more
resources to my other hobbies in life. Raph surely has other
interests as well, but it seems he's more fulfilled with his
professional life, and more able to give a consistent amount of
attention to this group.

I hope that makes sense and doesn't seem rash to you folks.

Best,
Clay

On Nov 29, 12:30 pm, Raph Frank <raph...@gmail.com> wrote:
> On Sun, Nov 29, 2009 at 6:38 PM, Warren D. Smith (CRV cofounder,
>

[Abd ul-Rahman Lomax]

Any vote not used to elect a seat, with Asset voting, is "spare." The
use of Asset voting with the Droop quota is appropriate for a (short)
fixed deadline, my position has been that no deadline is appropriate.
Not enough votes for a seat, no seat, until, at least, a majority of
electors have agreed upon some future course. Setting a very short
deadline so that people travelling or otherwise occupied with the
Thanksgiving weekend was arbitrary, and I'm not guaranteed to recognize it.

It takes time to negotiate vote transfers. There is still "ballot"
thinking here, with only a very small tweak.

There were 17 votes. For three seats, Hare quota is 5.666 votes. The
Assets are:

Abd 7
Clay 4
Warren 3
Sean 2
Raf 1

Droop quota is 4.25. Electing someone with 4 votes is like a
plurality election. Consider it this way.

Suppose, at this point, that the winner is the "top three." That
would produce Abd, Clay, Warren. However, note the votes that are
excess and short:

Abd 2.75 excess over seat Droop quota (4.25)
Clay (-0.25)
Warren (-1.25)
Sean 2
Raf 1

If we use the Hare quota (which expresses true representation):

Abd 1.333 excess over seat quota (>5.666)
Clay (-1.666)
Warren (-2.666)
Sean 2.000
Raf 1.000

Only 7 voters total, of the 17 voting, are represented by Clay and
Warren. With myself, it could be said that 14 are represented, but
clearly three are not, the same vote as Warren got, and the two seats
besides myself represent the same number of votes as my single seat.
Of course, I can shift the results by reassigning.

positive numbers represent voters not fairly represented, negative
numbers represent seats with excess power over fair representation.

Asset, especially with small numbers of voters, doesn't work properly
unless the quotas and vote transfers are exact.

By no standard except the arbitrary one set up by Clay is anyone
elected, so far, but me, should I choose to actually serve (I could
transfer all my votes, if I think this best for ESF.) Do we really
want our first election to be plurality-at-large? Without adequate
opportunity for reassignment, that's what it is.

Don't think that the problem here is with the personalities of the
candidates. Before I proceed, I want to hear from all the asset
holders. I will cooperate with them to generate the most
representative committee possible.

This is what I originally suggested: Hare quota. When two seats have
been settled, those two, by consensus, determine the method for
assignment of the remaining seat, aiming at maximum representation.
In no case, however, should the two select the third seat with the
opposition of a majority of the remaining assets.

We don't have two seats yet, just one. Using the Hare quota, the only
seat I could unilaterally create would be one for Clay. I prefer to
wait and see what the others want to do.

This is the thing about the Droop quota used with Asset. Clay is
second rank, and is only a quarter vote short of election by Droop
quota, and Warren is 1.25 vote short. However, suppose Sean and Raf
were to agree on Candidate X, and persuade X to serve, and me to
support X. X could be elected.

I will, however, temporarily assign 1.25 votes to Warren, which will
give him a Droop quota election, should he want it and should it be
decided that Droop is appropriate. That would leave me, if we stick
with Droop, an additional 1.5 votes, plus Sean and Raf have 3 votes
together. If Sean and Raf were to agree on one among them to serve,
or another to elect, that would resolve the election, as would also
happen if anyone transfers to Clay a quarter-vote.

Frankly, I have always found election day a bit of a letdown. All
that fuss, and all we get is a some raw numbers and elections. With
Asset, the drama has just begun, a process to expand representation.

Bottom line: Because it's been said that we should "eat our own dog
food," I'm not prepared to accept an election by less than, at least,
a Droop quota, which will likely give results *close* to fair. One
seat, perhaps, gets a bit of extra clout. With two seats with extra
clout, it's clearly unfair.

By the way, to those who expressed their confidence in me, thanks. I
appreciate it. I hope all had a good Thanksgiving holiday, for those
in the U.S., and just plain a good week, for the rest.

[Abd ul-Rahman Lomax]

Clay has a conflict of interest with respect to declaring the results
of the election, and should stand aside, unless we have substantial
consensus on the results (in which case the conflict would be merely
formal and moot). Electing the "richest three" is plurality-at-large,
with only a very limited time allowed to make it possibly better than that.

[Abd ul-Rahman Lomax]

No. Droop is already a majoritarian compromise in the last seat. To
declare an election with a candidate having less than a Droop quota
is election by plurality, not even majority.

I have transferred sufficient votes to Warren to allow his election
with the Droop quota, as well as my own. If the period of the
election is determined to have lapsed, the remaining seat has failed
to be elected. However, we will have, at that point, elected enough
members of the committee to begin to function with a limitation:
consensus of the two members. It would then be up to Warren and I to
determine the next step. I would prefer to find agreement among the
remaining electors, and I would contribute my remaining Assets to
facilitating that, if needed. Ideally, we do complete a Hare
election, with every voter feeling represented in the resulting panel.

This may not be a difficult task at all, but it's my desire to make a
point by pursuing it. I'd rather the remaining seat be elected by a
broad a consensus as possible among the remaining electors with
excess assets: myself, Clay, Raph, Sean. Clay does not quite have a majority.

excess, unused assets:

Abd 1.5 (after 1.25 votes transferred to Warren)
Clay 4
Sean 2
Raph 1

total votes unallocated: 8.5
majority: >4.25

[Abd ul-Rahman Lomax]

At 12:48 PM 11/27/2009, Sean Walker wrote:

Sure. However, the raw votes only elected one seat with even a Droop
quota. I have remedied that by ensuring that, if we use Droop, we
have two elected, myself and Warren. Droop amounts to a majority
election for the final seat. If a majority is required! And haven't
we been around and around the desirability of that as a minimum?

As I've stated, I prefer using the Hare quota, and then working to
find the widest possible consensus with the last seat elected. A
committee can function when it is a seat short. Here, likely because
of there being so few seats being elected, we don't even get two
seats with Hare, but I assume we'd get there quickly.

[Abd ul-Rahman Lomax]

At 03:43 PM 11/27/2009, clay shentrup wrote:
>Why don't you guys let me take you out for martinis, and we can talk
>about how best to redistribute your asset? :)

All's fair in love, war, and elections. Unless these electors made
promises not to sell their votes for a cheap shot.

[Warren Smith]

well, apparently the election is now clinched by Warren, Raph=Ivan, and Abd,
thanks to transfers from Clay to Ivan and Abd to Warren.
(Isn't that right?)

By the way, about Sean being a high-school student, I did not mean to
demean him,
it's actually impressive; and we need young blood so he's desirable.
But on the minus side, I think he's presumably less desirable as a
decision-maker, since less experience than any of us.

--
Warren D. Smith
http://RangeVoting.org <-- add your endorsement (by clicking
"endorse" as 1st step)
and
math.temple.edu/~wds/homepage/works.html

[Abd ul-Rahman Lomax]

At 06:02 PM 11/29/2009, clay shentrup wrote:

><<<<<<<<<<<< NOTICE OF TRANSFER >>>>>>>>>>>>>>>>
>
>I transfer all 4 of my votes to Ivan/Raph

This elects Ivan/Raph, thus, unless something changes, these will be
declared elected:

Abd, Warren, Raph.

I would ask Sean to formally assent to this, if he cares to, and my
congratulations to him for his showing in the election. That would
make the election effectively unanimous, which would be nice for starters.

Thanks, Clay. Asset can work this way. I could also have decided to
elect rather than serve, and I considered it.

All pending matters and issues which have been raised by this
election, which have been interesting, can be considered and decided
by the elected committee, but I assume that we will continue to
consult as broadly as possible.

[Raph Frank]

On Sun, Nov 29, 2009 at 11:02 PM, clay shentrup
<thebrok...@gmail.com> wrote:
> Of course, using the Score Variants scheme does require
> some software (manually is possible, but a little too demanding of
> people's time), which I'll hopefully have time to write in the not-too-
> distant future.

Well, that was the plan when I updated the phpBB software.

> Probably my biggest problem with RRO is the same as my problem with
> Abd's ratification notion.  The basic problem is that whatever number
> you choose is arbitrary. If you say "2 of 3 representatives must
> approve this decision for it to be ratified", it's not as if you have
> some Bayesian Regret metric that says 2 of 3 is better than 3 of 3, or
> vice versa.

Normally, it is absolute majority of the members or 2/3 of those
present. The 2/3 rule is just a simple super-majority.

However, the reason for absolute majority is that it means that the
votes of those who aren't present couldn't change the result.

The idea of having 2/3 required to end discussion is that it prevents
a majority from ignoring a minority (but it can still close discussion
with any absolute majority anyway).

> But deliberation only has so much potential to inform and
> allow reps time to re-evaluate the logic behind their views (aka
> "utility assessments").

Also, it gives them time to see how other people view the problem, and
if there is "community spirit", then they might change their mind
about their preferences.

I think your concept that non-voters should be assumed to vote for the
status quo is similar to the principle that discussion shouldn't just
end.

> Whereas asking people whether they want to accept
> the SV winner or hold a re-election, and then acting on the basis of
> quantity rather than intensity, may actually diminish the revealed
> preference ability of Score Voting.

If a reasonable number of voters accept that score voting is good for
the group, then the would vote for the SV winner even if they prefer
another option.

This is also a revealed preference. A voter who votes against the SV
winner in the 2nd round places a stronger value on his candidate
winning than on the general utility in the group. This is
demonstrating strength of feeling against the SV winner.

> It suggests to me that Abd would work to enact policies which serve
> his intuitions more than any scientific indicators of effectiveness.
> Call me left-brained, but I do not hold intuition in very high
> regard.  I want numbers and facts.

However, we don't have simulations which cover the issue at question.

The sims would have to have some voters who value the group more than
achieving their personal favourite.

For example, the process could be

Determine the SV winner and the condorcet winner. These 2 candidates
go to the 2nd round and each voter can vote for 1 or other.

In theory, the condorcet winner should win in any head to head. It is
unclear what would happen in practice.

> Getting Score Voting used in government just
> means so much that I've felt compelled to help in any way possible,
> even for the 1 in a million chance that I could make a difference.
> It's kind of like the irrationality of voting -- what are the odds
> that my effort will make a difference?
>
> But I'm eating into my 30's now,

Yeah, that feeling sucks :).

> You only live once,
> so I've got to try to make these dreams of mine happen.

Good luck, there is a life out there.

I hope you are not planning to go completely cold turkey?

> <<<<<<<<<<<<  NOTICE OF TRANSFER >>>>>>>>>>>>>>>>
>
> I transfer all 4 of my votes to Ivan/Raph
>
> <<<<<<<<<<<<<<<< END >>>>>>>>>>>>>>>>>>>

Well, as I said, not a massive amount I can do from this side of the
Atlantic (especially in an official capacity), but have no problem
being involved in discussions on the best way forward for the
provisional board/parliament.

(and thanks for the vote of confidence)

Certainly any official board would probably need to be present in the US.

As I said, my primary reason for standing was to set the precedent
that as many people as possible should stand in Asset elections.

I do think you were good at outreach and that is something we would lose.

> In so doing, I ask that he keep my Score Variants proposal as a
> serious consideration, perhaps suggesting improvements to it or even
> writing some prototype code to implement it.  And I hope that he will
> keep Abd's interminable philosophical rants in check with the rigor of
> hard science and mathematics. :)

Well, so far I prefer the ongoing poll approach rather than the secret
ballot at the end.

However, there could be a final SV ballot to finally make the decision
(with more than just 2 options).

With the on-going poll, the rule could be that discussion is
maintained for as long as the score for the status quo is over 50%
(and members who haven't voted are considered to vote max for the
status quo). This allows a majority to prevent discussion from ending
on an issue. With Robert's rules, the status quo would need to drop
below 1/3 (or there would need to be a separate end discussion option
which would need to go above 2/3 .. or even a "ignore the rules and
force discussion to end immediately" that just needs to go above 50%).

I think that even with the ongoing poll method, people still wouldn't
bullet vote.

> I hope that makes sense and doesn't seem rash to you folks.

Well, you have to do what you have to do.

[Abd ul-Rahman Lomax]

At 08:24 PM 11/29/2009, you wrote:
>well, apparently the election is now clinched by Warren, Raph=Ivan, and Abd,
>thanks to transfers from Clay to Ivan and Abd to Warren.
>(Isn't that right?)
>
>By the way, about Sean being a high-school student, I did not mean to
>demean him,
>it's actually impressive; and we need young blood so he's desirable.
>But on the minus side, I think he's presumably less desirable as a
>decision-maker, since less experience than any of us.

My intention would be to specially invite Sean to participate with
the committee.

Interesting how the election played out, eh?

[Sean Walker]

> This elects Ivan/Raph, thus, unless something changes, these will be
> declared elected:
>
> Abd, Warren, Raph.
>
> I would ask Sean to formally assent to this, if he cares to, and my
> congratulations to him for his showing in the election. That would
> make the election effectively unanimous, which would be nice for starters.

I, Sean Walker, hereby formally assent to the stated result of the
election of Abd, Warren, and Ralph.
Congratulations, all of you.

> By the way, about Sean being a high-school student, I did not mean to
> demean him,
> it's actually impressive; and we need young blood so he's desirable.
> But on the minus side, I think he's presumably less desirable as a
> decision-maker, since less experience than any of us.

No offense taken. Being as young as I am, of course I'm the least
experienced of this fine group. It's great to see the fruits of our
democracy, in a way that nobody likely suspected. Besides, it's not
like not being elected is such a big problem. It's simply a part of
our democratic process, and I look forward to our group getting things
done now that we have our committee.

[Abd ul-Rahman Lomax]

At 08:40 PM 11/29/2009, Raph Frank wrote:

>Certainly any official board would probably need to be present in the US.

The requirement may be only that a majority of board members for a
U.S. nonprofit corporation are in the U.S. I assume we will check.

>As I said, my primary reason for standing was to set the precedent
>that as many people as possible should stand in Asset elections.

Right. Vote-splitting doesn't affect asset elections, or not much.
They are also unpredictable, as we saw. We should work on Asset
process. My concept was *not* that there would be a specific
deadline, that reduces it to plurality, in fact, with only a bit more
flexibility under pressure. I'm coming to prefer Hare quota, as I've
said, at least until some impasse develops, and with some possible
practical systems, electors would continue to be able to vote.

*For advisory purposes*.

(Secret ballot elections interfere with negotiations. Electors can
negotiate openly, and can be approached to suggest compromises. So,
what I'd see is periodic asset elections that maintain the body of
electors, then that body continues to exist as an advisor to the
board. Not terribly necessary when small, but not difficult to do
either. Thus, if we want to, for the near future, estimate overall
position on some idea, Clay would still have his 4 votes, but Raph
would routinely exercise them unless Clay specifically casts them on
some question. The legal authority, over property or with respect to
legal responsibilities, rests with the formally elected board. But
what organization's board would completely neglect an advisory board
that collectively represented *all* the members sufficiently
interested to vote?)

[Sean Walker]

On Nov 29, 7:00 pm, Abd ul-Rahman Lomax <a...@lomaxdesign.com> wrote:
> My intention would be to specially invite Sean to participate with
> the committee.
>
> Interesting how the election played out, eh?

I would certainly be willing to provide my input in regards to
committee decisions, and I'm honored to know that you thought to
specifically invite me, although I suspect that such decisions would
be debated in public anyway, and as such I would provide input anyway.

Certainly a good way of showing how asset shakes things up a bit.

[clay shentrup]

On Nov 29, 5:40 pm, Raph Frank <raph...@gmail.com> wrote:
> > But deliberation only has so much potential to inform and
> > allow reps time to re-evaluate the logic behind their views (aka
> > "utility assessments").
>
> Also, it gives them time to see how other people view the problem, and
> if there is "community spirit", then they might change their mind
> about their preferences.

What do you mean "also"? That is encompassed by what I just said.
Voters can attain more data and redo their assessments of their data.
That data could include information about "community spirit".

> > Whereas asking people whether they want to accept
> > the SV winner or hold a re-election, and then acting on the basis of
> > quantity rather than intensity, may actually diminish the revealed
> > preference ability of Score Voting.
>
> If a reasonable number of voters accept that score voting is good for
> the group, then the would vote for the SV winner even if they prefer
> another option.

Why on Earth would you believe that? Strategic voting is real. It
happens all the time.

The goal of a good social choice mechanism is to behave well even when
voters try to game the system. Not to put in place a rule that
plausibly HARMS the quality even if used by sincere voters, and then
expect them to be sincere and not harm the system even worse by being
tactical.

If a majority is given the option to thwart a more passionate
minority, they might use it. So why even give them the option? Where
is the evidence that it's a good idea?

> This is also a revealed preference.  A voter who votes against the SV
> winner in the 2nd round places a stronger value on his candidate
> winning than on the general utility in the group.

So if a voter reveals to you that he would like the group to suffer so
that he can be happier, why would you grant his wish?

> This is
> demonstrating strength of feeling against the SV winner.

Not really. You don't have to demonstrate any specific amount of
utility. All you have to do is care more about your own happiness than
the group's -- that doesn't necessarily say anything about your
utility values.

> > It suggests to me that Abd would work to enact policies which serve
> > his intuitions more than any scientific indicators of effectiveness.
> > Call me left-brained, but I do not hold intuition in very high
> > regard.  I want numbers and facts.
>
> However, we don't have simulations which cover the issue at question.

Exactly. So then we can only guess about the Bayesian Regret of some
approval process. As high as SV's utility efficiency already is, it's
probable that this step will make things worse, not better.

> Good luck, there is a life out there.

But I can't find it because I'm typing tomes about voting theory while
my girlfriend frowns at me for not being social. :)

> I hope you are not planning to go completely cold turkey?

I went out of my way to indicate that won't be the case. But I may go
to a point where I'm only reachable by phone or direct email. That is,
if you want me to pitch in with some paperwork filing, computer
programming, or cash donations, contact me -- otherwise I'm not in the
loop.

> Well, as I said, not a massive amount I can do from this side of the
> Atlantic (especially in an official capacity), but have no problem
> being involved in discussions on the best way forward for the
> provisional board/parliament.

I don't think your location will matter too much. It will be more
about research and planning, and maybe even coding.

> (and thanks for the vote of confidence)

Don't get a big head.

> Certainly any official board would probably need to be present in the US.

This isn't a board per se. We're not incorporated yet.

> Well, so far I prefer the ongoing poll approach rather than the secret
> ballot at the end.

Like I said, the reps are free to poll each other to their hearts'
content prior to the SV election. This leaves it up to them to decide
whether they think that will be helpful to poll each other.

A series of polls would seem to converge to the Condorcet winner,
which is WORSE than the SV winner if we assume fairly sincere voters
otherwise (which I think we can, since that is more advisable with a
small number of voters). And that's optimistic -- what if the reps
decide to mislead each other with those polls?

I just don't see any scientific basis for thinking that imposed
polling sequences are helpful.

> However, there could be a final SV ballot to finally make the decision
> (with more than just 2 options).

Uhhh... that's what I said. :)

> I think that even with the ongoing poll method, people still wouldn't
> bullet vote.

The problem isn't bullet voting -- it's tactical voting.

I think you'd have to be crazy to think people would just keep voting
sincerely instead of tactically if given the information to do so.

[Raph Frank]

On Mon, Nov 30, 2009 at 6:55 AM, clay shentrup
<thebrok...@gmail.com> wrote:
> What do you mean "also"? That is encompassed by what I just said.
> Voters can attain more data and redo their assessments of their data.
> That data could include information about "community spirit".

The SV result may show that an unexpected option has support.

> Why on Earth would you believe that? Strategic voting is real. It
> happens all the time.

Well, it depends on how large the relative utility of the 2 candidates is.

In a Presidential election, you pretty much get to pick a Democrat or
a Republican. For most people, there is a big difference between the
2 candidates. Thus people use strategy (and plurality is a big deal
here). However, even then a few percent "throw their vote away".

However, it is likely that the condorcet winner and the SV winner
would be very similar in utility (and probably the same person).

Thus many would be willing to vote for group unity over the tiny
difference in utility between the 2 candidates.

This is especially important in a voluntary organisation.

> The goal of a good social choice mechanism is to behave well even when
> voters try to game the system.

The system would degrade to selecting the condorcet winner instead of
the SV winner. This is still a pretty good method of picking the best
variant. Thus the system degrades gracefully.

You could also add a rule that the SV results are used to break a
circular condorcet tie. This is in fact a proposed method for doing
it that has some beneficial characteristrics, in terms of finding an
honest condorcet winner.

>> This is also a revealed preference. A voter who votes against the SV
>> winner in the 2nd round places a stronger value on his candidate
>> winning than on the general utility in the group.
>
> So if a voter reveals to you that he would like the group to suffer so
> that he can be happier, why would you grant his wish?

If that is what the majority wants. I might decide to leave the group
if that happens though.

Also, it could just indicate his strength of feeling on the issue. SV
doesn't necessarily find this due to normalisation.

>> However, we don't have simulations which cover the issue at question.
>
> Exactly. So then we can only guess about the Bayesian Regret of some
> approval process. As high as SV's utility efficiency already is, it's
> probable that this step will make things worse, not better.

Unfortunately, the answer to this question is parameter dependent.

For example, you could have each user's utility function updated after the SV to

U_new(X) = A*(U_old(X)) + (1-A)*(SV(X))

where
U_old(X) is the voters utility for X
SV(X) is candidate X's average score

U_old(X) would have to be scaled so it matches whatever the voter used
on the ballot (some candidates might end up with better than max score
etc.)

Anyway, if A is 0 (altruist), the the SV candidate will win, and if A
is 1 (selfish), then the condorcet winner will win.

It could turn out that with an A of 0.05, that the SV winner almost
always wins. Also, thinking about it, the fact that some people vote
for 3rd party candidates shows that A is not 0, in RL. However, that
is related to honesty in voting rather than valuing the group
consensus.

Also, with spherically symmetric voters, the condorcet winner is the
same as the SV winner (subject to random noise). It might actually be
hard to set up a voter configuration where the 2 are different.

I think with random utility for each candidate, the 2 are the same too
(maybe Warren can comment), as it is effectively spherically symmetric
just with one dimension for each candidate.

Under strategy, SV is actually a pretty reasonable condorcet finding method.

>> (and thanks for the vote of confidence)
>
> Don't get a big head.

If I had managed to sabotage Warren's campaign (secretly :) ), them I
would have received his votes too.

> A series of polls would seem to converge to the Condorcet winner,
> which is WORSE than the SV winner if we assume fairly sincere voters

Under strategy, the 2 methods are the same.

However, it might turn out that there is 1 option with 52% support and
another with 68% support. The group might deicde to go with the 68%
one, even though the 68% support one isn't actually a condorcet
winner.

>> However, there could be a final SV ballot to finally make the decision
>> (with more than just 2 options).
>
> Uhhh... that's what I said. :)

Yeah I know, I meant I think it is a reasonable thing to do. However,
I think by the end there would only be 1 option that is seriously
considered.

> I think you'd have to be crazy to think people would just keep voting
> sincerely instead of tactically if given the information to do so.

Well, some honest voters will help tip the balance. In fact, if we
assume pure strategy, we might as well just use approval voting.

Would you zero out an option that you though was reasonably good in
order to force a condorcet victory for some other option?