Special Resolution 9

[Dave Mansfield]

This is the three changes to the Articles.

* Change to membership removal process
* Removal of the "no further places candidate"
* Replace reference to general meeting, with directors election.

I'm not sure I feel the same way about each change. Is it too late to unbundle them?

[A S]

Other organisations have a procedural motion "That the current motion be voted on in parts", but that is in their constitution, so I'm not sure if there is anything we can do now. I guess it will be a case of all or nothing.

--
You received this message because you are subscribed to the Google Groups "London Hackspace" group.
To unsubscribe from this group and stop receiving emails from it, send an email to london-hack-sp...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Russ Garrett]

I would be happy to take a motion to amend the changes at the meeting.

Can I ask which of the changes concerns you? I will have a full
justification for these changes shortly.

Russ

> --
> You received this message because you are subscribed to the Google Groups
> "London Hackspace" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to london-hack-sp...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

--
Russ Garrett
ru...@garrett.co.uk

[A S]

Russ,

I presume then that resolutions 1-3 will not be open to amendment, whereas the others will be?

[Russ Garrett]

I will take any amendments provided they are not frivolous, vexatious,
or defamatory.

Russ

On 16 December 2015 at 14:32, 'A S' via London Hackspace

[Dave Mansfield]

It's the middle one that I'm less happy with.

I'm a fan of the: "I'd rather leave this position unfilled than have X take this position"

It's nuclear option, but useful when there's not enough candidates, or other shortlist problems.

As such I've voted that way in university sabbatical elections, and in the puppygate fiasco of this year's Hugos Awards.

[Martin (Crypt)]

I think the issue with this is its confusing to people.  The way its worked in previous elections is that the 'No further places' option was considered another candidate, so unless that candidate got a significant number of votes, which has never been the case, those votes are carried to the next candidate, even if this was not the voters intention.

Given that our voting is handled by an external service, this leaves 3 options.  

1) Write a new voting system to replace this one.  This would never be truly independent, and although we could make it secure, there may be accusations that the voting system is run by members with an conflict of interest

2) Find a better service.  I've not looked for this, but I assume that will be difficult

3) Remove the 'No Further Places' option, and make it clear to voters how the system works.

It seems like option 3 is the one preferred at the moment, and its probably a good option if no other better service can be found.

[David Murphy]

I think the problem re the middle one is worse in the current state.

Currently "no further places" appears to not do what most people would intuitively believe it does: stop moving your vote to the next person on the list. I was surprised by this myself, I thought the order in which I placed candidates after "no further places" was irrelevant but what actually happens is that "no further places" is treated like a candidate similar to "re-open-nominations" and your vote continues to be moved down your list to candidates.

I gather it would significantly complicate/change the voting system to align the behavior with intuition so eliminating no-further-places seems preferable on the basis that unexpected behavior should be considered a bug.

In my opinion:

worst:current state.

better:removing the option.

best:changing voting system such that votes die with no-further-places.

Anyone please correct me if I'm factually incorrect in the above.

On Wed, Dec 16, 2015 at 2:45 PM, Dave Mansfield <nos...@davemansfield.com> wrote:

[Mr Ed]

If you need someone to second the motion for unbundling these, I second the motion for the same reasons.

I would rather keep the "no further places" wording for the same reasons. Though I'd be interested to hear the reasons for the change, as there may be something I've not considered and so might change my opinion.
-Ed

[Russ Garrett]

Right, I do (or at least did) agree with you which is why I originally
introduced the "No Further Places" option. However, my mind has been
turned and I now disagree with it. I'll try and state why:

It doesn't work. Surely given all the controversy about the most
recent election, NFP would have had a reasonable showing? It was
actually the first candidate to be eliminated. It's not a special
candidate, so it invariably gets eliminated in the first round as very
few people put it as their first preference. I am having difficulty
constructing any plausible election where it would not be eliminated.
It's entirely methodologically unsound. I was a prat for suggesting it
in the first place.

It is confusing. For every election I field dozens of questions and
complaints from people who don't understand what it does, or don't
understand whether ranking candidates after NFP makes a difference (it
does, but not much). I think the majority of voters don't understand
how to use it correctly. STV is a complex system to begin with, but
NFP makes it even more complex.

Jeff from OpaVote/OpenSTV also has similar reckons about this, and he
knows more about voting systems than me (note that the implementation
he's describing is slightly different from the one we use, but the
conclusions are the same):

http://blog.opavote.com/2015/09/just-say-no-to-none-of-above.html

I do have some more ideas about how to improve the trustee voting
system, however they will not be ready in time for this EGM. I'm going
to try and write them up afterwards.

Russ

[Ben Clifford]

On Wed, 16 Dec 2015, Mr Ed wrote:

> I would rather keep the "no further places" wording for the same
> reasons. Though I'd be interested to hear the reasons for the change,
> as there may be something I've not considered and so might change my
> opinion.

(From my personal interest in voting systems,) what do you think voting
for "no futher places" does? (eg to your preferences ranked below NFP on
your ballot; and if NFP is elected say position 3 out of 4 vacancies)

--

[Tim Reynolds]

Hi Dave,

Do you mean your understanding of it is "any candidated ranked below NFP
don't get my vote at all" ?

If so, that is incorrect and a widely held belief. It's resulted in some
people being very confused with how votes are transferred (myself included)

On 16/12/2015 14:45, Dave Mansfield wrote:
> It's the middle one that I'm less happy with.
>
> I'm a fan of the: "I'd rather leave this position unfilled than have X
> take this position"
>
> It's nuclear option, but useful when there's not enough candidates, or
> other shortlist problems.
>
> As such I've voted that way in university sabbatical elections, and in
> the puppygate fiasco of this year's Hugos Awards.
>
> On Wednesday, 16 December 2015 14:29:59 UTC, Russ Garrett wrote:
>
> I would be happy to take a motion to amend the changes at the meeting.
>
> Can I ask which of the changes concerns you? I will have a full
> justification for these changes shortly.
>
> Russ
>
> On 16 December 2015 at 14:18, Dave Mansfield

> <nos...@davemansfield.com <javascript:>> wrote:
> > This is the three changes to the Articles.
> >
> > * Change to membership removal process
> > * Removal of the "no further places candidate"
> > * Replace reference to general meeting, with directors election.
> >
> > I'm not sure I feel the same way about each change. Is it too
> late to
> > unbundle them?
> >
> > --
> > You received this message because you are subscribed to the
> Google Groups
> > "London Hackspace" group.
> > To unsubscribe from this group and stop receiving emails from it,
> send an

> > email to london-hack-sp...@googlegroups.com <javascript:>.

> > For more options, visit https://groups.google.com/d/optout

> <https://groups.google.com/d/optout>.
>
>
>
> --
> Russ Garrett
> ru...@garrett.co.uk <javascript:>

>
> --
> You received this message because you are subscribed to the Google
> Groups "London Hackspace" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to london-hack-sp...@googlegroups.com

> <mailto:london-hack-sp...@googlegroups.com>.

[Ben Clifford]

On Wed, 16 Dec 2015, Tim Reynolds wrote:

> Hi Dave,
>
> Do you mean your understanding of it is "any candidated ranked below
> NFP don't get my vote at all" ?
>
> If so, that is incorrect and a widely held belief. It's resulted in
> some people being very confused with how votes are transferred (myself
> included)

I think there's an option in opavote to have truncated ballots, where you
don't have to list everyone.

But I think in the system being used, "any candidate ranked below NFP
didn't get my vote at all" is the same as "I will allow the rest of the
hackspace electors to rank you remaining candidates, because I am not
going to" rather than something stronger/more negative.

--

[Dave Mansfield]

My understanding was that "NFP" is treated like any other candidate.

So if the result of the election for 4 seats is: A, B, NFP, C, D, E. Then A and B are elected, and 2 seats remain empty.

> <mailto:london-hack-space+unsub...@googlegroups.com>.

[Russ Garrett]

On 16 December 2015 at 15:17, Ben Clifford <be...@hawaga.org.uk> wrote:
> I think there's an option in opavote to have truncated ballots, where you
> don't have to list everyone.

Yes, and that is certainly a possibility for us to use next year. I
did decide that combining that with NFP would be too complex though.

I processed the votes from the recent election to truncate the ballots
after NFP, but it didn't change the results.

--
Russ Garrett
ru...@garrett.co.uk

[Tim Reynolds]

Right! Sorry, discussing this is a nightmare.

> > an email to london-hack-sp...@googlegroups.com <javascript:>
> > <mailto:london-hack-sp...@googlegroups.com

> <javascript:>>.
> > For more options, visit https://groups.google.com/d/optout
> <https://groups.google.com/d/optout>.
>
> --

> You received this message because you are subscribed to the Google
> Groups "London Hackspace" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to london-hack-sp...@googlegroups.com

> <mailto:london-hack-sp...@googlegroups.com>.

[Dave Mansfield]

One scenario I have in mind, is where the number of candidates nominated <= number of posts. Hasn't happened yet, but it could do in the future..

Especially with Special Resolution 5, this removes a final safeguard.

I agree with Jeff's post that actively recruiting candidates is a better solution, but that's not something that can be catered for in the rules.

To try to prevent confusion, could one of the other NOTA/SFC/RON/ACS names form Jeff's post be used instead? 

[Mark Steward]

The problem with the last option (as I see it) is that expired votes then have no input into later rounds. Unpopular people still get voted in, just less predictably, and anyone who thinks the list of candidates is inadequate gets a weaker vote. Unfortunately, I think people prefer this because impotent sufferance is more socially acceptable than having to quantify opposition.

We don't want our voters to feel put off by the voting system, but in my view the fairest option on all counts is to ensure people rank all candidates. Perhaps "No Further Places" would work if renamed to an unelectable "Like/Dislike Boundary"? Or, to solve the expired vote problem, if we combined choosing candidates to rank *and* No Further Places?

Mark

[Ben Clifford]

On Wed, 16 Dec 2015, Dave Mansfield wrote:

> My understanding was that "NFP" is treated like any other candidate.
> So if the result of the election for 4 seats is: A, B, NFP, C, D, E.
> Then A and B are elected, and 2 seats remain empty.

That's not like any other candidate!

If NFP is like any other candidate, then it takes up its position on the
board alongside A, B and C. Somehow. Maybe a doll?

--

[Mark Steward]

I believe the proposed change merely removes the requirement for NFP. It doesn't prevent us adding more options.

Mark

[Russ Garrett]

On 16 December 2015 at 15:22, Dave Mansfield <nos...@davemansfield.com> wrote:
> My understanding was that "NFP" is treated like any other candidate.
>
> So if the result of the election for 4 seats is: A, B, NFP, C, D, E. Then A
> and B are elected, and 2 seats remain empty.

Yes, in theory that is how it should work.

In practice, the way STV works makes this extremely unlikely to
happen. NFP will naturally get a very small share of top-preference
votes, which means it has a high chance of being eliminated in the
first few rounds. Once it's eliminated, it can't do anything.

--
Russ Garrett
ru...@garrett.co.uk

[Dave Mansfield]

Being extremely unlikely to happen is a feature, not a bug, in my opinion. Generally we'd rather have someone fulfilling the role, than leaving it empty. 

It is a last resort. Especially if not enough candidates stand.

[Tim Reynolds]

The barrier to standing is so low, and the vote order is randomised. I'm
not sure anyone is better than no one.

On 16/12/2015 15:59, Dave Mansfield wrote:
> On Wednesday, 16 December 2015 15:30:08 UTC, Russ Garrett wrote:
>
> On 16 December 2015 at 15:22, Dave Mansfield

> <nos...@davemansfield.com <javascript:>> wrote:
> > My understanding was that "NFP" is treated like any other candidate.
> >
> > So if the result of the election for 4 seats is: A, B, NFP, C, D,
> E. Then A
> > and B are elected, and 2 seats remain empty.
>
> Yes, in theory that is how it should work.
>
> In practice, the way STV works makes this extremely unlikely to
> happen. NFP will naturally get a very small share of top-preference
> votes, which means it has a high chance of being eliminated in the
> first few rounds. Once it's eliminated, it can't do anything.
>
>
> Being extremely unlikely to happen is a feature, not a bug, in my
> opinion. Generally we'd rather have someone fulfilling the role, than
> leaving it empty.
> It is a last resort. Especially if not enough candidates stand.
>

> --
> You received this message because you are subscribed to the Google
> Groups "London Hackspace" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to london-hack-sp...@googlegroups.com

> <mailto:london-hack-sp...@googlegroups.com>.

[Ingvar Mattsson]

On Wednesday, 16 December 2015 15:30:08 UTC, Russ Garrett wrote:

The "hack" that the Hugo voting does (for "no award", rather than 

"no further places") is that after a potential winner is selected, the 

ballots are divided into the two piles "winner before (possible) No 

Award" and "No Award before (possible) winner" and the potential 

winner needs to have a larger pile. This could, in theory, be 

extended to work with multiple elections (for each elected person, 

do the division across "all votes", on not passing the bar, they're 

simply not elected). 

I don't know to what extent this is something that would be simple 

to do in OPA, nor am I sure how easy it is to understand. But I 

seem to recall I spent a fair bit of time in Spokane to explain the 

Hugo voting system (and some proposed amendments) to people, 

and that is one I find relatively easy to understand.

[David Murphy]

@Dave

Imagine a scenario where there are 150 voters and 100 have very strong feelings about only wanting their favorite single candidate.

those 100 all vote evenly for 5 different candidates first and put NFP second and they hate every other candidate.

NFP gets zero first place votes and is eliminated as a candidate.

The 5 different candidates with 20 votes each get eliminated and all 100 votes trickle down past NFP(no longer existing) to other candidates.

Indeed I'm pretty sure it would be possible to construct a "worst-case" where every single elected candidate is ranked bellow every voters NFP without NFP reaching the threshold.

So I don't think it works well for your goal.

--

You received this message because you are subscribed to the Google Groups "London Hackspace" group.

To unsubscribe from this group and stop receiving emails from it, send an email to london-hack-sp...@googlegroups.com.

[Dave Mansfield]

On Wednesday, 16 December 2015 16:08:16 UTC, David wrote:

@Dave

Imagine a scenario where there are 150 voters and 100 have very strong feelings about only wanting their favorite single candidate.

those 100 all vote evenly for 5 different candidates first and put NFP second and they hate every other candidate.

NFP gets zero first place votes and is eliminated as a candidate.

The 5 different candidates with 20 votes each get eliminated and all 100 votes trickle down past NFP(no longer existing) to other candidates.

Indeed I'm pretty sure it would be possible to construct a "worst-case" where every single elected candidate is ranked bellow every voters NFP without NFP reaching the threshold.

So I don't think it works well for your goal. 

If everyone votes in that way, then I'd agree that the STV vote reflects the collected will of the mass; i.e. we don't want anyone.

That scenario is however the exact opposite of the current situation, where "NFP" gets so few votes that it always gets eliminated first.

[David Murphy]

@Dave

But that's my point, in the scenario I outline NFP would get eliminated first and someone that almost everyone ranked after NFP could be elected. 

The worst-case is somewhat specific though.

--

[Yvan Janssens]

So, I am not sure if I classify as "what most people think", but reading this
thread made me realise that I totally misunderstood the NFP option, and that
the effective vote I casted was different than the intended vote I had.

I understood that the NFP option meant that all people below that option are
people I don't want to be trustee, and that my vote wouldn't be cast down to
those. As a result, I inuitively ranked the options I chose from least
desirable to most desirable. My ballot looked something like this:

* top choice
* second choice
...
< nfp >
* least desirable choice
* second to least desirable choice
...

Reading this thread did explain a lot about how the voting process works
though, so I really appreciate that this has been brought up - otherwise I
wouldn't have known that I am using the system the wrong way.

Y.