LP Texas & Approval Voting

[Arthur Thomas IV]

Hello all. I have been an advocate for Approval Voting for many years now. I am on the LP Texas state committee and currently serving as Rules chair for preparation for our 2016 state convention.

James Holland (both of us are in San Antonio) is a rules committee member as well and we want to work on establishing quite clear and simple rules stating absolute preference for using approval voting. I would really like to make a statement through LP Texas in how we function that AV is a better system. I believe the LP should be a strong advocate and hopefully other state affiliate will follow and AV will adopt a more solidly clear preference within the LP. I think if we are going to advocate for election reforms then we should be using it ourselves without question. 

We are still in review so nothing as been written yet but we have two major issues that we have to deal with. (Ignoring the fact that we have to get anything past the delegates but that is an internal party matter). 

1. NOTA: The LP uses None of the Above in our elections. We do not believe that just because candidates are on the ballot that one of them must be selected. I am a strong proponent of treating NOTA just as another option on the ballot although this might be a small hurdle against some that want to treat it as a special case (If you vote for NOTA you cannot make another selection). 
2. Majority Votes: I do not think majority requirements are necessary but I think we may have to add them in for some instances. 

Why I am writing the group is to see if anyone would like to give an advice or input into the matter. Our current rules have voting sprinkled throughout and when it comes to voting in conventions is a quite tangled mess. (if anyone is brave enough to look: http://www.lptexas.org/state-rules) I certainly don't mind discussing any draft version either later in case people want to point out any gotchas or caveats that we might not be aware of. We are very eager to see a straight forward and reliable voting system used as an example.

There was some unfortunate history at our last convention with using approval voting that put a "bad taste" in some peoples mouth. We had a very contentious election. We had a voting confirmation group come in (I forget their name) and assist with our voting. Them being confused by AV and our rules being so confusing caused a lot of issues and really delayed the vote counting to an extreme amount. This caused people to blame AV unfortunately. I am really hoping to get around that and get people comfortable with the idea before this next convention. 

Vendors and groups can get booths there but I am not in charge of that so I don't know how it works. If any of yall are near Texas (the 2016 convention will be in San Antonio in April) I hope you can come as well!

Also thanks to everyone that works for the center for election science and the website. The video is especially helpful for giving to people new to the idea. I can try to keep information updated if anyone has any interest. Thanks!

[Warren D Smith]

With approval (or score) voting, indeed virtually any voting system,
NOTA could simply be added as an extra "candidate"
just as capable of "winning."

It also is possible to treat NOTA asymmetrically, i.e. differently
than the real candidates,
by, for example, regarding NOTA as the winner if the candidate who otherwise
would have won, gets average score (or percent approving him) below T,
where T is some pre-agreed numerical threshold.

The latter incidentally could satisfy your "majority" desires, by,
e.g, with approval,
making T=50% so that NOTA would win if nobody can obtain majority approval.
I warn you that it is fairly common for nobody to obtain majority
approval, e.g. French presidential 2002:
http://www.rangevoting.org/FrenchStudy.html
so, if that is your desire, you better ask yourself... would it really
have been
better for France to elect nobody in 2002?

Also happened again in France 2007 (but much closer this time):
http://www.rangevoting.org/French2007studies.html

About approval & score voting,
http://www.rangevoting.org
http://www.rangevoting.org/Approval.html


--
Warren D. Smith
http://RangeVoting.org <-- add your endorsement (by clicking
"endorse" as 1st step)

[Arthur Thomas IV]

Very interesting. I had not thought about NOTA being a 'failed majority' type possibility. I can already very confidently say that people would be against that. I would advocate NOTA being an affirmative choice for sure. I will pass this by others though and see what they think! Thanks.

Yeah I think non majorities would be more natural and I am against the false idea of majority requirement. Non majorities, when they appear, are just an honest representation of peoples choices. If people feel uncomfortable putting a non majority in office then they shouldn't question the voting system in this case but the idea that they are electing people that appeal to so few people. 

NOTA has been always added as an extra "candidate". We have a long tradition of that. There is just a small battle of treating it as separated. Many wish to make it so that if you vote for NOTA then you cannot make another choice which violates the idea of AV and brings back game play into the system. This will be probably one of the hardest hurdles in convincing people. 

Libertarians are a wild bunch. Many of us would be ok with electing nobody if we could! :P haha.

Thanks so much. Definitely more to think about.

--
You received this message because you are subscribed to a topic in the Google Groups "The Center for Election Science" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/electionscience/sz-_YZSS-UU/unsubscribe.
To unsubscribe from this group and all its topics, send an email to electionscien...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Arthur M. Thomas IV

amtho...@gmail.com

[Jan Kok]

Hi Arthur,

Thanks for helping to spread the use of Approval Voting. Definitely, please do keep us informed.

LP Colorado has been using Approval Voting at its state conventions since about 2008.

My opinion is that NOTA should be treated as just another candidate as much as possible. If NOTA wins or ties for winner, no actual candidate is elected. I don't recommend using Approval Voting in multiple winner elections, but if you do use it that way, and NOTA wins a spot, any candidates who get fewer votes than NOTA should not be elected.

I don't see a need to prevent voters from approving NOTA plus other candidates in the same race. Libertarians are all about minimizing restrictions, right? :-) I can think of some (unlikely) situations where it might make sense to vote that way.

[Warren D Smith]

QUESTION: In LP elections, how often does NOTA actually win?
(Just interested.)

At one point I wrote an essay about how Score Voting, aka Range Voting,
was the "most libertarian voting method." Is it?
Or was I nuts?... Anyway, here is the essay:
http://www.rangevoting.org/ForLibs.html

[William Waugh]

Jan Kok "I don't recommend using Approval Voting in multiple winner elections,"...

Although there are things to be said in favor of other multiwinner methods (already under discussion in this forum), scoring Approval ballots as Reweighted Range Voting (RRV) seems to me as not among the worst ways to go.

By the way, I'd like to caution against abbreviating Approval Voting as "AV". Some writers use "AV" for "Alternative Vote" meaning IRV or STV or something like that in the UK context.

[Arthur Thomas IV]

I have no historical data on NOTA wins. I know that in 2012 national convention won several rounds on the chair vote I believe. (I could be wrong on the position). Even if NOTA doesn't threaten winning often it is an ever present option that many are very vocal about. It definitely a unique and proud tradition of the party. Its really our only option for keeping bad candidates off the ballot too. NOTA unfortunately has a weaker presence at our last state convention over rules confusion and not being able to choose NOTA along with another choice. The approval vote ballot (after a very long count) came out a tie. This caused another vote and another ridiculously long count. This had an unpleasant experience that was blamed on approval voting. 

I wish I could say the same excitement about approval voting and I am hoping to change that. It has been used for years but our rules make it difficult to understand to say the least. I would agree that approval voting is extremely libertarian in concept. Its the most expressive without restriction of options. Unfortunately states and counties may be very flippant in using it (again I blame our rules). The status quo is harder to counter even sometimes in the anti-status quo party though and people to default to certain voting habits.

I think your "philosophical interlude" was spot on. Its hard to discuss Duverger's Law with people though. 1) I am bad at it and 2) their eyes glaze over really fast haha. I continue to try!

Thanks for the advice William on the name. I will try to remember not to abbreviate to "AV". Its better marketing to get the name out there sure anyway so I should make a habit of it. 

--
You received this message because you are subscribed to a topic in the Google Groups "The Center for Election Science" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/electionscience/sz-_YZSS-UU/unsubscribe.
To unsubscribe from this group and all its topics, send an email to electionscien...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Jan Kok]

On Wed, Aug 12, 2015 at 9:41 PM, William Waugh <2knuw...@snkmail.com> wrote:

Jan Kok "I don't recommend using Approval Voting in multiple winner elections,"...

Although there are things to be said in favor of other multiwinner methods (already under discussion in this forum), scoring Approval ballots as Reweighted Range Voting (RRV) seems to me as not among the worst ways to go.

What I "don't recommend" is counting the approvals and declaring the top N most-approved candidates as winners.

Approval Voting ballots used as input to RRV is OK. It gives you proportional representation, so it's a reasonable method for choosing multiple winners to serve on a committee, for example. The main drawback to RRV is that it requires a program (that would take several hours to write and test) to do the tallying quickly and accurately. In contrast, plain Approval, Score, and "quasi-proportional" methods such as Cumulative Voting can be tallied trivially with a spreadsheet.

[Clay Shentrup]

I strongly agree with treating NOTA as another "candidate". In reality, it truly is an outcome or a "state of the world" to which a voter attaches a certain value.

E.g. maybe you have these sincere preferences:

Libertarian = 5

NOTA = 3

Republican = 1

Democrat = 0

A sensible tactical vote would be to vote NOTA and Libertarian (or even Republican+NOTA+Libertarian), just the same way you would if NOTA were an actual candidate.

San Francisco Bike Coalition (a pretty influential group that endorses candidates in local elections) does this, using a 0-9 Score Voting ballot. You can rate the NOTA option and all the actual candidates together.

[Clay Shentrup]

I would say that for multi-winner Plurality-ballot elections, best to worst is:

- RRV (aka Proportional Approval Voting)

- Approval Voting

- Plurality Voting

Consider a 3-winner election with left/center/right factions roughly equal in number, with several candidates from each faction.

RRV would elect Left, Center, and Right.

Approval Voting would elect Center 1, Center 2, and Center 3. Not as good, but still ideological centered in the same general position as the electorate.

Plurality Voting could conceivably elect Left 1, Left 2, Center—or Right 1, Right 2, Right 3! This would be WORSE than having all centrists, because then the ideological center would be severely skewed from that of the electorate.

[Arthur Thomas IV]

Is there a good example of Proportional Approval Voting somewhere? I haven't run across this term. I want to be sure I understand it. 

--

You received this message because you are subscribed to a topic in the Google Groups "The Center for Election Science" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/electionscience/sz-_YZSS-UU/unsubscribe.
To unsubscribe from this group and all its topics, send an email to electionscien...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Jan Kok]

On Wed, Aug 12, 2015 at 9:41 PM, Warren D Smith <warre...@gmail.com> wrote:

QUESTION: In LP elections, how often does NOTA actually win?
(Just interested.)

In the 2006 LP Colorado convention, there were a couple of elections in which NOTA came in second place. I think one of them was for the governor candidate. When she was questioned during the nominating/election process, she didn't know what the Second Amendment (to the US Constitution) was - a bit of ignorance that was concerning to many of the delegates. Anyway she won the nomination, but NOTA came in second place with something like 30% of the votes.

Now, Colorado election law says that if more than one candidate gets at least 30% of the votes, all those candidates have to appear on a primary ballot, to be voted on by party members in a statewide primary election. People were giggling about the possibility that this candidate would appear on a primary ballot against NOTA. The party chair duly reported the convention election results to the Secretary of State, who promptly ruled that NOTA was not a person, so no primary election was necessary for the LP's governor nominee.

[Warren D Smith]

.

I wish I could say the same excitement about approval voting and I am
hoping to change that. It has been used for years but our rules make
it difficult to understand to say the least. I would agree that
approval voting is extremely libertarian in concept. Its the most
expressive without restriction of options.

--well, I was trying to argue score voting aka range voting (i.e. each
voter scores each candidate on an 0-to-9 scale, highest average wins;
if a voter does not score a candidate then that candidate's average is
unaffected by that voter) is "the most libertarian" voting system,
more so than approval which insists on only the endpoints of the scale
as allowed scores.

--Approval voting at meetings can be done by go thru options, all
approvers hold up their hands (or a red card) for each, count the
hands. This takes essentially the same amount of time as a plurality
vote -- except for a plurality vote somebody might be cheating by
holding up hand 2 times, whereas with approval that's legal.

--"RRV" and "asset voting" are two multiwinner voting methods discussed here:
http://rangevoting.org/RRV.html
http://rangevoting.org/Asset.html
The whole area of multiwinner voting methods is still embryonic in my opinion.
Asset voting is simple, maybe the simplest good multiwinner system.
RRV is pretty complicated. It has been used for the "best visual
effects" OSCAR film award for several years.

[Clay Shentrup]

On Wednesday, August 12, 2015 at 11:04:03 PM UTC-7, Arthur Thomas IV wrote:

Is there a good example of Proportional Approval Voting somewhere? I haven't run across this term. I want to be sure I understand it.

https://www.youtube.com/watch?v=_jS7b-0PV9E

I unfortunately/ironically goofed at one point in that, but it was a quick and dirty demo. The underlying math in the spreadsheet was totally correct.

[Steve Cobb]

CES gathered most of the voting-related text of its bylaws into a single article, 3 Voting. You can find the bylaws here:

http://www.electology.org/#!transparency/c65g

Also a bit of “By-Law Amendment Language” is offered here:

http://www.electology.org/#!tools/c1jc4

However, it is all for single-winner decisions.

We expressly include NOTA. The majority requirement is imposed by Robert's Rules of Order and other parliamentary authorities, but we allow for the decision basis to be lowered.

[Jameson Quinn]

I realize this thread has gone on for a couple of days now, and there have been some good ideas. But without disparaging those ideas, I want to bring up some different ones. 

2015-08-12 17:06 GMT-04:00 Arthur Thomas IV <amtho...@gmail.com>:

Hello all. I have been an advocate for Approval Voting for many years now. I am on the LP Texas state committee and currently serving as Rules chair for preparation for our 2016 state convention.

This is great, and thanks a lot for reaching out.

 

James Holland (both of us are in San Antonio) is a rules committee member as well and we want to work on establishing quite clear and simple rules stating absolute preference for using approval voting. I would really like to make a statement through LP Texas in how we function that AV is a better system. I believe the LP should be a strong advocate and hopefully other state affiliate will follow and AV will adopt a more solidly clear preference within the LP. I think if we are going to advocate for election reforms then we should be using it ourselves without question. 

We are still in review so nothing as been written yet but we have two major issues that we have to deal with. (Ignoring the fact that we have to get anything past the delegates but that is an internal party matter). 

1. NOTA: The LP uses None of the Above in our elections. We do not believe that just because candidates are on the ballot that one of them must be selected. I am a strong proponent of treating NOTA just as another option on the ballot although this might be a small hurdle against some that want to treat it as a special case (If you vote for NOTA you cannot make another selection). 
2. Majority Votes: I do not think majority requirements are necessary but I think we may have to add them in for some instances. 

For both NOTA and majority votes, you might be interested in a median-based system. Here are the rules for a system that I'll call "majority choice approval"¹:

1. Voters grade each candidate using 3-6 categories. I'd suggest using either letter grades (A/B/C/D/F) or words (Excellent/Good/Acceptable/Unacceptable). The important thing is that the lowest category is considered "below NOTA", whereas the second-lowest is "above NOTA". Thus, NOTA is not a separate candidate.

2. To count the votes, get a tally for each grade for each candidate.

3. To see who won, first consider only the top grade as approvals, and see whether any candidate has majority approvals at that threshold. If not, lower the "approval threshold" (by adding in lower grade tallies) until some candidate(s) have a majority. 

4. If two candidates reach a majority together, the winner is whichever got the biggest majority.

(In voting theory terms, this is one of a class of similar systems called "graded median" or "modern Bucklin" systems. Terminology can be a bit confusing here. Other systems in this class include Majority Judgment, Graduated Majority Judgment, etc.)

Note that if no candidate has a majority above the lowest grade, then you can consider NOTA to have won.

Counting this system is a little bit more complicated than approval, but it can still easily be done with tally marks on paper and a bit of simple addition.

Why am I suggesting this rather than just adding a majority threshold to approval? Because I find that one of the biggest objections people have to approval is that, if they want to be reasonably strategic, it doesn't let them express who their true favorite. With this system, in the end, all ballots are counted as a 1 or 0 for each candidate, just as in approval, so you get all the theoretical advantages; but in the process, people get to be more expressive, which leaves them happier. It also tends to find a threshold where most voters end up having used optimal strategy.

 

Why I am writing the group is to see if anyone would like to give an advice or input into the matter. Our current rules have voting sprinkled throughout and when it comes to voting in conventions is a quite tangled mess. (if anyone is brave enough to look: http://www.lptexas.org/state-rules) I certainly don't mind discussing any draft version either later in case people want to point out any gotchas or caveats that we might not be aware of. We are very eager to see a straight forward and reliable voting system used as an example.

The system I'm suggesting is somewhat novel. However, I'd be willing to help you write the bylaws and/or the "how to" document if you're interested. Also note that it's quite similar to Bucklin voting, which was used in over a dozen US cities around 1920, and to Majority Judgment, which has been used in a number of competitions and/or private elections in France.

....

One other thing that's been brought up in this thread is whether some of the elections you're talking about might actually be multi-winner, and if so, what system would be good. I think that for internal party elections, where you want to develop some degree of unity, full PR might not be the best idea; but on the other hand, you want some representation for the largest factions. 

The MCA system I described above is easy to adjust to give this kind of result. (Side note: If you're willing to do ballot-reweighting, you can actually get a fully proportional system that reduces to MCA for single-winner; but I'm assuming you want something simpler.) Say you're electing 4 candidates. Then instead of lowering the threshold until some candidate has more than 1/2, you can just lower it until there's more than 1/5 for the first candidate, then 1/4 for the second, then 1/3, and then 1/2 (as with single-winner MCA) for the last. This is NOT a truly proportional system, because votes are not "used up" once they help elect one candidate; but (as long as there isn't a highly-organized minority facing a badly-divided majority), it will give approximately proportional results, but also guarantee that there are some winners with broader support.

Finally: If you're NOT interested in using MCA, and you do want a fully-proportional system that uses approval ballots (as is being discussed elsewhere in this thread), may I suggest you look at the "E Pluribus Hugo" system that's being developed for the Hugo science fiction awards?

 

There was some unfortunate history at our last convention with using approval voting that put a "bad taste" in some peoples mouth. We had a very contentious election. We had a voting confirmation group come in (I forget their name) and assist with our voting. Them being confused by AV and our rules being so confusing caused a lot of issues and really delayed the vote counting to an extreme amount. This caused people to blame AV unfortunately. I am really hoping to get around that and get people comfortable with the idea before this next convention. 

Vendors and groups can get booths there but I am not in charge of that so I don't know how it works. If any of yall are near Texas (the 2016 convention will be in San Antonio in April) I hope you can come as well!

Also thanks to everyone that works for the center for election science and the website. The video is especially helpful for giving to people new to the idea. I can try to keep information updated if anyone has any interest. Thanks!

--
You received this message because you are subscribed to the Google Groups "The Center for Election Science" group.
To unsubscribe from this group and stop receiving emails from it, send an email to electionscien...@googlegroups.com.

[Arthur Thomas IV]

Thanks for all the feedback so far!

The only multiwinner votes I can think of are for delegate spots at our conventions. For instance we have N delegate positions in a county and only that many delegates (ignoring alternates) can fill those seats. We used approval voting in our county last year and it worked well. Top vote getters got the seats. Any ties could be resolved by the chair or some random method if needed really. 

Jameson, I am intrigued by your idea because I have encountered the "but I want to pick my favorite" before. It does seem to be a kid gloves version of range voting. My only initial concern is complexity still. Not only in writing rules about it but getting people to understand. 

I plan to sit down with my fellow sub committee member and go over all this to see if we can work out a best path here. I really appreciate the input a lot and while some is immediately over my head I look forward to learning more!

--
You received this message because you are subscribed to a topic in the Google Groups "The Center for Election Science" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/electionscience/sz-_YZSS-UU/unsubscribe.
To unsubscribe from this group and all its topics, send an email to electionscien...@googlegroups.com.

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

[Jameson Quinn]

2015-08-14 13:09 GMT-04:00 Arthur Thomas IV <amtho...@gmail.com>:

Thanks for all the feedback so far!

The only multiwinner votes I can think of are for delegate spots at our conventions. For instance we have N delegate positions in a county and only that many delegates (ignoring alternates) can fill those seats. We used approval voting in our county last year and it worked well. Top vote getters got the seats. Any ties could be resolved by the chair or some random method if needed really. 

Jameson, I am intrigued by your idea because I have encountered the "but I want to pick my favorite" before. It does seem to be a kid gloves version of range voting. My only initial concern is complexity still. Not only in writing rules about it but getting people to understand. 

"You give each candidate a grade. We count all the grades at or above some threshold as approvals, and set that threshold as high as we can while still getting a majority."

"That way, if a majority agree with you on your favorite, we don't have to look any further; but you also get a chance to add some fallback options in case your favorite doesn't have a majority."

If your audience are super nerdy, you can also talk about how this system uses the median, which is a more robust statistic than the mean.

As to the relationship to score voting: MCA has been shown to be not quite as good as range if everyone honestly rates all the candidates, but it has several countervailing advantages:

1. More robust against one-sided strategy (because most people's honest ballots naturally end up being ideal from a strategic perspective).

2. Works well with only a few grading categories, so less agonizing over "should I give X an 83 or an 87?"

3. For similar reasons, easier to do a paper tally.

4. Closer to systems that have actually been used in political elections, including Progressive-era US cities and prenapoleonic revolutionary Geneva. (Warren would probably use Sparta's "loudest cheer" system and honeybee colonies as examples on the other side. Personally I see those as not really score voting, but YMMV.)

[William Waugh]

This is a response to the mention by Jameson Quinn of a median-based system.

As I have mentioned in this forum before, I tried simulating a case using a median-based system. As a result I am tentatively convinced that in regard to enabling any electoral strategy to break a two-party entrenchment, the ranking or rating opportunity that median-based systems offer does not contribute, except inasmuch as a voter can leave a candidate's name off completely or include the candidate's name. It seems best to treat median-based systems as Approval if one encounters them as a voter. I support Range Voting with granularity 101 and giving non-mentioned candidates the lowest score, for single-winner elections. A voter faced with coarser granularity should use probability. Using probability is less convenient to practice and harder to explain, and that is why to advocate for 101 granularity over mere Approval and equivalents.

On Friday, August 14, 2015 at 12:51:05 PM UTC-4, Jameson Quinn wrote
https://groups.google.com/d/msg/electionscience/sz-_YZSS-UU/9XaHd9Q-BwAJ

[Warren D Smith]

On 8/16/15, William Waugh <2knuw...@snkmail.com> wrote:
> This is a response to the mention by Jameson Quinn of a median-based
> system.
>
> As I have mentioned in this forum before, I tried simulating a case using a
> median-based system. As a result I am tentatively convinced that in regard
> to enabling any electoral strategy to break a two-party entrenchment, the
> ranking or rating opportunity that median-based systems offer does not
> contribute, except inasmuch as a voter can leave a candidate's name off
> completely or include the candidate's name. It seems best to treat
> median-based systems as Approval if one encounters them as a voter. I
> support Range Voting with granularity 101 and giving non-mentioned
> candidates the lowest score, for single-winner elections. A voter faced
> with coarser granularity should use probability. Using probability is less
> convenient to practice and harder to explain, and that is why to advocate
> for 101 granularity over mere Approval and equivalents.

--WDS:
That by W.Waugh all sounds interesting but I do not understand it.
Perhaps he can explain in more detail.

Why should median or mean-based range voting be better about
breaking 2-party domination? It is not immediately obvious why,
and if one just is better in that respect from computer sims without a reason
(which it could be, I suppose) then I'd like to know more about what the
computer did and how this was assessed.

And I also do not know what Waugh meant here about "probability."

If he had earlier clearer posts, perhaps provide a hyperlink to them?

I don't know how this meshes with Waugh, but...
I actually feel there IS reason to believe that average-based-range voting is
superior to approval voting for the purpose of breaking 2-party domination.
I have 2 reasons:
A. experimental fact, based on poll evidence in approval-style and
score-style polling
including as exit polls... third parties always seem to do better with score
than with approval, relative to the 2 major parties. I've seen that
in enough polls to
be pretty convinced of it by now. These differences can be enormous.
B. hypothesized explanation, I call this the "nursery effect"...
introduction here:

http://www.rangevoting.org/NurserySumm.html

[William Waugh]

<< And I also do not know what Waugh meant here about "probability." >>

I meant that a voter simulates finer grained Score voting in a coarser grained Score election by voting probabilistically. For example suppose the election is among Worst, Compromise, and Favorite, where based on opinion polls and the like, I expect Favorite would receive about 1% in a plurality election, but suppose we have Approval. Then I approve Favorite, and I extract a random number from a computer or some dice throws or something like that and approve Compomise with 99% probability.

On Monday, August 17, 2015 at 12:00:56 PM UTC-4, Warren D. Smith (CRV cofounder, http://RangeVoting.org) wrote https://groups.google.com/d/msg/electionscience/sz-_YZSS-UU/Su5JmtUnCAAJ

[Warren D Smith]

On 8/24/15, William Waugh <2knuw...@snkmail.com> wrote:
> << And I also do not know what Waugh meant here about "probability." >>
>
> I meant that a voter simulates finer grained Score voting in a coarser
> grained Score election by voting probabilistically. For example suppose the
>
> election is among Worst, Compromise, and Favorite, where based on opinion
> polls and the like, I expect Favorite would receive about 1% in a plurality
>
> election, but suppose we have Approval. Then I approve Favorite, and I
> extract a random number from a computer or some dice throws or something
> like that and approve Compomise with 99% probability.

--aha. Thus effectively simulating F=100, C=99, W=0 using score voting, but
using approval voting to do it?

Many of the median-based schemes have tiebreaking mechanisms that
actually would yield the same winner as average-based score voting,
if all voters behaved this way with approval. However, if we had, say,
score with {0,1,2,3,4} as the only permitted scores, and the voter behavior
were F=4, C={96% 4 and 4% 3 at random}, W=0
then, well... I believe there exist examples of this general ilk
where median-based schemes will disagree
with average-based schemes about the winner.

[Arthur Thomas IV]

ok, its been a few months but we have some language written up to replace our existing language. Sadly there are lots of potential rule changes so these likely will not make it onto the floor at the convention, but at least they will be in the report. I am posting them here for review and comment.

One of the things that affects our voting in the LP is the inclusion of NOTA. We must be able to say "NO" to a candidate even to the point of leaving the seat empty.

This would be under our section for single seat offices (public/party):

1. Approval Voting Procedure. Each delegate may vote for as many of the candidates as he or she approves. All votes for the candidates and “None of the Above” (NOTA) shall be tallied, and whichever receives the most votes will win.
2. NOTA procedure for public office. NOTA shall be an option in all elections for public office. All candidates for a public office shall be placed on a single ballot along with NOTA. If NOTA receives more votes than the candidate with the most votes, then no candidate will be nominated for that position. A Convention may decline to nominate any candidate for public office even if there are persons seeking the Party’s nomination for that office.
3. NOTA procedure for party office. NOTA shall be an option in all elections for party office. All candidates for a party office shall be placed on a single ballot along with NOTA. If NOTA receives more votes than the candidate with the most votes, then either nominations may be reopened by majority vote or that office shall be declared vacant and shall be filled as provided elsewhere in these Rules.
4. Order of nominations. In voting on the nomination of candidates for public office, the offices shall be considered in the reverse of the order in which they appear on the general election ballot.
5. Opt out. The delegates of the convention may decide by majority vote to use whatever voting method they prefer to elect candidates for public and party offices as long as NOTA is included.

This is for multi seat races:

1. Each ballot shall contain the full list of all nominees. Each nominee shall have the following options adjacent to their name on the ballot: Yes, No, Abstain.
2. Each voter may vote in the manner he or she chooses for each nominee.
3. All Nominees that obtain more yes votes than no votes are Eligible Nominees. Eligible Nominees will be ranked in order of highest yes count to lowest yes count. In the case of a tie in yes votes, the higher preference shall be the Eligible Nominee with less no votes. If a tie occurs when Eligible Nominees have the same yes and no votes, the order shall be determined at random or by whatever means the body prefers.
4. A ballot selection for abstain, multiple selections, or no selection shall be considered an abstention for that Nominee.
5. Nominees will be seated in ranked order until all seats are filled or until there are no longer eligible nominees.

Given out completely messy our current rules are we wanted to keep it short and sweet.

[Andy Jennings]

Arthur,

Thank you for your work on this.  It's great to see concrete proposals like this come forward.  Good luck getting it implemented, if not this year then at some point in the future.

Let us know if there are specific supporting materials that would be helpful.

~ Andy

[William Waugh]

On Wednesday, August 12, 2015 at 5:06:48 PM UTC-4, Arthur Thomas IV wrote https://groups.google.com/d/msg/electionscience/sz-_YZSS-UU/sZRJoHIRAQAJ

The single-winner section looks like a great plan.

A couple of notes in regard to the multi-winner section:

1. It does not attempt to provide proportional representation (PR), but has rather a winner-take-all character. I can see where you might prefer to keep it simple instead of going for PR. As you describe it, it is still a miles better system than any that limits how many candidacies a voter can support, or any that requires ranking.

2. If write-ins are allowed, there could be a dark-horse problem, given that you allow abstention with respect to a given candidacy.

[Arthur Thomas IV]

Thanks for the input.

Technically people can be nominated from the floor in almost all the cases (except public candidates), but there are no blank write-in spots.

This isn't really proportional representation because its really a bunch of single individuals competing for multiple open seats. There are no seats to be divided proportionally between groups. Basically we nominate a bunch of people from the floor and people vote on who should get the available seats.  Does that make sense?

--
You received this message because you are subscribed to a topic in the Google Groups "The Center for Election Science" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/electionscience/sz-_YZSS-UU/unsubscribe.
To unsubscribe from this group and all its topics, send an email to electionscien...@googlegroups.com.

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

[William Waugh]

On Monday, February 22, 2016 at 10:43:44 PM UTC-5, Arthur Thomas IV wrote:

...

This isn't really proportional representation because its really a bunch of single individuals competing for multiple open seats. There are no seats to be divided proportionally between groups. Basically we nominate a bunch of people from the floor and people vote on who should get the available seats.  Does that make sense?

 

No. Your group can still do PR if it chooses to. It would suffice to follow the procedure Clay illustrates in his video, above. 

[Arthur Thomas IV]

Don't you need factions for that to work though? 

I do not fully understand the example yet. 

--

You received this message because you are subscribed to a topic in the Google Groups "The Center for Election Science" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/electionscience/sz-_YZSS-UU/unsubscribe.
To unsubscribe from this group and all its topics, send an email to electionscien...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Clay Shentrup]

On Monday, February 22, 2016 at 11:55:39 PM UTC-8, Arthur Thomas IV wrote:

Don't you need factions for that to work though?

Factions aka "diversity of opinion". Which you definitely have.

[William Waugh]

It is not necessary for anyone in the group to recognize, identify, discern, name, or have conscious awareness of the factions. Factions may exist in the different valuations the voters have regarding the candidacies. If such factions exist, the voting system may be able to find them by itself. There is no need for the humans to feed any party or faction or tendency identifiers to the voting system -- only the votes themselves.

Tallying proceeds in rounds, where each round finds one winner. Finding the first winner is straightforward -- it is the candidate who received the most count of approvals. For the subsequent rounds, you weight the votes based on the winners already found and who approved them. Those votes get a lower weight than the full weight they originally had. The grounds for this is that those voters already have some representation in the set of already-found winners. So the second winner is who gets the highest score based on the weighted votes. Approvals of candidates already determined as winners don't count in this comparison. After the second round, the tally determines a new set of weights for the votes based on the two winners already determined. Conduct additional rounds like that until sufficient seats are filled.

[Arthur Thomas IV]

Thanks, I understand the goal of using that a bit better now. Reducing the "influence" of those already represented in the decision making process. Its very interesting. Definitely a bit to complex deal with in the rules right now.

Some people are put off by basic approval voting because of bad experience. Some people running the elections handled it oddly and ate up quite a bit of time which people blamed on AV :( 

In the general public, I have been surprised at people knee jerk reaction to it as well. I can understand being a bit suspicious of someone wanting to change a voting method but the outright rejection because of that is odd. 

Anyway, I will keep working on it here in Texas :) Thanks for the help!

On Tue, Feb 23, 2016 at 10:55 AM, William Waugh <2knuw...@snkmail.com> wrote:

It is not necessary for anyone in the group to recognize, identify, discern, name, or have conscious awareness of the factions. Factions may exist in the different valuations the voters have regarding the candidacies. If such factions exist, the voting system may be able to find them by itself. There is no need for the humans to feed any party or faction or tendency identifiers to the voting system -- only the votes themselves.

Tallying proceeds in rounds, where each round finds one winner. Finding the first winner is straightforward -- it is the candidate who received the most count of approvals. For the subsequent rounds, you weight the votes based on the winners already found and who approved them. Those votes get a lower weight than the full weight they originally had. The grounds for this is that those voters already have some representation in the set of already-found winners. So the second winner is who gets the highest score based on the weighted votes. Approvals of candidates already determined as winners don't count in this comparison. After the second round, the tally determines a new set of weights for the votes based on the two winners already determined. Conduct additional rounds like that until sufficient seats are filled.

--
You received this message because you are subscribed to a topic in the Google Groups "The Center for Election Science" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/electionscience/sz-_YZSS-UU/unsubscribe.
To unsubscribe from this group and all its topics, send an email to electionscien...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.