These forums have been archived and are now read-only.

The new forums are live and can be found at https://forums.eveonline.com/

Jita Park Speakers Corner

 
  • Topic is locked indefinitely.
 

The voting reform discussion

First post First post
Author
Previous Topic Next Topic
CCP Xhagen
C C P
C C P Alliance
#421 - 2012-09-24 15:09:32 UTC
Thanks again for continuing this.

And Inquisitor Kitchner, you bring up very good points.

Under the Schulze method E and A are elected, under FPTP it is C and A and with the STV it is C and A (like you elegantly point out).

But, after running through the numbers for the Schulze method, this is the results.

5 ACBED (meaning, 5 voters have order of preference: A > C > B > E > D)
5 ADECB
8 BEDAC
3 CABED
7 CAEBD
2 CBADE
7 DCEBA
8 EBADC

1st preference
a: 10 (22%)
b: 8 (18%)
c: 12 (27%)
d: 7 (16%)
e: 8 (18%)

2nd preference
a: 10 (22%)
b: 10 (22%)
c: 12 (27%)
d: 5 (11%)
e: 8 (18%)

3rd preference
a: 10 (22%)
b: 8 (18%)
c: 0 (0%)
d: 8 (18%)
e: 19 (42%)

4th preference
a: 8 (18%)
b: 14 (31%)
c: 5 (11%)
d: 10 (22%)
e: 8 (18%)

5th preference
a: 7 (16%)
b: 5 (11%)
c: 16 (36%)
d: 15 (33%)
e: 2 (4%)

__
(candidate X: the % of voters that choose him for)

a: 22% in the top 1, 78% in bottom four
b: 18% in the top 1, 82% in bottom four
c: 27% in the top 1, 73% in bottom four
d: 16% in the top 1, 84% in bottom four
e: 18% in the top 1, 82% in bottom four

a: 44% in the top 2, 56% in bottom three
b: 40% in the top 2, 60% in bottom three
c: 54% in the top 2, 46% in bottom three
d: 27% in the top 2, 73% in bottom three
e: 36% in the top 2, 64% in bottom three

a: 66% in the top 3, 34% in bottom two
b: 58% in the top 3, 42% in bottom two
c: 54% in the top 3, 46% in bottom two
d: 54% in the top 3, 46% in bottom two
e: 78% in the top 3, 22% in bottom two

a: 84% in the top 4, 16% in bottom one
b: 89% in the top 4, 11% in bottom one
c: 65% in the top 4, 36% in bottom one
d: 67% in the top 4, 33% in bottom one
e: 96% in the top 4, 4% in bottom one

This shows that in the election example above the more you consider how people order on the list, the stronger candidates E and A stand out. A is selected due to fewer people wanting that candidate in the last two places than B.

The thing about election systems is that we have a bunch of information in the votes - changing the election system to get as correct of a result as possible from the current voters is (and pardon my french) a no-brainer.

CCP Xhagen | Associate Producer | @strangelocation
Dersen Lowery
The Scope
#422 - 2012-09-24 15:56:58 UTC
Inquisitor Kitchner wrote:
What you'll end up with is having the top elected person simply being the guy everyone hates the least. Whereas A & C are actually most people's top two preferences E struggles to get above third.


As an aside, this is exactly how singles are chosen for radio play. Songs with extreme ratings with focus groups, whether 1 star or 5 stars out of 5, are set aside, and songs that consistently average 3s are released.

It also tends to be how any non-athletic contest sorts out, with the candidate who is least inoffensive to the greatest number of people winning.

I'm neither endorsing nor defending the idea, just pointing out that outcomes like those rendered of the Shulz method are pervasive and familiar.

Inquisitor Kitchner wrote:
Do you really want to use a system where nearly everyone's lower choice gets elected?


I don't know that it's necessarily worse than a system where the FOTM gets elected.

Inquisitor Kitchner wrote:
Should you be worried if you change the voting system to the extent where one large group can vote in all 14 candidates of their choosing rather then just one? Yes. However much I like my own alliance if you got 14 of us to run the CSM on our own then you'd never get a conflicting opinion (well, not fundamentally conflicting) and that just means certain issues wouldnt be addressed.


This. The one system in the US that most closely resembles so-called "direct democracy" is the California ballot initiative, and that has a long and sordid history of being gamed and manipulated into ramming through propositions that would wither under the scrutiny of any decent legislature.

We have a Senate here for a reason, as hard as that reason may be to see sometimes.

Proud founder and member of the Belligerent Desirables.

I voted in CSM X!
Lunaleil Fournier
Viziam
Amarr Empire
#423 - 2012-09-24 22:16:45 UTC
Dersen Lowery wrote:

Inquisitor Kitchner wrote:
Do you really want to use a system where nearly everyone's lower choice gets elected?


I don't know that it's necessarily worse than a system where the FOTM gets elected.



Of course it's worse. Candidates with lower choices get in over people with higher choices? That's not a voting system that is representative of the people's wishes.

If that's the way things are calculated, I'd only select one candidate and submit the ballot, if I participate at all.
L'Acuto
Garoun Investment Bank
Gallente Federation
#424 - 2012-09-24 22:27:19 UTC  |  Edited by: L'Acuto
STV (Sexually Transmitted Vote) - who wants to **** and who wants to get ****ed?
Dersen Lowery
The Scope
#425 - 2012-09-24 23:56:17 UTC  |  Edited by: Dersen Lowery
Lunaleil Fournier wrote:
Of course it's worse. Candidates with lower choices get in over people with higher choices? That's not a voting system that is representative of the people's wishes.


It's the system that represents the most wishes of the most people. It has the side effect of placing at a disadvantage people who are intensely liked by one group and intensely disliked by the other, which is a vulnerability that can be gamed, but the other system can vault someone that one group loves and the rest of the voters intensely dislike.

So, again, I don't see how it's worse. It's just a question of to what degree you want consensus candidates, and to what degree you want bloc candidates.

nb: I don't see anything particularly wrong with bloc candidates.

Proud founder and member of the Belligerent Desirables.

I voted in CSM X!
Pipa Porto
#426 - 2012-09-25 00:03:25 UTC
Dersen Lowery wrote:
Lunaleil Fournier wrote:
Of course it's worse. Candidates with lower choices get in over people with higher choices? That's not a voting system that is representative of the people's wishes.


It's the system that represents the most wishes of the most people. It has the side effect of placing at a disadvantage people who are intensely liked by one group and intensely disliked by the other, which is a vulnerability that can be gamed, but the other system can vault someone that one group loves and the rest of the voters intensely dislike.

So, again, I don't see how it's worse. It's just a question of to what degree you want consensus candidates, and to what degree you want bloc candidates.

nb: I don't see anything particularly wrong with bloc candidates.


So, you're advancing a system that rewards anonymity?

Say there are 3 candidates and 2 groups.
Group 1 likes Candidate A and hates B
Group 2 likes Candidate B and hates A
Nobody knows Candidate C

Group 1 votes ACB
Group 2 votes BCA
C votes for himself, CBA or CAB

Nobody knows who C is, nobody voted for C, but C wins.

EvE: Everyone vs Everyone

-RubyPorto
Scooter McCabe
Thunderwaffe
Goonswarm Federation
#427 - 2012-09-25 01:18:50 UTC
CCP Xhagen you are a far braver man than I to actually walk through the mathematics of the Schulze method. But I hoped as you walked through that hypothetical you did see the potential for dissatisfaction and that the two methods there is the potential for a candidate that would be traditionally elected under our current system to find themselves mathematically disqualified. I applaud you for showing an example of what the outcome of a potentially tight race might look like. While tight races are exciting, in the end we know someone is not walking away happy.

Its easier to example and accept that a candidate came up a few votes short. The discontent is quiet and becomes quiet acceptance because of the transparency and fairness in the democratic system. Under the method outlined you are going to run into election results where a candidate would have out performed the other, and that's going to lead to some very vocal protestations. I don't know if these systems carry with them the kind of transparency and sense of fairness that act as a built in buffer to the discontent caused by an election loss. That's not CCP's fault that these proposed systems would be met with skepticism, many of us live under a democratic system in the real world, its what we know.

As a mathematical concern the Schulze method being Condorcet method of elections, has two main problems. First if someone does not meet the Condorcet Criterion to elect a clear winner the votes can be tallied multiple ways to elect different candidates each time. It also can hilariously lead to a rock/paper/scissors scenario where each beats the other and no clean winner exists requiring some method of tie breaking, but in that scenario does CCP want to be the one to break the tie. I know CCP is trying to make everyone happy and your looking for a silver bullet, but in election systems this isn't it.

So if we did go with some form of the Condorcet method what do you do to address the 800 pound gorilla in the room? Also the even bigger question remains how do we get more people involved in the elections?

One thing I am certain we agree on is that increasing the size of the CSM, having separate CSMs, or CSM chairs straight-jacketed to niche advocacy would produce a disjointed experience for players throughout EVE, not to mention the hellish bureaucratic nightmare it would make for everyone involved.
Pipa Porto
#428 - 2012-09-25 01:40:47 UTC  |  Edited by: Pipa Porto
CCP Xhagen wrote:
The thing about election systems is that we have a bunch of information in the votes - changing the election system to get as correct of a result as possible from the current voters is (and pardon my french) a no-brainer.


Then you run into the significant risk of accusations of tampering with the election (because choosing the voting system after the vote is held is the definition of election tampering) to support a specific candidate or agenda. That's a giant steaming pile that CCP would do well to avoid diving headfirst into.

Here's how the accusations would run:
A would win in a FPTP system.
B would win in an STV system.
C would win in whatever other system.

If you end up picking FPTP, B and C start yelling.
Pick STV, A and C start yelling, etc.


Changing the voting system well before the election is fine, though it won't have nearly the impact on representing the playerbase's wishes that improving turnout would. Changing the voting system after the election is dirty pool.

Forcing people to vote doesn't work, as that, at best*, inserts noise into the results from people trying to click through the process. Same with direct encouragement like rewards for voting.

The CSM candidates need to be doing things to improve turnout. Things like effective campaigning, effective polling, and organizational efforts to increase turnouts. Goonswarm wins because they have high turnout among their members and present capable candidates.


By the way, discussing changes to the voting system when some sixth of the electorate (the sixth who won the election, to boot) are without representation is fairly ridiculous and reeks of the current CSM simply wanting to remain in power. Their initial proposal being designed expressly to do that makes the stench worse. You're going to have to wait until CSM 8 to have a productive discussion of this because this CSM has already painted themselves as craven careerists (Trebor and the CSMs who defended the proposal) or cowards/absentees (those who didn't immediately denounce it), or both (Hans, who managed to do both, defending it and then running away very quickly as soon as the tide turned).


*at worst, you get a significant bias towards name recognition.

EvE: Everyone vs Everyone

-RubyPorto
Dersen Lowery
The Scope
#429 - 2012-09-25 02:50:47 UTC
Pipa Porto wrote:
[So, you're advancing a system that rewards anonymity?

Say there are 3 candidates and 2 groups.
Group 1 likes Candidate A and hates B
Group 2 likes Candidate B and hates A
Nobody knows Candidate C

Group 1 votes ACB
Group 2 votes BCA
C votes for himself, CBA or CAB

Nobody knows who C is, nobody voted for C, but C wins.


I'm presuming that if nobody knows them, they won't even get votes. If C is everybody's second choice, then C wins. I'm agnostic about which is "better." I just don't see how it's obvious that this result is worse than A or B getting elected, when a significant portion of the player base hates them.

I honestly don't care that much either way. CSMs 6 and 7 have good people on them, so nothing about the current system is especially broken. I just don't get the hate for the other system.

Proud founder and member of the Belligerent Desirables.

I voted in CSM X!
Pipa Porto
#430 - 2012-09-25 03:07:35 UTC
Dersen Lowery wrote:
Pipa Porto wrote:
[So, you're advancing a system that rewards anonymity?

Say there are 3 candidates and 2 groups.
Group 1 likes Candidate A and hates B
Group 2 likes Candidate B and hates A
Nobody knows Candidate C

Group 1 votes ACB
Group 2 votes BCA
C votes for himself, CBA or CAB

Nobody knows who C is, nobody voted for C, but C wins.


I'm presuming that if nobody knows them, they won't even get votes. If C is everybody's second choice, then C wins. I'm agnostic about which is "better." I just don't see how it's obvious that this result is worse than A or B getting elected, when a significant portion of the player base hates them.

I honestly don't care that much either way. CSMs 6 and 7 have good people on them, so nothing about the current system is especially broken. I just don't get the hate for the other system.


Actually take a look at my example. C got 1 positive vote and won the election because A and B were polarizing figures while C voted for himself and shut the hell up. C is the second choice because the people who hate A or B will put the candidate they hate at the bottom and fill in the middle with whatever's left.

The bigest problem I had with CSM 7 until recently was that they were largely ineffective or absent. Now, as demonstrated in the other thread, my problem is that they're cravenly careerist, attempting to rig the vote to keep themselves in power and disenfranchise their opposition.

Bad politicians isn't something that changing the voting system can fix. It's just something that happens when you vote on incomplete information, which is all the time, so...

EvE: Everyone vs Everyone

-RubyPorto
Yeep
The Scope
Gallente Federation
#431 - 2012-09-25 10:20:31 UTC
Dersen Lowery wrote:

I'm presuming that if nobody knows them, they won't even get votes. If C is everybody's second choice, then C wins. I'm agnostic about which is "better." I just don't see how it's obvious that this result is worse than A or B getting elected, when a significant portion of the player base hates them.


Because Eve is a game and in order to maintain balance for one party to gain another has to lose. I'd rather have CSM members who are honest about that than shameless panderers who appeal to everyone but will have to break half their promises post-election.
CCP Xhagen
C C P
C C P Alliance
#432 - 2012-09-25 14:53:49 UTC
Scooter McCabe wrote:
CCP Xhagen you are a far braver man than I to actually walk through the mathematics of the Schulze method. But I hoped as you walked through that hypothetical you did see the potential for dissatisfaction and that the two methods there is the potential for a candidate that would be traditionally elected under our current system to find themselves mathematically disqualified. I applaud you for showing an example of what the outcome of a potentially tight race might look like. While tight races are exciting, in the end we know someone is not walking away happy.

Its easier to example and accept that a candidate came up a few votes short. The discontent is quiet and becomes quiet acceptance because of the transparency and fairness in the democratic system. Under the method outlined you are going to run into election results where a candidate would have out performed the other, and that's going to lead to some very vocal protestations. I don't know if these systems carry with them the kind of transparency and sense of fairness that act as a built in buffer to the discontent caused by an election loss. That's not CCP's fault that these proposed systems would be met with skepticism, many of us live under a democratic system in the real world, its what we know.

As a mathematical concern the Schulze method being Condorcet method of elections, has two main problems. First if someone does not meet the Condorcet Criterion to elect a clear winner the votes can be tallied multiple ways to elect different candidates each time. It also can hilariously lead to a rock/paper/scissors scenario where each beats the other and no clean winner exists requiring some method of tie breaking, but in that scenario does CCP want to be the one to break the tie. I know CCP is trying to make everyone happy and your looking for a silver bullet, but in election systems this isn't it.

So if we did go with some form of the Condorcet method what do you do to address the 800 pound gorilla in the room? Also the even bigger question remains how do we get more people involved in the elections?

One thing I am certain we agree on is that increasing the size of the CSM, having separate CSMs, or CSM chairs straight-jacketed to niche advocacy would produce a disjointed experience for players throughout EVE, not to mention the hellish bureaucratic nightmare it would make for everyone involved.


I fully understand your concerns and that is why I'm bringing this up well in advance of the next election; to have people poking the ideas to see if, and what, flaws come up.

In regards to the transparency - we would release both the counting code before the election and the voting data (i.e. how votes were cast and for what members). We release the voting data now anyway, i.e. by providing the list of candidates and how many votes they got. So we would strife for as much transparency as we can actually muster. On that note I don't foresee releasing daily data about votes during the election though, only in one chunk after it.

Regarding tallying the votes and tiebreakers - we will simply commit to clear rules and methods before the election so that if a situation comes up everyone will know how it will be handled.

So, by being as honest, transparent and clear as possible I would think that the 800 pound gorilla in the room would simply be afraid of me.

CCP Xhagen | Associate Producer | @strangelocation
CCP Xhagen
C C P
C C P Alliance
#433 - 2012-09-25 14:57:14 UTC
And please don't read my previous reply as 'WE ARE GOING TO IMPLEMENT A NEW ELECTION SYSTEM' - I'm a reformist, not a revolutionist and while the current election system produces 'A RESULT', most people agree that it is sufficient but not the best system there is. So we _can_ still use it, but I'd like to upgrade it (despite the objections of 'if it ain't broken don't fix it' - no one (including me) has showed that it is broken and I fully acknowledge that).

CCP Xhagen | Associate Producer | @strangelocation
Snow Axe
Republic Military School
Minmatar Republic
#434 - 2012-09-25 15:38:24 UTC
CCP Xhagen wrote:
And please don't read my previous reply as 'WE ARE GOING TO IMPLEMENT A NEW ELECTION SYSTEM' - I'm a reformist, not a revolutionist and while the current election system produces 'A RESULT', most people agree that it is sufficient but not the best system there is.


[citation needed]

"Look any reason why you need to talk like that? I have now reported you. I dont need to listen to your bad tone. If you cant have a grown up conversation then leave the thread["
Lord Zim
Gallente Federation
#435 - 2012-09-25 15:44:00 UTC
CCP Xhagen wrote:
and while the current election system produces 'A RESULT', most people agree that it is sufficient but not the best system there is.

What's the problem with the current election system?

Cyno's lit, bridge is up, but one pilot won't be jumping home.

RIP Vile Rat
Jackie Fisher
Syrkos Technologies
#436 - 2012-09-25 15:50:27 UTC
For some reason this thread makes me think of great solutions in engineering.

Fear God and Thread Nought
CCP Xhagen
C C P
C C P Alliance
#437 - 2012-09-25 15:54:05 UTC
Snow Axe wrote:
CCP Xhagen wrote:
And please don't read my previous reply as 'WE ARE GOING TO IMPLEMENT A NEW ELECTION SYSTEM' - I'm a reformist, not a revolutionist and while the current election system produces 'A RESULT', most people agree that it is sufficient but not the best system there is.


[citation needed]

Fair point.

It is true that it is considered sufficient by several people, at least I'm inferring that from the number of posts saying 'don't change the voting system'.

CCP Xhagen | Associate Producer | @strangelocation
Snow Axe
Republic Military School
Minmatar Republic
#438 - 2012-09-25 16:02:32 UTC  |  Edited by: Snow Axe
Forgive me if this comes off confrontational, but shouldn't the first step have been trying to get a real idea of what people ACTUALLY think of the current voting system before even considering a discussion about changing it?

Or better yet, not even think about changing it until you've got your voting numbers where you want them to be (or at least to where you think it's as good as it's going to get). You don't decorate your house until you're finished building it.

"Look any reason why you need to talk like that? I have now reported you. I dont need to listen to your bad tone. If you cant have a grown up conversation then leave the thread["
CCP Xhagen
C C P
C C P Alliance
#439 - 2012-09-25 16:07:00 UTC
Lord Zim wrote:
CCP Xhagen wrote:
and while the current election system produces 'A RESULT', most people agree that it is sufficient but not the best system there is.

What's the problem with the current election system?


1. People only vote for one candidate (per account) for a council that has fourteen seats.
2. A lot of votes are truncated because they go to candidates that don't make it in. Don't misunderstand me, I realize that is how an election works. Pointing towards point one, allowing people to express their preferences by selecting more than one candidate would create a better representation of the voters' preferences on the council.
3. Strategic voting is extremely easy. Again, don't misunderstand. There is nothing wrong with strategic voting, but making it as easy as possible means that instead of manipulating people to vote for you, you manipulate the system by utilizing the votes you already have in the manner to achieve the results you want.
4. A side effect of a 'mathematically stronger' system could be an increased belief in the election - I know it is a bit far fetched and doesn't really belong on this list. Cry

CCP Xhagen | Associate Producer | @strangelocation
CCP Xhagen
C C P
C C P Alliance
#440 - 2012-09-25 16:09:07 UTC
Snow Axe wrote:
Forgive me if this comes off confrontational, but shouldn't the first step have been trying to get a real idea of what people ACTUALLY think of the current voting system before even considering a discussion about changing it?

Or better yet, not even think about changing it until you've got your voting numbers where you want them to be (or at least to where you think it's as good as it's going to get). You don't decorate your house until you're finished building it.

But I CAN talk about decorating my house before it is built. And talking about changing the election system also brings out peoples' opinion about the current election system.

CCP Xhagen | Associate Producer | @strangelocation