Election Results - January 2005

It looks like we have enough votes in to call the election. It should be noted that the Clown Party (CP) is not eligible to hold seats in the RA, as they didn't meet the minium membership requirement. In the voting analysis post following this one, I'll discuss why the CP was included (but not counted) along with a discussion of our current voting system.

The SDF has earned 4 seats and the MPP has earned 3 seats. I'll be contacting members individually according to the faction rankings to see if they want to serve (some may want to keep current government posts). In the future I'll include that question in the voting process so we can release names with the seat results.

Below is a summary of the results for the vote. As you know, factions were rated from 1 to 3 with more "points" going to those who were higher rated. The equation which calculates the score gives the highest ranked faction a "1" and the lowest ranked faction a "0". Factions in between receive fractional amounts. The equation is

score = (number_of_factions - rank) / (number_of_factions - 1)

where a rank of "1" gives the highest score. Because we had only two eligible factions, the Borda method of ranking had no effect.

The Sainte-Laguë method was used to distribute RA seats. The table used to calculate the seats is shown below as well. Please see the Wikipedia entry for details on how the method is applied.

Faction Vote (Borda Count)     Score    Percent
  Social Democratic Faction       14      60.8%
  Moderate Peoples Party           9      39.1%

Faction Seats (Sainte-Lague Method)
  Votes         SDF     MPP     Seat
 --------------------------------------------
  Seat 1        14.000  9.000   SDF
  Seat 2        4.667   9.000   MPP
  Seat 3        4.667   3.000   SDF
  Seat 4        0.933   3.000   MPP
  Seat 5        0.933   0.600   SDF
  Seat 6        0.133   0.600   MPP
  Seat 7        0.133   0.086   SDF
 --------------------------------------------
  Seats         4       3 
  Percent       57.1%   42.8% 

Thanks to everyone who voted!

Up next ... the analysis.

~Ulrika~

Below is a quick analysis of the vote. I think it's very important for everyone to read and understand it, as I'd like to make some changes to the system, inform people about the benefits and detriments of tactical voting, and discuss ethical issues. If there is anything that is not understood, be sure to ask questions.

As a disclaimer, I do not know who casted what vote. The votes are completely anonymous. I do know who voted because their name shows up in the password list after they receive a PIN from the voting booth but that is not associated with the results in any way.

Tactical voting
I added the Costume Party (CP) to the ballot, even though they were ineligible, to see how pervasive tactical voting would be. Tactical voting occurs when voters misrepresent their sincere preferences in order to gain a more favorable outcome. An example is members of the MPP ranking the CP over the SDF to gain an advantage.

The effect of this tactical voting can alter the vote dramatically, especially with small groups such as ours. Unfortunately, sometimes in small groups tactical voting can have the opposite effect intended. I'll follow up with more details in my next post entitled "Tactical Voting".

Also, there was even evidence of tactical voting within factions themselves when voters were asked to rank members. A single person burying a competitor in a long list had a dramatic effect on the ranking of the person.

Tactical voting is a problem in all deterministic voting systems (nonrandom), however it is especially acute in ranked systems such as the Borda count. Thus I'd like to change the method of ranking to mitigate the effects of tactical voting. It will require some discussion but I'm leaning towards Cloneproof Schwartz Sequential Dropping (CSSD). After we settle into our next lease, we'll revisit this topic.

Destructive Voting
I was quite surprised but two people actually ranked the CP first. It was something I had not considered! With our current system, if the CP were not disqualified, it would have gained two seats on the RA matching the MPP! This is due to the way that the Sainte-Laguë method amplifies the strength of minority parties.

Because of the potential of destructive voting and our small group size, I recommend that we change from the Sainte-Laguë method to either the d'Hont method or a modified version of the Sainte-Laguë method that uses a different series of divisors. We'll revisit this one later as well.

Voting Fraud
Finally, I have evidence that at least one person engaged in alt voting, which is not allowed. I don't know who it is nor do I think it really matters, as the elimination of a vote from either side makes no difference. I cannot discourage alt voting enough for reasons I don't need to go into here.

Next up ... tactical voting.

~Ulrika~

In this post, I'll show what would have happened if the Clown Party (CP) had not been disqualified due to membership limits (three parties didn't make the limit) and take a closer look at the effects of tactical voting.

Based on the votes we received this is how the election would have turned out using a Borda count and the Sainte-Laguë method.

Faction Vote (Borda Count)     Score    Percent
  Social Democratic Faction       16      46.3%
  Moderate Peoples Party        14.5      42.0%
  Costume Party                    4      11.5%

Faction Seats (Sainte-Lague Method)
  Votes         SDF     MPP     CP      Seat
 --------------------------------------------
  Seat 1        16.000  14.500  4.000   SDF
  Seat 2        5.333   14.500  4.000   MPP
  Seat 3        5.333   4.833   4.000   SDF
  Seat 4        1.067   4.833   4.000   MPP
  Seat 5        1.067   0.967   4.000   CP
  Seat 6        1.067   0.967   1.333   CP
  Seat 7        1.067   0.967   0.267   SDF
 --------------------------------------------
  Seats         3       2       2 
  Percent       42.8%   28.5%   28.5% 

As you can see, even though the CP received only 11.5% of the ranked vote, they took 28.5% of the seats.


If you eliminate the Borda count and go with a single-vote count with the Sainte-Laguë method, this is the result:

Faction Vote (Borda Count)     Score    Percent
  Social Democratic Faction       12      52.1%
  Moderate Peoples Party           9      39.1%
  Costume Party                    2      8.69%

Faction Seats (Sainte-Lague Method)
  Votes         SDF     MPP     CP      Seat
 --------------------------------------------
  Seat 1        12.000  9.000   2.000   SDF
  Seat 2        4.000   9.000   2.000   MPP
  Seat 3        4.000   3.000   2.000   SDF
  Seat 4        0.800   3.000   2.000   MPP
  Seat 5        0.800   0.600   2.000   CP
  Seat 6        0.800   0.600   0.667   SDF
  Seat 7        0.114   0.600   0.667   CP
 --------------------------------------------
  Seats         3       2       2 
  Percent       42.8%   28.5%   28.5% 

Here the group with only 8.7% of the single-vote count captured 28.5% of the seats. This is why I'm questioning the Sainte-Laguë method. It would likely work better in larger groups but for our small group it seems to promote small groups at the expense of the medium-sized groups with unexpected efficacy!


A good compromise would be using a modified Sainte-Laguë method with a divisor that increments by 1 instead of by 2. The results of a modified Sainte-Laguë method are shown below:

Faction Vote (Borda Count)     Score    Percent
  Social Democratic Faction       16      46.3%
  Moderate Peoples Party        14.5      42.0%
  Costume Party                    4      11.5%

Faction Seats (Sainte-Lague Method)
  Votes         SDF     MPP     CP      Seat
 --------------------------------------------
  Seat 1        16.000  14.500  4.000   SDF
  Seat 2        8.000   14.500  4.000   MPP
  Seat 3        8.000   7.250   4.000   SDF
  Seat 4        2.667   7.250   4.000   MPP
  Seat 5        2.667   2.417   4.000   CP
  Seat 6        2.667   2.417   2.000   SDF
  Seat 7        0.667   2.417   2.000   MPP
 --------------------------------------------
  Seats         3       3       1 
  Percent       42.8%   42.8%   14.2% 

This is more representative.


Finally, here is a demonstration that will be of great interest to everyone. The combination of the MPP tactical votes, which rated the CP second and the SDF last (members of the SDF didn't do this for some reason) in unison with the destructive votes that rated the CP first dropped the MPP to the same number of seats as the CP. Had the CP been in the election, the tactical voting would have had the paradoxical effect of costing the MPP a seat in the RA, making them equal with the faction they ranked second! Let that be a lesson to you tactical voters.

Faction Vote (Borda Count)     Score    Percent
  Social Democratic Faction       12      52.1%
  Moderate Peoples Party           9      39.1%
  Costume Party                    2      8.69%

Faction Seats (Sainte-Lague Method)
  Votes         SDF     MPP     CP      Seat
 --------------------------------------------
  Seat 1        12.000  9.000   2.000   SDF
  Seat 2        4.000   9.000   2.000   MPP
  Seat 3        4.000   3.000   2.000   SDF
  Seat 4        0.800   3.000   2.000   MPP
  Seat 5        0.800   0.600   2.000   CP
  Seat 6        0.800   0.600   0.667   SDF
  Seat 7        0.114   0.600   0.667   CP
 --------------------------------------------
  Seats         3       2       2 
  Percent       42.8%   28.5%   28.5% 

Fascinating, isn't it?

I'm amazed that even in the first vote we had suspected fraud, tactical voting between both factions and members, and destructive voting. Sounds like politics have come to SL.

~Ulrika~

if you know who the alt was that voted they should be outed

I admit that I really have no interest in the Neuraltenberg project except as an example of political wranglings. Even though the Clown Party was - self-admittedly - a joke Party, I find it interesting that they were simultaneously on the ballot and disqualified yet would have taken a number of seats had they not been disqualified.

I do not dispute the Projeckt's autonomy in disqualifying a party, but think this an example of those who control the voting process control the outcome. I expect there was a point to this exercise, but haven't fully parsed the explanations yet.

Posting before seeking understanding? You wouldn't be Malachi, if you didn't.

~Ulrika~

I don't actually know who it was. I built so much anonymity into the system that it would be difficult to identify them without a lot of coding and I am so over coding right now.

Plus, this was really just a test election. I think it has more value in rooting out the bugs in our voting system and providing lessons for future voters.

~Ulrika~

Thanks for all your work on setting up and running the voting system Ulrika.

Although I'm not sure fascinating is quite the right word for all the analysis.

When will we know the names of the elected?

Why thank you sir! I looove infrastructure. Wait 'till you hear what I have next on my agenda. You'll all flip.

I hope to release the list of names by Friday sometime. I have heard back from four members, am still waiting for three, and need to contact a few more. In the future I'll automate this task.

~Ulrika~

I'm reading all this mind-boggled.

Looks like Ulrika is a math professor with an expertise in election statistics! I never knew!

Contrary to the above post, I feel extremely confident that with this kind of understanding of the election process, we have the most profoundly fair one possible. I only hope that, as time goes on, the rest of us can contribute a similiar level of expertise in our own areas.

Thank you Ulrika.

Sudane

My apologies for posting this so late. Unfortunately, my lovely PowerBook's power supply blew up this weekend, and I was unable either to log in on Second Life as much as I liked, as well as being unable to read my emails

Having fixed that, here are the official results of the Representative Assembly (RA) elections:

Gwyneth Llewelyn SDF Leader
Pendari Lorentz MPP Leader
Roberta Dalek SDF
Sudane Erato SDF
Cristiano Midnight MPP
Talen Morgan MPP
Chandra Page SDF

Congratulations to all elected members

The RA had their first meeting last Sunday and we discussed mostly one single issue: how the Neualtenburger Projekt is going to continue beyond our alloted time (as you know, our lease expired "officially" last week). This will be the first and most important task for the RA on the next weeks. At the same time, we will discuss the necessary changes to make the Projekt economically sound - mostly by renting space in Neualtenburg and eventually changing the way the City Vendors will/should be used by Guild members. All this will probably occupy our time for the immediate future.

Besides the "formal in-world discussions" on the Senate, the members of the RA have also exchanged informal emails. This means that much of the work gets coordinated while most of us are working IRL and we can meet together just to do some voting or quick opinion polls. We can assure the citizens of Neualtenburg that our work at the RA "goes on" even if you don't see us in-world all the time. Think of our off-world work as "comittees", preparing work outside the RA, and coming together with all our homework done, and do the voting on the many proposals.

Gwyneth,

Congratualtions to you and your collegues in the First Representative Assembly of the City of NeualtenBurg.

Is it time for a party?

Did someone say party?

I'm in! Let's set a time and place. We should have fireworks too!

~Ulrika~

... will that be a political party?