Election reform bill - counting methods
That following the refusal of the SC to consider the submission with regard to the recent July 2008 election and the concern expressed about the 'after the fact decision as to which counting method to adopt' the following bill is submitted for RA consideration.
That this current election result be considered void.
That the RA legislate on the method of counting to be used
That the election be rerun using the agreed single counting method
****************************
The submission to the SC re the July 2008 CDS election
That the election was run with two methods of counting the result
That the two methods were believed by the election official (Jon Seattle) to not make any difference (see forum post below April 27th and Feb 5th 2008)
Had the RA known that the two counts could make a difference it would have had the chance to legislate on the accepted method.
In this election the two methods delivered a different result despite the previous prediction - this was therefore unexpected by the SC and their representative
That this placed CDS and the SC in a difficult position i.e. having to decide 'after that fact and when the two possible results were known' which method to use
That the difference in seat allocation could make a large political difference to the coming CDS RA session.
That the SC does not represent all factions in CDS being one independant and two DPU members
That the method chosen by the SC favours DPU (this to be fair may only be the appearance of bias)
That based on comments by Flyingroc all previous elections 'historically used a variant of the Borda system (aside from the past two elections) that weighted the lowest-ranked with 0 points' - which indicates that a new precedent was set for the last two elections.
What method of counting gives the most representative government.
Therefore it is submitted
That this current election result be considered void.
That the SC urgently request the RA to legislate on the method of counting to be used
That the election be rerun using the agreed single counting method
Extracted evidence from chat and forums
July 2008 Election Resuls (crosspost)
by Claude Desmoulins on Sun Jul 20, 2008 10:13 pm
Please congratulate the new LRA, Justice Soothsayer of the CSDF
Other CSDF members of the new RA are Arria Perrault and one seat TBD.
DPU members of the new RA are Rubaiyat Shatner and 2 seats TBD
The NuCare member of the new RA is Bells Semyorka
Last term a precedent was set with the SP that factions winning more seats in the election that they had candidates on their lists have the opportunity to fill those seats. The situation has occurred again this time.
Also , the rare circumstance arose wherein a change from zero based Borda to One Based Borda would have altered the outcome of the election to a 3-2-2 division of seats. The SC , based on the stronger historical precedent of Zero-Borda use in the CDS, chose this calculation method.
Transcripts of the SC certification discussion and comments will follow.
Claude Desmoulins
Dean of the SC
Posts: 580
Joined: Thu May 25, 2006 4:28 pm
Location: Neufreistadt
Private message
Re: July 2008 Election Resuls (crosspost)
by Jon Seattle on Mon Jul 21, 2008 12:07 am
Here are the numbers. I will prepare a more detailed matrix later. These are zero-based Borda scores.
CSDF (First Ranked = 28, Borda Score = 61, seats = 3)
----
Justice Soothsayer
Arria Perreault
DPU (First Ranked = 8, Borda Score = 54, seats = 3)
---
Rubaiyat Shatner
NuCARE (First Ranked = 12, Borda Score = 29, seats = 1)
Re: July 2008 Election Resuls (crosspost)
by Claude Desmoulins on Mon Jul 21, 2008 1:58 am
Bromo,
DISCLAIMER : I SPEAK ONLY FOR MYSELF
Deciding such a thing before the fact is the kind of policy making exercise the SC usually tries to avoid. Especially since the issue of 1-borda v 0-borda was raised at the last general election, there was a considerable time during which a policy making body could have clarified the 1 borda/zero-borda question. To the extent there was a precedent the SC followed it (AFAIK, in seven prior general elections zero borda was used six times and one borda once)
I think you are right that there are a number of issues here.
Re: July 2008 Election Resuls (crosspost)
by Bromo Ivory on Mon Jul 21, 2008 12:08 pm
Jon Seattle wrote:Since my role in this is just as a volunteer providing a technical service, I am very careful not to make policy decisions. Constitutionally, I work for the SC under their supervision. Because this had been a question before, the counting method did make a difference this time, after the polls closed I submitted the question to the SC. They, quite correctly, confirmed that the zero-based counts are correct. There was, contrary to Bromo's assertion, no decision making based on outcome.
#1 - I am not accusing you of anything, I am only commenting upon the original post by Claude. I am not sure what "not making policy making" you are talking about - last time I checked you aren't a SC member? In my mind, as long as you weren't expressing opinions on the method of Borda counting before decisions were being made or revealing votes to non SC members until the SC had made up their mind, I would have no issue (and the first part simply because you are privy to votes, too).
#2 - I read the original post and it seemed to me a decision was made. My issue is that the SC had all the votes and then had to make a decision on how to count them. This is backwards. Even if the RA did not clarify anything, the SC needs to have made a decision before the polls open and then stick to it. We can have 11,000 posts back and forth on how "yes it was" and "no it wasn't" which is tedious and tiring - but we don't even to have a discussion if the method of tallying votes were decided before the votes were cast.
The base fact was the votes came in and then the SC decided how to count them. According to the posts by Claude. This is an enormous problem and in a real life country this would make the elections illegitimate. I agree that the SC was placed in an awkward decision, but the remedy was very simple.
I won't even go into other issues - which is the "ghost" RA seats and the minimum faction membership requirements. While the SC has declared this "precedent" I do think it is a bad percedent because we allow non faction members to rank those in a faction. This brings up a good legal question as well "What is a faction member?"
Zero-Based Borda Counts
by Jon Seattle on Sun Apr 27, 2008 1:12 am
Flying Roc asked about zero-based Borda counts. As in the prior election, it does not make a difference. But if you are curious here they are:
DPU (Borda Score = 56, seats = 1)
---
Flyingroc Chung
CSDF (Borda Score = 48, seats = 1)
----
Gwyneth Llewelyn
NuCARE (Borda Score = 25, seats = 0)
------
Cindy Ecksol
The software applies the Sainte-Laguë with the numbers from scratch (for each method of computing). In neither case did it change the outcome. For a very long and detailed look at the difference please see the postings on this issue that occurred after the prior election:
viewtopic.php?f=1&t=1585&st=0&sk=t&sd=a&start=15#p9693
I much prefer the zero-based method, but much of the literature I find on the subject uses one-based Borda and the current law is that the constitution be read literally. Claude has a copy of the (anonymous, random ordered) votes and the software used to do the calculations. It would be a good idea to change to zero-based calculations for future elections.
Jon Seattle
I need a hobby
Posts: 547
Joined: Mon May 29, 2006 1:18 pm
Location: Evanston IL
Private message
Re: A Final Report on the Election
by Jon Seattle on Tue Feb 05, 2008 8:01 pm
Its a non-issue. The SC and I have already had this discussion with Sleazy. I implemented Broda counts as the constitution requires, using the published definition from seven different sources. Borda counts typically start with one. And it did not matter one bit in this election. (See http://en.wikipedia.org/wiki/Borda_count .)
1. The difference between 1 and 0 based Borda counts is to add a constant to every count for every faction.
2. In this election, I programmed both methods to make sure, and there was no difference in outcome.
3. The scores published were the consistent with the published definition, and there is unlikley to be a difference, but I would prefer we use zero-based counts in future elections. However, the constitution will have to be changed to allow this.
Here is more on each point:
1. The difference between 1 and 0 based Borda counts is to add a constant to every count.
Re: Zero-Based Borda Counts
by Flyingroc Chung on Sun Apr 27, 2008 2:35 am
Although the constitution is not exactly clear on this, we have historically used a variant of the Borda system (aside from the past two elections) that weighted the lowest-ranked with 0 points. The idea was that in a two-faction election, we would end up with a straight St. Lague allocation of seats. Here's an old post of Ulrika's (back in 2005) where she stated as much:
http://forums.secondlife.com/showthread.php?t=56741
Ulrika posted the equation we used as:
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)
Since AFAIK, Ulrika was essentially the SC at that time, I've in the past assumed that this way of counting is constitutional. When we did the first 3-faction elections in 2006, Ulrika raised the counting issue again, and note that in this discussion, we assumed that last place had 0 points:
http://forums.secondlife.com/showthread.php?t=82946
In the Jan 2007 elections, I again reiterated that we used a 0-based counting (with a slight change to use integer rather than fractional counting):
viewtopic.php?f=1&t=844&p=5231#p5231
So we have here on the one hand historically one way of counting borda scores, and on the other hand a different one based on Jon's more literal interpretation of the constitution (0-based borda scores are unusual, which can possibly be interpreted as not really a borda count; I'm not sure I agree with this, but *shrug*. (note also that 0-based and 1-based ranking do not matter in a winner-take-all environment))
Since Jon prefers 0-based counting, and since we have historically done 0-based counting anyway, I'm going to propose that we use the 0-based count from now on. I think we can do this as regular law (no need for a constitutional amendment) by defining a standard ranking for borda counts in the CDS.
**************************************
ThePrincess Parisi: claude when do we announce the seats
[2008/07/19 13:12] Claude Desmoulins: tOMORROW AFTERNOON SO WE CAN sorry
[2008/07/19 13:12] ThePrincess Parisi: why not today
[2008/07/19 13:12] ThePrincess Parisi: its always today
[2008/07/19 13:12] Claude Desmoulins: sort our a couple of wrinkles.
[2008/07/19 13:12] ThePrincess Parisi: dpu
[2008/07/19 13:12] Claude Desmoulins: No, It used to be Sundays
[2008/07/19 13:12] ThePrincess Parisi: ?
[2008/07/19 13:13] ThePrincess Parisi: no its the day the electinos are over
[2008/07/19 13:13] ThePrincess Parisi: always
[2008/07/19 13:13] ThePrincess Parisi: same day
[2008/07/19 13:13] ThePrincess Parisi: within hours
[2008/07/19 13:13] ThePrincess Parisi: last two terms
[2008/07/19 13:13] Claude Desmoulins: There are also one borda-zero borda issues on which the SC needs to weigh in
[2008/07/19 13:13] ThePrincess Parisi: what does that mean
[2008/07/19 13:14] ThePrincess Parisi: a tie or what
[2008/07/19 13:14] You: what the problem?
[2008/07/19 13:14] You: with the count?
[2008/07/19 13:14] Claude Desmoulins: Depending on whether you use 3-2-1 or 2-1-0 results change
[2008/07/19 13:14] ThePrincess Parisi: well they cant change it mid steream
[2008/07/19 13:14] You: cant change after the fact
[2008/07/19 13:14] ThePrincess Parisi: are you kidding
[2008/07/19 13:15] Claude Desmoulins: no
[2008/07/19 13:15] You: so what was the start count method?
[2008/07/19 13:16] Claude Desmoulins: there's some historical use of 1 bodra and some of zero
[2008/07/19 13:16] You: when did it changed and why
[2008/07/19 13:16] Claude Desmoulins: constitution and code are silent.
[2008/07/19 13:17] Claude Desmoulins: the election software was recoded.
[2008/07/19 13:26] IM: Jon Seattle: Ah, the software is very flexible, it just counts both ways and lets the SC decide. It always has done that.
[2008/07/19 13:26] IM: ThePrincess Parisi: what a total joke
[2008/07/19 13:26] IM: ThePrincess Parisi: are you kidding the SC decides which way the votes are counted.. lmao
[2008/07/19 13:26] IM: MT Lundquist: what they decide how to take the result AFTER THE FACT
[2008/07/19 13:27] IM: ThePrincess Parisi: and we pay money for this democracy.. can you imagine in RL?
[2008/07/19 13:27] IM: ThePrincess Parisi: im laughing so hard.. i hope you all are reading this !!!
[2008/07/19 13:27] IM: Jon Seattle: Up to now there has been no difference between the two menthods of counting.
[2008/07/19 13:27] MT Lundquist: this needs to go the RA next week
[2008/07/19 13:27] MT Lundquist: we need legislation to make this NOT A JOKE
[2008/07/19 13:27] Jon Seattle: Both FR and I agree on the proper method and we posted about it months ago.
[2008/07/20 15:02] MT Lundquist: Claude I wish to formally appeal the election in CDS based on the fact that the results were necesarily decided AFTER the fact
[2008/07/20 15:02] Second Life: User not online - message will be stored and delivered later.
[2008/07/20 15:02] MT Lundquist: there i wish the SC to consider that the election be rerun when the method of counting is fixed
[2008/07/20 15:02] Second Life: User not online - message will be stored and delivered later.