Wikipedia:Requests for comment/Arbitration Committee Elections December 2019

2019 Arbitration Committee Elections

Status as of 15:13 (UTC), Saturday, 20 July 2024 ()

• Thank you for participating in the 2019 Arbitration Committee Elections. The certified results have been posted.
• You are invited to leave feedback on the election process.

---

The purpose of this request for comment is to provide an opportunity to amend the structure, rules, and procedures of the December 2019 English Wikipedia Arbitration Committee election and resolve any issues not covered by existing rules. 21:02, 31 August 2019 (UTC)

Background: In the case of proposals that change existing rules, or that seek to establish new ones, lack of consensus for a change will result in the rules from the 2018 election remaining in force. Some issues are not covered by the existing rules but will need to be decided one way or another for the operation of the election, in those cases it will be up to the closer to figure out a result, even if there is no clear consensus, as they have had to in the past.

ACERFC decisions to date

• Election: There will be 13 arbitrators, six seats on "Tranche Alpha" from the 2018 elections and seven seats on "Tranche Beta". A maximum of seven 2-years term are up for elections each year, with the seventh seat switching tranche if necessary. Any additional vacancies which open before voting begins will be filled by a 1-year term. A minimum 50% support is required to be elected for a 1-year term, and a minimum 60% support is required for a 2-years term. Successful candidates with the lowest support percentages are given the 1-year term if any. If there are more vacancies than candidates with the required minimum, the extra seats will remain unfilled.
• Candidates: Registered account with 500 mainspace edits, editor in good standing "that is"/"and is" not under block or ban, meets Foundation's Access to nonpublic personal data policy, and has disclosed alternate accounts (or disclosed legitimate accounts to Arbcom).
• Electoral Commission: A RFC to appoint 3 Electoral Commission officials who will solve disputes and problems during the election. Commission should only intervene when there is a problem that needs resolving, and either discussion isn't working, the rules are unclear, or there isn't time. Open to anyone who is over 18, meets Foundation's Access to nonpublic personal data policy, and otherwise be eligible to vote. Appointments to the Commission should be confirmed by the Arbitration Committee per the CheckUser policy.
• Timeline:
  • ACE RFC: (30 days of September)
  • Electoral Commission RFC: 7 days nominations, 7 days evaluation, selection by 7 days after close of evaluation. (October)
  • Nominations: 2nd Sunday of November (10 days)
    • Nomination is hard deadline for creation and transclusion of nomination statement. How to handle any site-wide disruption is at the discretion of the Electoral Commission
  • Fallow period: (5 days)
  • Voting period: (14 days)
  • Scrutineering: No deadline for releasing or announcing the results.
• Guides: Allowed but with some strong suggestions. Must be allowed reasonable visibility.
• General Guide: Wikipedia:5 Minute guide to ArbCom Elections created and advertised.
• Voting: A voter needs 150 mainspace edits by 1 November and registered an account before 28 October, not currently blocked at the time of voting. Voting system of (Support/Abstain/Oppose) will be used with percentages calculated via Support/(Support + Oppose). Secret ballots via SecurePoll will be used.
• Scrutineering: 3 functionaries from outside en.wiki as scrutineers.
• Ordering: The order of candidates are software randomised on the candidate page, and on the ballot.
• Warning: Potential candidates are warned of risks from standing for election with message similar to that on WP:CUOS2015 to be incorporated into the candidate instructions page.
• Questions: No standard questions for every candidates. No limits on the number of questions, but candidates are not obligated to answer every questions. Electoral Commission (as a group not individually) have the discretion to remove offensive (eg. WP:POLEMIC-style statements) or off-topic questions from question pages, following discussion among the Electoral Commission members. While other editors can obviously remove clear vandalism, egregious personal attacks, etc., the determination of what is inappropriate or off-topic is clearly to be left to the Commission.
• Advertising: Traditional notices posted to various community noticeboards, watchlist notice and/or central notice banner for election in general (not individual candidates), Mass Message - eligible voters, have edited last 12 months before nominations. Extra care should be taken in wordings of advertising to make sure it's neutral.
• Blocking: Blocking policy applies normally, but a candidate shouldn't be disqualified for being blocked (except for sockpuppetry) after nominating him/herself.

References

Structure: This RfC is divided into portions, each of which contains a discussion point for the community. The standard RfC structure will be used, in which any user may make a general statement that other users may endorse if they so agree. The points will be listed in the table of contents below, along with the users who have made statements.

Per the consensus developed in previous requests for comment, the electoral commission timetable is as follows:

• Nominations: Saturday 00:00, 5 October – Friday 23:59, 11 October (7 days)
• Evaluation period: Saturday 00:00, 12 October – Friday 23:59, 18 October (7 days)
• Commission selection: completed by Friday 00:00, 25 October

Per the consensus developed in previous request for comments, the arbitration committee election timetable is as follows:

• Nominations: Sunday 00:00, 3 November – Tuesday 23:59, 12 November (10 days)
• Setup period: Wednesday 00:00, 13 November to Sunday 23:59, 18 November (5 days)
• Voting period: Tuesday 00:00, 19 November to Monday 23:59, 2 December (14 days)
• Scrutineering: Begins Tuesday 0:00, 3 December

Anyone is free to raise any new topics that they feel need to be addressed by adding them as level two headers.

Duration: In order to preserve the timeline of the election (see above), we should aim to close this RfC as soon as 30 days have passed, i.e. on or after September 30, 2019. The results will determine the structure, rules, and procedures for the election.

---

Use the following format below; post a new statement at the BOTTOM of the section in which you want to make a statement. Endorse by adding a hash symbol (#) and your signature.

===Statement #N by [[User:USERNAME|USERNAME]]===
Comment ~~~~

==== Users who endorse statement #N: ====

#~~~~

Users should only edit one summary or view, other than to endorse.

Points of discussion

Number of arbitrators

Statement #1 by Rschen7754

Proposal

• Elect enough arbitrators to bring the committee back to a size of 15. As of this writing, that is 8 2-year terms and 1 1-year term. Additional resignations/removals would be replaced by more 1-year terms.
• Leave the minimum percentages of support at 50% for a 1 year term and 60% for a two year term.

Rationale

Well, that didn't work - we wound up with 10 arbitrators, quite a few of those with large periods of inactivity, and a lot of difficult issues. Maybe it would have worked out better had 2019 been more like 2018 or 2017, but oh well. This restores the committee back to the usual size, but keeps the minimum support requirements that addresses the concerns of fewer candidates running and "less suitable" candidates getting elected. Rschen7754 21:24, 31 August 2019 (UTC)

Users who endorse statement #1 (number of arbitrators):

1. As proposer. I know I'm jumping the gun a bit here, but I proposed the original reduction, and I want to do my part in making things right. --Rschen7754 21:28, 31 August 2019 (UTC)
2. I agree, but I also wonder if we want to increase the committee further, by perhaps another 2 arbitrators. This would be to allow all our arbitrators to have some 'holiday' time away from the committee without having to feel that they're neglecting their position etc. Nick (talk) 21:34, 31 August 2019 (UTC)
3. I agree, though I don't agree with a larger increase. There were reasons for the reduction but we've seen the risks. I think the negatives of potentially long conversations from too many ARBs are outweighed by the issues from insufficient viewpoints being provided in decision making, due to resignations/inactive. Nosebagbear (talk) 22:26, 31 August 2019 (UTC)
4. Kevin (aka L235 · t · c) 22:29, 31 August 2019 (UTC)
5. I might even be open to a larger increase, but this seems like a reasonably conservative move for now. StudiesWorld (talk) 22:59, 31 August 2019 (UTC)
6. Per Nosebagbear's final sentence. Thryduulf (talk) 23:24, 31 August 2019 (UTC)
7. This is a step in the right direction, and I'm hopeful that we'll be able to increase the size of the Committee as Nick suggested in the future. I don't think this would be an appropriate time to do so yet, though. Supporting as proposed by Rschen7754. OhKayeSierra (talk) 01:49, 1 September 2019 (UTC)
8. We could use a few more arbs. Reducing the size of the committee was a good idea last year, but it didn't work out the way we hoped. – bradv🍁 01:59, 1 September 2019 (UTC)
9. More arbitrators increases the chances that the committee better reflects the community as a whole. Calidum 02:32, 1 September 2019 (UTC)
10. Pharaoh of the Wizards (talk) 03:52, 1 September 2019 (UTC)
11. More arbs hopefully means greater diversity of views while achieving a greater spread of the workload and allowing scope fo recusals/breaks. I think adding two more would have the potential to make ArbCom too unwieldy. Peacemaker67 (click to talk to me) 04:40, 1 September 2019 (UTC)
12. SQL^{Query me!} 04:52, 1 September 2019 (UTC)
13. Beyond My Ken (talk) 05:19, 1 September 2019 (UTC)
14. CThomas³ (talk) 05:56, 1 September 2019 (UTC)
15. TheSandDoctor ^Talk 06:10, 1 September 2019 (UTC)
16. –Dave | Davey2010^Talk 11:05, 1 September 2019 (UTC)
17. --qedk (t 桜 c) 11:18, 1 September 2019 (UTC)
18. To add a perspective from inside the committee: we had two long-term inactive arbs, and typically 2-3 temporarily inactive at a given time, so functionally we were working with 8-9 at the beginning of the year and now after the resignations 6-7. My impression from the beginning of my term was that we were struggling to keep up with the day-to-day workload, leaving some larger issues unaddressed, and that was before our busy summer. I was surprised by how much of the workload is not visible on-wiki: appeals, drafting, functionary oversight, refining internal processes, etc. Those things typically don't require much discussion and would simply benefit from having more hands. I'd say restoring the committee size to 15 is essential. More could be nice but is probably too large a step now, given the amount of empty spot to fill anyway. To compensate, it would help if voters looked for a strong commitment to activity from candidates. – Joe (talk) 11:58, 1 September 2019 (UTC)
19. The current size of the committee is too small to effectively handle spikes in case activity. While the case activity this year does not appear to be representative of the activity in average years, the committee should be sufficiently large to meet proposed deadlines (cf. Wikipedia talk:Arbitration/Requests/Case/Antisemitism in Poland/Proposed decision), or at a minimum, communicate changes to them in a timely fashion. The possibility of edge cases in resignations and recusals should also be taken into account, and should not put the committee into duress. — Newslinger talk 12:44, 1 September 2019 (UTC)
20. -- Tavix ^(talk) 00:25, 2 September 2019 (UTC)

Comments (number of arbitrators #1):

Input from current arbitrators on workload is critical here to determine what number we aim for. Cas Liber (talk · contribs) 23:32, 31 August 2019 (UTC)

Statement #2 by Rschen7754

Leave the committee at a size of 13.

Users who endorse statement #2 (number of arbitrators):

1. There is absolutely no evidence that more arbitrators increases throughput, and based on what I've heard of some former arbitrators it might actually decrease it as more people need to state their opinion for stuff to proceed. As-is it seems like the Committee is large enough to be representative. Jo-Jo Eumerus (talk, contributions) 07:44, 1 September 2019 (UTC)

Statement #N by USERNAME

Text of proposal ~~~~

Users who endorse statement #N (number of arbitrators):

1. ~~~~

Procedures for emergency elections

Statement #1 by TonyBallioni

In the case where the arbitration committee calls emergency or interim elections, the same rules as the past Arbitration Committee Election shall apply with the following modifications:

• The time period for nominations shall be one week.
• Voting shall commence as soon as possible upon the close of nominations.
• The previous election commission will be reappointed. If they are unable or unwilling to perform their job, members of the current functionaries team who are not arbitrators will be asked to volunteer without the need for confirmation.
• Arbitrators elected under this abbreviated process shall have their terms automatically expire on 31 December unless they are confirmed in the next general arbitration committee election.
• No interim elections may expand the size of the committee beyond that authorized in the previous election RfC.
• The Election Commission is empowered to resolve by majority vote situations unforeseen in the previous election RfC that may prevent emergency or interim elections from being held.

Rationale

We saw this year that while the Arbitration Policy has a clause for emergency elections, it’s toothless because the elections take roughly 3 months to run, so by the time you have a new arbitrator, it’s time for the next general ACE. This despite this year being the only time in recent memory where emergency elections may have been helpful. This proposal solves that by creating an abbreviated procedure for snap elections when needed without the need for a new RfC, and limits the length of service because of the abbreviated procedures, so you don’t end up with someone serving a 21 month term without the traditional scrutiny. TonyBallioni (talk) 21:37, 31 August 2019 (UTC)

Users who endorse statement #1 (emergency elections):

1. This was a real impediment to calling an election. This makes a snap election much easier to administer, which is as it should be. Best, Barkeep49 (talk) 22:21, 31 August 2019 (UTC)
2. Nick (talk) 21:46, 31 August 2019 (UTC)
3. This is sensible as it avoids launching a three-month process to fill what would presumably be an immediate need in a snap election. Mz7 (talk) 22:03, 31 August 2019 (UTC)
4. Obviously as I proposed it. TonyBallioni (talk) 22:28, 31 August 2019 (UTC)
5. Kevin (aka L235 · t · c) 22:29, 31 August 2019 (UTC)
6. Definitely resolves an issue, with only minor negatives. Taken to mean replacing resigned ARBs or removed ARBs, not those who just happen to be inactive. Nosebagbear (talk) 22:29, 31 August 2019 (UTC)
7. El_C 22:40, 31 August 2019 (UTC)
8. * Pppery * _{it has begun...} 22:50, 31 August 2019 (UTC)
9. Very sensible. Thryduulf (talk) 23:27, 31 August 2019 (UTC)
10. Umm...this seems sensible. Cas Liber (talk · contribs) 23:28, 31 August 2019 (UTC)
11. StudiesWorld (talk) 23:59, 31 August 2019 (UTC)
12. OhKayeSierra (talk) 01:51, 1 September 2019 (UTC)
13. This is a sensible idea, and wouldn't require any changes to ARBPOL. – bradv🍁 02:01, 1 September 2019 (UTC)
14. Support, though I would also support giving the community the ability to call for a special election if it deemed one necessary. Calidum 02:34, 1 September 2019 (UTC)
15. Pharaoh of the Wizards (talk) 03:52, 1 September 2019 (UTC)
16. SQL^{Query me!} 04:52, 1 September 2019 (UTC)
17. Beyond My Ken (talk) 05:19, 1 September 2019 (UTC)
18. This is a sensible provision. --TheSandDoctor ^Talk 06:12, 1 September 2019 (UTC)
19. Indeed a very sensible idea. –Dave | Davey2010^Talk 11:08, 1 September 2019 (UTC)
20. --qedk (t 桜 c) 11:17, 1 September 2019 (UTC)
21. This proposal makes the committee more resilient against edge cases in resignations. — Newslinger talk 12:59, 1 September 2019 (UTC)
22. Kudpung กุดผึ้ง (talk) 21:39, 1 September 2019 (UTC)

Comments (emergency elections statement #1)

Would this replace inactive arbitrators or only resigned ones? If replacing inactive arbitrators as well, then the number of elected arbitrators should be allowed to exceed the number of arbitrators selected in the last elections RfC up to the number of inactive arbitrators. I would further stipulate that as an arbitrator returns to active duty, the interim arbitrator that received the fewest votes should be dismissed from the committee. isaacl (talk) 22:00, 31 August 2019 (UTC)

I endorsed it under the idea it would only replace resigned Arbs. I think dealing with inactive ones should be under a different statement. Best, Barkeep49 (talk) 22:04, 31 August 2019 (UTC)

It’d be resigned or removed for inactivity. Having inactive arbitrators is not in itself a bad thing if they may return to activity. The committee has removed for inactivity before, so if they are at the stage or calling interim or emergency elections, that is something they could consider. TonyBallioni (talk) 22:08, 31 August 2019 (UTC)

For any proposed interim election format, I suggest including an emergency ripcord allowing the arbitration committee to define an emergency procedure, should circumstances warrant. By their very nature, it's hard to predict exactly what will be needed in emergencies, and I think some flexibility is desirable. isaacl (talk) 22:00, 31 August 2019 (UTC)

@Isaacl: What sorts of decisions and procedures do you imagine that the election commission could not handle but ArbCom could? Best, Kevin (aka L235 · t · c) 22:27, 31 August 2019 (UTC)

I’d very strongly oppose that. Elections, even emergency or interim, should be conducted as independently from the committee as possible. TonyBallioni (talk) 22:08, 31 August 2019 (UTC)

Right now the arbitration committee possesses the authority to trigger interim elections. This already vests the flexibility to deal with emergencies in its hands. Do you propose changing this? isaacl (talk) 22:24, 31 August 2019 (UTC)

@Isaacl: I agree with @TonyBallioni:. I'm debating to take it to the stage of adding an additional suggested change that if the number of active arbitrators drops below, say, 8, for more than a month, it would automatically trigger an emergency interim election. Nosebagbear (talk) 22:37, 31 August 2019 (UTC)

I believe that would require a change to the arbitration policy, as it currently vests the power to initiate an interim election entirely in ArbCom. Best, Kevin (aka L235 · t · c) 22:39, 31 August 2019 (UTC)

The arbitration committee could choose to modify its own procedures to specify triggers for an interim election. However it wouldn't be bound to the triggers in theory, since the committee could retract them at will. Doing so would, of course, run the risk of losing community trust. isaacl (talk) 03:29, 1 September 2019 (UTC)

No. I don't suggest taking it away from them. I oppose giving them the power to regulate how the election they call is run. The existing committee should not be in the business of creating the rules for electing fellow members. TonyBallioni (talk) 22:48, 31 August 2019 (UTC)

I'm considering the case where the number of arbitrators falls below a viable quorum, some urgent situation arises which the community wants to have handled by arbitrators, and there is a lack of time to follow the proscribed interim election process. The default way around, of course, is for the community to hold a request for comments discussion deciding on how to deal with the situation, but it can be hard to get a consensus agreement in a short time frame. To be honest, I think most people understand that establishing consensus requires patience, with the resulting consequence that action has to be delayed to let the consensus process play out. If that means waiting for interim elections to be completed, or if near the year's end, waiting for the regular election to be held, so be it. isaacl (talk) 03:20, 1 September 2019 (UTC)

• How long will the voting period last? To be honest, I suspect if this were to happen then there would be a lot blow-back saying "there wasn't enough time", either for voting, or for vetting the candidates. — Ched (talk) 22:15, 31 August 2019 (UTC)
  • @Ched: I'm guessing that since the rules will remain the same except where changed above, two weeks will still be the voting period. And since (hopefully) there won't be more than three or four arbs being replaced, the arbitrators elected will receive a generally higher proportion of the vote than in our annual elections, where there could be eight or nine arbs being elected at once; perhaps that will dampen any objections on the basis that the candidates were insufficiently vetted. Best, Kevin (aka L235 · t · c) 22:18, 31 August 2019 (UTC)

    While some voters do strategically vote I don't think we know how frequently this is done to suggest fewer elections would lead to a higher overall candidate approval. Best, Barkeep49 (talk) 22:20, 31 August 2019 (UTC)

    I guess I'm presupposing that the same candidates run in an interim election as in an annual election; the top three or four candidates in our elections tend to have a higher level of support than the top seven or eight. Best, Kevin (aka L235 · t · c) 22:22, 31 August 2019 (UTC)

    I would hope the interim election would bring us a diferent variety of candidates. For instanace, a former arb might be willing to do six interim months of ArbCom but not a full 1/2 year term. Best, Barkeep49 (talk) 22:37, 31 August 2019 (UTC)

    (General reply to this thread) While I'm sure my proposal has issues, any method of holding interim or emergency elections would. The current issue is that it is next to impossible to hold them because by the time the need is identified, the process to set up elections would take too long for holding them to be practical (in any real world scenario, the need isn't going to be apparent until around 1 June at the earliest). A shortened process of some sort is needed, otherwise holding elections when needed is not an option.

    In my view, having a hard cap on the length of the terms that an interim election will produce is a check on the period for vetting being shorter. We aren't going to end up with someone serving for 21 months out of this. It would be 4-6 months until the next general ACE, and I think that's low enough of a risk for me to be comfortable. TonyBallioni (talk) 22:45, 31 August 2019 (UTC)

    Uh, doesn't six months=1/2 year? Liz ^{Read! Talk!} 22:50, 31 August 2019 (UTC)

    I believe my friend was referring to "one or two years" as a term Kevin (aka L235 · t · c) 22:58, 31 August 2019 (UTC)

• @TonyBallioni: would you see this otherwise following the mechanisms of the general election? (e.g. The creation of an election commission, configuration of SecurePoll, etc?) — xaosflux ^Talk 00:46, 1 September 2019 (UTC)
  • @Xaosflux: the election commission established in the previous ACE would be the effective election commission unless any members were not available/didn't want to. In which case, non-arb functionaries would volunteer (the third bullet point). Having functionaries serve as backup is mainly for convenience as they've already signed the NDA and typically have the trust of the community, as well as having more frequent interaction with stewards than most of en.wiki.
    Re: secure poll, yeah, that'd have to be configured, which is why I changed the start of voting from "immediately" after the close of nominations to "as soon as possible". My vision here isn't for a perfect framework for a frequent event, but for a workable framework for an infrequent one. TonyBallioni (talk) 01:09, 1 September 2019 (UTC)

• @TonyBallioni: Newbie question, but what does: The Election Commission is empowered to resolve by majority vote situations unforeseen in the previous election RfC that may prevent emergency or interim elections from being held. refer to? –MJL ‐Talk‐^☖ 02:02, 1 September 2019 (UTC)
  • @MJL: every arbitration committee election has an election commission made of three individuals elected by the community to oversee it. The commissioners are ineligible to run in the election themselves. The bullet point you quote basically means that if something comes up that couldn't have been predicted in September of the year before an interim election, the commissioners can decide how to move forward with holding a timely election.
    Having been around for a number of these, WikiLawyering over the exact rules of the election is pretty much par for the course, and this just says that the commission has the authority to make sure an election happens if things don't go perfectly, which given that it would be an interim/emergency election, is pretty likely. TonyBallioni (talk) 02:29, 1 September 2019 (UTC)
• I'd rather see an automatically-triggered election. I'm dubious that ArbCom would ever call them. They haven't yet.--Wehwalt (talk) 05:59, 1 September 2019 (UTC)

  @Wehwalt: We discussed it after the resignations this summer and decided not primarily because of how much time and effort it takes (rendering it pointless, as Tony says). – Joe (talk) 12:18, 1 September 2019 (UTC)

  Joe Roe, Understandable. That's why I favor a trigger, if say Arbs are down five members by June 30, say. That way, the procedures are in place and it's less of a fuss all around. Wehwalt (talk) 15:59, 1 September 2019 (UTC)

  Joe Roe And I meant more in general than this particular ArbCom in particular. Wehwalt (talk) 16:48, 1 September 2019 (UTC)

Statement #2 by TonyBallioni

Do not establish emergency election procedures. TonyBallioni (talk) 21:40, 31 August 2019 (UTC)

Users who endorse statement #2 (emergency elections):

1. ~~~~

Statement #N by USERNAME

Text of proposal ~~~~

Users who endorse statement #N (emergency elections):

1. ~~~~

Blocked candidates

Statement #1 by TonyBallioni

Users that are indefinitely blocked or site banned at the start of the voting period shall be ineligible for election to the arbitration committee. Candidates who are only temporarily blocked at the time of the start of voting are still eligible for election. TonyBallioni (talk) 00:07, 1 September 2019 (UTC)

Users who endorse statement #1 (Blocked candidates):

1. Wasn't planning on making any additional proposals until I read the green box above. This has been how it was interpreted in the past (see the Dysk ArbCom run of two years ago), but since the big green box at the top of the screen tells me otherwise, sure, let's formalize this. TonyBallioni (talk) 00:07, 1 September 2019 (UTC)
2. SQL^{Query me!} 04:53, 1 September 2019 (UTC)
3. The problem of a "rogue administrator" (see below) is unlikely to occur and, if it did, would presumably result in emergency desysopping and restoration of the candidates affected to the ballot. See WP:IAR, WP:BURO, WP:COMMONSENSE Beyond My Ken (talk) 05:23, 1 September 2019 (UTC)
4. Formalizing this makes sense in my view. As Beyond My Ken stated, the "rogue administrator" - while a valid concern, which the community rightfully is concerned about - presumably would be resolved rather quickly and the candidates affected restored, thus no longer making this clause applicable/causing an issue. --TheSandDoctor ^Talk 06:16, 1 September 2019 (UTC)
5. per BMK & SandDoc. –Dave | Davey2010^Talk 11:14, 1 September 2019 (UTC)
6. Kudpung กุดผึ้ง (talk) 21:45, 1 September 2019 (UTC)

Statement #2 by TonyBallioni

No changes should be made to the rules for blocked candidates. TonyBallioni (talk) 00:07, 1 September 2019 (UTC)

Users who endorse statement #2 (Blocked candidates):

1. I don't entirely oppose statement 1, but we need to make sure that a rogue admin can't just indef all the candidates and then disqualify everyone. --Rschen7754 00:10, 1 September 2019 (UTC)
2. No need for added red tape in this instance, IMHO. I'd like to think that the community is already fully capable of weeding out blocked users as unfit to serve on ArbCom through their vote. OhKayeSierra (talk) 02:01, 1 September 2019 (UTC)
3. Per OhKayeSierra. * Pppery * _{it has begun...} 03:01, 1 September 2019 (UTC)
4. Per OhKayeSierra. Unnecessary red tape and it's highly unlikely for a candidate so problematic to win a seat. feminist (talk) 09:29, 1 September 2019 (UTC)
5. I agree OhKaye, this just seems unnecessary. If effect, if a blocked candidate was somehow considered suitable as an Arbitrator, some really strange circumstance has arisen that we might want to consider. Nosebagbear (talk) 12:17, 1 September 2019 (UTC)
6. If the community has trust in a blocked user, they should be allowed to serve. StudiesWorld (talk) 12:43, 1 September 2019 (UTC)
7. Agreed with OhKaySierra and StudiesWorld. I have trust in the community to take blocks properly into account with their voting. CThomas³ (talk) 16:41, 1 September 2019 (UTC)

Comments

• Yeah, Rschen7754, I thought of that too. I considered adding a clause like "if an appeal is ongoing and Election Commission may allow the candidates name to appear on the ballot" or something of that sort, but thought simpler was better. This isn't a proposal I feel particularly strong on, but I thought it worth bringing up at the least. TonyBallioni (talk) 00:14, 1 September 2019 (UTC)

• If an appeal is granted, restoration of the right to vote can be part of it. We don't need to go hog-wild and cover every possobility. Beyond My Ken (talk) 05:25, 1 September 2019 (UTC)

Statement #N by USERNAME

Text of proposal ~~~~

Users who endorse statement #N (Blocked candidates):

1. ~~~~

Voter activity requirements

Statement #1 by xaosflux

Historically: "A voter needs 150 mainspace edits by 1 November and registered an account before 28 October".

Problem to solve: We are maintaining voter rolls of users that have long left the project.

Proposal: Change to "Eligible voters must have made 150 mainspace edits by 1 November and one live edit (in any namespace) within two years of 1 November, and have registered an account before 28 October."

— xaosflux ^Talk 00:37, 1 September 2019 (UTC)

Users who endorse statement #1 (Voter activity requirements):

1. Proposer, — xaosflux ^Talk 00:38, 1 September 2019 (UTC)

   Personally, I'm fine with 1 year or 2 years - since this was a new requirement I started it with a bit wider margin, but as far as the size of the rolls I don't think it will be a huge difference. — xaosflux ^Talk 15:13, 1 September 2019 (UTC)

2. This is very generous compared to the steward elections. Rschen7754 01:00, 1 September 2019 (UTC)
3. I'm generally a fan of having a significantly diverse pool of voters, but yeah, 1 edit in the last two years seems sensible enough. TonyBallioni (talk) 01:01, 1 September 2019 (UTC)
4. –MJL ‐Talk‐^☖ 01:19, 1 September 2019 (UTC)
5. Pharaoh of the Wizards (talk) 01:25, 1 September 2019 (UTC)
6. SQL^{Query me!} 04:53, 1 September 2019 (UTC)
7. I would prefer 1 year to 2 years, but... Beyond My Ken (talk) 05:26, 1 September 2019 (UTC)
8. Two years is more than adequate. CThomas³ (talk) 06:16, 1 September 2019 (UTC)
9. Yes, and one live edit is not even sufficient.--Ymblanter (talk) 07:46, 1 September 2019 (UTC)
10. Reasonable as a start. Second choice after Statement #3. feminist (talk) 09:31, 1 September 2019 (UTC)
11. I too would prefer a year but 2 years is better than nothing at all. –Dave | Davey2010^Talk 11:15, 1 September 2019 (UTC)
12. --qedk (t 桜 c) 11:16, 1 September 2019 (UTC)
13. I'm not opposed to moving forward the date of required registration, but I don't think it is necessary. StudiesWorld (talk) 12:44, 1 September 2019 (UTC)
14. Third choice (to statements #4 and #3). This is better than the status quo, but does not go far enough to prevent low information voters from skewing the results. One year would be a greater improvement, but I would prefer requiring a higher number of live edits (than one edit) within one year of 1 November. — Newslinger talk 13:07, 1 September 2019 (UTC)
15. Third choice (behind my second choice of statement #3, and my first choice for statement #4, where I also briefly mention a possible more radical alternative).Tlhslobus (talk) 14:25, 1 September 2019 (UTC)
16. Third choice to statement four and statement 3. Barkeep49 (talk) 16:10, 1 September 2019 (UTC)
17. Prefer 1 year , but I'd go with 2. Kudpung กุดผึ้ง (talk) 21:54, 1 September 2019 (UTC)

Comments (statement #1, voter activity requirements)

• Suggested copyedit for grammar and style: "Eligible voters must have made 150 mainspace edits by 1 November and one live edit (in any namespace) within two years of 1 November, and have registered an account before 28 October." Best, Kevin (aka L235 · t · c) 00:45, 1 September 2019 (UTC)

  @L235: OK, replaced above. — xaosflux ^Talk 00:47, 1 September 2019 (UTC)

Statement #2 by xaosflux

No changes should be made to this qualification for voters. — xaosflux ^Talk 00:37, 1 September 2019 (UTC)

Users who endorse statement #2 (Voter activity requirements):

1. Users who have been inactive for two years are unlikely to show up to vote. This change is pointless. * Pppery * _{it has begun...} 01:40, 1 September 2019 (UTC)

   @Pppery: this is primarily to reduce the overhead in managing the eligible voters list for securepoll and the subsequent step-1 filtering for the mass-messaging exercise, I don't expect it to have any outcome on the results of the election. — xaosflux ^Talk 02:01, 1 September 2019 (UTC)

Comments (statement #2, voter activity requirements)

• Question: How much of a burden will adding another qualification be in the preparation of the voter rolls? Newyorkbrad (talk) 02:16, 1 September 2019 (UTC)

Statement #3 by Calidum

Proposal: Change to "Eligible voters must have made 150 mainspace edits by 1 November and one live edit (in any namespace) within two years of 1 November, and have registered an account before 1 October."

Users who endorse statement #3 (Voter activity requirements):

1. As proposer. Limiting suffrage to accounts created before 1 October decreases the likelihood of sock puppetry. As far as I can tell, the current 28 October deadline was arbitrarily decided. Calidum 03:33, 1 September 2019 (UTC)
2. OhKayeSierra (talk) 04:45, 1 September 2019 (UTC)
3. First choice. feminist (talk) 09:32, 1 September 2019 (UTC)
4. I even say you should make it 1 September, the date of this RFC. 4nn1l2 (talk) 10:28, 1 September 2019 (UTC)
5. Second choice (to statement #4) This seems better, I'd still be happier for 1 edit within the previous 12 months, but that's a minor issue. But it's still something Nosebagbear (talk) 12:18, 1 September 2019 (UTC)
6. Second choice (to statement #4). Once again, I would prefer requiring a higher number of live edits (than one edit) within one year of 1 November. — Newslinger talk 13:14, 1 September 2019 (UTC)
7. Second choice (to my support for statement #4, where I also briefly mention a possible more radical alternative).Tlhslobus (talk) 14:20, 1 September 2019 (UTC)
8. Second choice Barkeep49 (talk) 16:10, 1 September 2019 (UTC)

Statement #4 by Newslinger

Proposal

Eligible voters must have made 150 mainspace edits by 1 November and 10 live edits (in any namespace) within one year of 1 November, and have registered an account before 1 October. — Newslinger talk 13:34, 1 September 2019 (UTC)

Rationale

10 is the number of edits required to attain autoconfirmed status, and ensures that voters are at least nominally active in the community within the past year. It also serves as a defense against sockpuppetry and compromised accounts, as some bogus editing patterns can be detected within 10 edits. This is a stricter version of Calidum's statement #3, which is in itself a stricter version of Xaosflux's statement #1.

Users who endorse statement #4 (Voter activity requirements):

1. First choice, as proposer. — Newslinger talk 15:03, 1 September 2019 (UTC)
2. My preferred choice under the current system (tho in theory I might prefer a system like the Church of England's Synods of bishops, priests and laity, or in our case Admins, Editors, and Readers, but that's another day's work, which I won't go into here anytime soon as I'm too busy with other stuff, but I'm briefly mentioning it here in case others might wish to think about it). Tlhslobus (talk) 14:14, 1 September 2019 (UTC)
3. First choice, I think a degree of activity is worthwhile, though I can't imagine we have too many individuals who'd fall afoul of it. Nosebagbear (talk) 15:06, 1 September 2019 (UTC)
4. First choice. Barkeep49 (talk) 16:09, 1 September 2019 (UTC)

Statement #N by USERNAME

Text of proposal ~~~~

Users who endorse statement #N (Voter activity requirements):

1. ~~~~

Voting system

Statement #1 by SoWhy

Proposal

Instead of supporting / opposing individual candidates, voters should be able to assign preferences to each candidates using a ranked voting system.

Rationale

Ranked voting allows voters to support multiple candidates more flexibly and allows them to specify which candidate they want to get the spot if two or more are tied. As determined when this was raised in the context of the 2018 election, mw:Extension:SecurePoll allows ranked voting as well as straight support/oppose voting to be used, so it's not a technical problem (discussion on talk page from last year). Which ranked voting system to use is a matter for another discussion, this proposal is only to gauge whether there is support for a ranked voting system at all. Regards SoWhy 09:32, 1 September 2019 (UTC)

Users who endorse statement #1 (Voting system):

1. fawiki is already using Schulze method (available on SecurePoll) for three years (2016, 2017, and 2018) and will use this method again in 2019. WMF is also using a preferential voting system for their elections. See m:Affiliate-selected Board seats/Resolution 2019 (#9). 4nn1l2 (talk) 10:19, 1 September 2019 (UTC)
2. --qedk (t 桜 c) 11:13, 1 September 2019 (UTC)
3. Ranked voting allows voters to convey their preferences more precisely. — Newslinger talk 13:58, 1 September 2019 (UTC)
4. Definitely support receiving more nuanced feedback from voters. CThomas³ (talk) 16:34, 1 September 2019 (UTC)

Comments (Statement #1 by SoWhy)

• I do not have a strong opinion on whether a preferential system is better than a direct vote, but we as a community have very little experience with preferential voting, and most users come from real life from countries not using preferential voting. If we just replace a direct voting by a preferential voting it will be extremely confusing for many voters. My suggestion would be to run a trial in parallel with a direct vote and see how it goes. (The results will not be affected by the trial). When I was still active in the Russian Wikipedia, the Schulze method for ArbCom elections was run in parallel with the actuall vote (which is there just onwiki) for quite some time, sometimes it gave different results (may be it still running now), but in the end of the day it was decided that the Schulze method is not a good alternative, and it has never been adopted as the main voting method. Without a trial, nobody would guess how it worked.--Ymblanter (talk) 11:07, 1 September 2019 (UTC)

  If you are not okay with Schulze method, there are other alternatives. My suggestion is Meek STV (phab:T117127) which is currently not supported by the SecurePoll. Regarding the community experience, it's just a simple ordering. Many online communities hold their elections using more sophisticated methods, such as Stack Overflow [1] [2] or RationalWiki [3]. 4nn1l2 (talk) 12:05, 1 September 2019 (UTC)

  My point is not that I am happy or not happy with any form of ranked voting, which is up to discussion, but that most of them are confusing for the first-time voter, and the result might be unexpected. My conclusion is all of them are too risky to run as the main method without a prior trial.--Ymblanter (talk) 12:27, 1 September 2019 (UTC)

• So, at the risk of sounding anglocentric, this is not something most English speakers will be familiar with as, generally speaking, it’s not how elections are conducted in most English-speaking regions (Yes, I’m aware Australia and New Zealand don’t use FPP.) I think the potential for people not knowing what’s going on is something that seriously needs to be considered here. TonyBallioni (talk) 12:20, 1 September 2019 (UTC)
  • It get that it's unfamiliar. But I disagree with the idea that it would be confusing. I am confident that voters are smart enough to rank things in order of preference. While this might not be common in elections, it's not uncommon in other areas, especially on the web. After all, you don't have to be an expert on the different ranked voting system to know how to assign preferences to candidates. And

1. Bob
2. Alice
3. Daniel
4. Chad

does not seem more complicated to me than

• Bob: Support
• Alice: Support
• Daniel: Neutral
• Chad: Oppose

Regards SoWhy 12:35, 1 September 2019 (UTC)

Well, this is not so easy. Imagine that I oppose Chad and also another candidate, Eve. I think Chad is just not up to the job, for example he is just coming off a one-week incivility block. He has no chances to be elected. Eve is possibly up to the job, I see some issues with her and do not want her elected, but others may think the issues are not serious enough and she might be elected anyway. This means in the ranked vote, to have better chances of my outcome achieved, I must rank Chad above Eve, even though I normally think Eve would be a better arbitrator than Chad. And there multiple issues like this, which are not obvious to first-time voters.--Ymblanter (talk) 12:45, 1 September 2019 (UTC)

@Ymblanter: It seems that the Duggan–Schwartz theorem says any fair method to choose a limited set of winners is going to be vulnerable to some sort of tactical voting. Thus all the debates over different voting methods, as people try to decide whether one method's vulnerability to different kinds of tactical voting is "worse" than another's. Anomie⚔ 13:54, 1 September 2019 (UTC)

I would prefer a system that allows voters to specify their true preferences more granularly over a system in which tactical voting strategies are more obvious. — Newslinger talk 14:47, 1 September 2019 (UTC)

I mean, I agree with you that if you think about it, it will make sense. I guess my concern is I think that a significant number of people will go “This isn’t what I’m familiar with. Close window.” I’ve always been very big on expanding the voter base for ACE, and I’m afraid this would have the opposite effect. TonyBallioni (talk) 12:58, 1 September 2019 (UTC)

I'm not sure I agree with you here, Tony. I think anyone going to the trouble of voting in the ArbCom elections will at least spend a few minutes trying to figure how out they are supposed to participate. As long as we make the ballot straightforward and easy to use I don't think we'll have a problem with chasing people away just because it's different. The ballot was new to everyone the first time.

As a first-time voter last year I actually felt somewhat constrained in my ability to provide useful feedback beyond "yes, no, and my opinion won't count." Having some kind of order preference provides another level of feedback regarding strength of support or oppose. I do share Barkeep's concerns that force-ranking of all candidates suppresses our ability to require a minimum of support and potentially leave a slot empty if needed, but I think we can work through that issue. CThomas³ (talk) 16:19, 1 September 2019 (UTC)

Can secure poll allow you to only rank as many candidates as you choose? If so any candidates omitted could be in effect "oppose". But I don't know the practical implications of doing that - does it impose problems for closing the election. Best, Barkeep49 (talk) 16:22, 1 September 2019 (UTC)

@Barkeep49: Yes, it is possible to rank only some candidates when voting with SecurePoll. And I'm telling you that as someone who has administered several elections on Votewiki [4]. 4nn1l2 (talk) 16:43, 1 September 2019 (UTC)

That's the sort of thing I am hoping we can do, yes. But I also don't know the practical implications, so I fell short of proposing a solution. CThomas³ (talk) 16:43, 1 September 2019 (UTC)

@Barkeep49: The current code for the ranked choice ballot accepts rankings from 1–999, and (if "Require that all options be ranked." is not set on the election) interprets any unranked options as being assigned rank 1000 (i.e. tied for last place). Also it allows tie rankings and doesn't require that numbers used are sequential. Anomie⚔ 18:04, 1 September 2019 (UTC)

• What about using a combination where we still have support/neutral/oppose for each candidate and one must meet specific minima, but the actual ordering of candidates is by some form of preferential voting? StudiesWorld (talk) 12:47, 1 September 2019 (UTC)
  • We probably shouldn't try to invent our own voting method, there are a lot of different concerns that go into "fairness" that take some pretty rigorous study/mathematical proofs to determine. Anomie⚔ 13:00, 1 September 2019 (UTC)

    I'm not saying invent anything complicated, just a sanity check to avoid anything caused by unfamiliarity, such as the situation Tony described. My thought would be: Each candidate is voted on as support/oppose/neutral and ranked. Any candidate who did not meet certain percentages for S/O/N is thrown out and the remaining candidates are selected from according to rank. I've done enough with voting systems to know that this could have some unforeseen consequences, but it seems better than running two independent votes. StudiesWorld (talk) 13:07, 1 September 2019 (UTC)

    Rather than having separate ballots with an arbitrary percentage cutoff for approval then ignoring the number of appoves in favor of ranked votes to fill the seats, we might get similar results by just including "None of the above" as an option on the ranked ballot. If the rankings end up as Carol, Alice, None of the above, Dave, Eve, Bob, then only Carol and Alice would be elected and the rest of the seats would remain vacant. I think that'd be equivalent to everyone voting "support" for people they rank above "None of the above", neutral for anyone they rank as a tie with "None of the above", and oppose for the rest. Anomie⚔ 14:45, 1 September 2019 (UTC)

• One item of potential concern here would be the ability to hold a second RFC to determine the specific method to be used and then to have code for that method written and deployed in time for the election. After investigating, I note it's possible to configure SecurePoll for an election using the Schulze method but then ignore the Schulze method results in favor of an analysis of the pairwise victory matrix (for methods using that concept) or a dump of the ballot rankings themselves using the chosen method's procedures instead. Anomie⚔ 15:41, 1 September 2019 (UTC)
• In general I'm a fan of ranked voting systems. However, I'm a bit nervous of the practical implications here. It seems better, for instance, to have an arb slot empty than to fill it with someone who would have, under the current system, only received 49.9% or less net support. This is especially true because it's likely we'll be electing a large number of arbs this year. Best, Barkeep49 (talk) 16:07, 1 September 2019 (UTC)
• Noting that a form of preferred voting (a modified STV) was used for the recent Affiliate-selected WMF board seats. There was significant confusion about the system to the point that several affiliates had to "re-vote"; it was also not nearly as conclusive as one might have hoped, in that it would have taken only one single change in order on a single ballot to have rendered a different result. The Schulze method is included on SecurePoll as an historical artifact; it was used for some BoT elections in the 2000s, but a review of the system when I joined the Election Committee identified that the "wrong" Schulze method was linked to SecurePoll (it is the one designed to give a single winner rather than multiple winners), and it has a very significant weakness in that leaving a candidate unranked entirely has a different impact than ranking a candidate with a number far, far below the number of candidates. As well, do take a look at the history of the article; it's "maintained" by Schulze himself as one of the most obvious COIs I've ever seen on Wikipedia. While I could be convinced that some sort of STV might be a reasonable choice, this is NOT the system I would choose, and I think it would be wrong to select a system where the most obvious source of information is a COI article on our own project. If we want to consider some for of STV or other ranking voting system, it will take at least a year to investigate systems, come up with a decision, and arrange for the software to be written/tested and ready to go. I'd go so far as to say that this is something that should be done in concert with a number of other projects to come to a joint decision. Risker (talk) 16:25, 1 September 2019 (UTC)

  This is all really helpful and presents good reasons for not changing our method. However, I would like to point out that the proposed process for actually doing so isn't practical - you're talking about a huge investment of time from some number of editors, across multiple projects, for maybe the chance of implementing. If that is the only way it could happen, then it won't happen. The best I've come up with for addressing this is a non-binding commitment this year to implement ranked choice next year as at least that would show some community support and provide some momentum behind the efforts. Best, Barkeep49 (talk) 16:35, 1 September 2019 (UTC)

  Risker, I do share your concerns with selecting a voting method, and your timing concerns. I am treating my "support" vote above as more of "if practical for 2019, but for sure in future elections." My concern is if we vote this down solely because we don't think we can do it in time, we'll just have the same timing problem next year when the same issue arises. Perhaps we do need to set this aside for now, but I'd like to at least formally set the process in motion here somehow. CThomas³ (talk) 16:39, 1 September 2019 (UTC)

  Well, it is specifically the 2019 RFC. I cannot see how you can secure endorsement for this for 2020 in the 2019 RFC. I suggest a full, considered, thought through RFC in January. That gives you 9 months. Leaky caldron (talk) 16:47, 1 September 2019 (UTC)

  Yes, but what we decide here is clearly not only for the 2019 elections. Our decisions here set how we conduct all future elections until such time as we change them again. CThomas³ (talk) 16:51, 1 September 2019 (UTC)

  Except in January everyone is tired of ArbCom elections. In essence because rules carry over from year to year it would be possible to enact the language now and absent a consensus to change it, it would go in effect in 2020. It would be no guarantee but would also provide some impetus for the work to be done to actually getting done. Best, Barkeep49 (talk) 17:01, 1 September 2019 (UTC)

  What Leaky caldron said. Risker (talk) 16:49, 1 September 2019 (UTC)

  Please note that there were only 122 voters in m:Affiliate-selected Board seats/2019/Results, so it is not strange that only one vote could have changed the results. For English Wikipedia ArbCom Election, there will be many more voters (2,118 last year). Please note that any single-winner voting method can be used for a multi-winner election; you only need to identify the first winner, then remove him/her, and tally the results again to determine the second winner, and so on. Yes, the results may not be proportional, but this is also the case for the current system. "It has a very significant weakness in that leaving a candidate unranked entirely has a different impact than ranking a candidate with a number far, far below the number of candidates". The cardinal numbers are not important at all; the only important factor is ordinal numbers. The candidates who are not included in the voter's list will be placed at the lowest rank possible. I do not call it a weakness. 4nn1l2 (talk) 17:46, 1 September 2019 (UTC)

  @Risker: I agree that it implements the single-winner Schulze method rather than something like Schulze STV designed for proportional representation in multi-winner elections. I'm curious about it has a very significant weakness in that leaving a candidate unranked entirely has a different impact than ranking a candidate with a number far, far below the number of candidates. Looking at the code, I see nothing in the code that would treat leaving options unranked any differently from giving them all a rank of 999. The code seems to do almost exactly that (it uses 1000 instead of 999, while only 1–999 are available to the ballot itself). I'm also curious what a potential COI with respect to the Wikipedia article has to do with the validity of the voting method, particularly when the COI does not seem to be at all hidden. Anomie⚔ 18:33, 1 September 2019 (UTC)

  It has to do with the Condorcet pairings. First off, only someone who carefully researches the system would know that it accepts rankings up to 999, and treats no ranking as 1000. Any *ranked* candidate gets compared to any other *ranked* candidate, so what ranking a voter uses has an impact. Many of our users who have experienced some sort of STV/ranking electoral system are used to being required to rank either ALL candidates or at minimum a certain number of candidates; it's how the most frequently-used systems work. In the Schulze system, the best way to "oppose" a candidate is to leave them entirely unranked (= a ranking of 1000/1000 candidates); however, most people will tend to "rank" their most unwanted candidate with a number that shows some sort of disdain. We saw this in at least one past board election using Schulze, where candidates who should have been at the very bottom were artificially elevated because people assigned them a rank instead of leaving them unranked. Incidentally, one of the reasons that the board election committee abandoned Schulze was that there were complaints that voters had no way of saying "this person is entirely inappropriate for the role" (i.e., opposing a candidate). I'm not sure this community is ready to give up that opportunity, given that our only other "voting" process (RFA/B) is completely designed around S/N/O. Risker (talk) 18:49, 1 September 2019 (UTC)

  Are you saying that people somehow assumed that ranking someone at 999 would put them below unranked candidates? Probably despite actual documentation to the contrary? And that's a fundamental failing in the polling method rather than the people voting? I have to disagree. If someone really wants to vote 999 to indicate disdain despite that no one will ever see the vote, they should be sure to rank everyone else above 999. Anomie⚔ 19:00, 1 September 2019 (UTC)

  I don't want to pollute this RfC with a bunch of mathematical theory, but the reality is that the methodology and its impacts have never been well described, not even by the creator of the system. Previous board election committees have variously stated things like "if you rank a bunch of candidates as 1,2,3,16, the system will interpret that as 1,2,3,4", which is true if there are only 4 candidates, but should be false if there are more than 4 candidates; the problem is, the people responsible for the system do not seem to know whether or not it's true or false. A non-mathematician, non-psephologist should be able to easily understand what different permutations can do to the vote; it's not possible for this system. It's also very unclear what really happens if one ranks two candidates at the same number, because the "success" point is winning the matched pair, not tying it. There are other, better systems if one absolutely *has* to have something other than S/N/O. In order to keep this to a single post, I'll reply to the "none of the above" point below. Keep in mind that voters today can oppose every candidate, or only some candidates; "none of the above" isn't quite where we would want to go, unless there are going to be as many "none of the above" slots as there are candidates. They could probably be labeled "Not Candidate A", "Not Candidate B" etc and then allowed to be ranked the same as other votes. Oh wait. "Not Candidate A" could potentially "win" the election, so that isn't going to be optimal either. Risker (talk) 21:40, 1 September 2019 (UTC)

  One option for being able to indicate an actual "oppose" would be to include "None of the above" as an option to be ranked. If the results end up as Carol, Alice, None of the above, Dave, Eve, Bob, then only Carol and Alice would be elected and the rest of the seats would remain vacant. Anomie⚔ 19:00, 1 September 2019 (UTC)

  RfA/Bs are not elections; they are evaluations. An election is a competition in which multiple candidates vie for a limited number of seats. I have never seen a "real" election with oppose votes. 4nn1l2 (talk) 19:05, 1 September 2019 (UTC)

  Well then, this isn't really an election, either. While there would be a limited number of seats available, only candidates that meet certain support criteria would be seated, so it's not about the top-8 candidates filling 8 empty seats. It's the top-8 candidates who obtain a minimum level of support filling those seats; in that way, it's a lot closer to RFA/B (which also requires a minimum level of support) than most run of the mill elections. Under the current system, it is entirely possible to have Arbcom start off the year with empty seats because an insufficient number of candidates achieved 50% community support. I think this is a good thing, because I don't want someone that only 30% of the community thought would be a good arb filling the seat. They're not political posts, they're work assignments - just like RFA/B. Risker (talk) 00:50, 2 September 2019 (UTC)

Statement #2 by Nosebagbear

No changes should be made to the style of voting system used. Nosebagbear (talk) 12:24, 1 September 2019 (UTC)

Users who endorse statement #2 (Voting System):

1. Due to the confusion/lack of need (see comments for details) Nosebagbear (talk) 12:24, 1 September 2019 (UTC)
2. Per my comments above. Giving the cultural/national backgrounds of most en.wiki editors, this would result in confusion. TonyBallioni (talk) 12:28, 1 September 2019 (UTC)
3. Calidum 14:16, 1 September 2019 (UTC)
4. There is not sufficient time between the close of this RFC and the start of the election to investigate, select, and develop/test the software for a new voting system. See my comments above with respect to the serious deficiencies of the Schulze method that is currently available on SecurePoll. Risker (talk) 16:30, 1 September 2019 (UTC)
5. This, of all years, is not the time to be messing around with voting methods. We have had issues with the mechanics of the election at least once in the last few years and I think the important thing for 2019 is to keep selection simple. Leaky caldron (talk) 16:37, 1 September 2019 (UTC)
6. Per Risker and Leaky caldron. Newyorkbrad (talk) 19:05, 1 September 2019 (UTC)
7. If it ain't broke, don't fix it. –Dave | Davey2010^Talk 23:41, 1 September 2019 (UTC)

Comments (Statement #2 by Nosebagbear)

• I get the reasoning used, but while non-FPP systems are fairly understandable for 1 candidate elections, they can get very confusing for multi-candidate (electing multiple individuals) setups. I feel there'd be significant confusion and I'm not sure how accurate a trial run would be for giving a consideration on how it would be - I suspect only those comfortable with it would use it. Nosebagbear (talk) 12:24, 1 September 2019 (UTC)

  I don't get it. Ranked voting only makes sense with multi-candidate setups. If you only have two candidates, you cannot use ranked voting. Regards SoWhy 12:36, 1 September 2019 (UTC)

  I believe Nosebagbear is distinguishing between elections to select a single candidate versus more than one. Newyorkbrad (talk) 13:18, 1 September 2019 (UTC)

  I was, yes, thank you Nosebagbear (talk) 14:00, 1 September 2019 (UTC)

  The current system (support/oppose/neutral) already deviates significantly from the voting systems used in most political elections, and the suggested voting strategy in Wikipedia:5-minute guide to ArbCom elections § Voting process ("The influence your ballot has on the results of the elections is maximized when you select Support or Oppose for every candidate, and Support approximately the same number of candidates as there are vacancies.") is unintuitive. I am unconvinced by the argument that ranked voting would be even more confusing. — Newslinger talk 14:39, 1 September 2019 (UTC)

  The ballot itself is pretty simple: "Rank each candidate in order of your preference, with 1 being the most preferred." The current preferential ballot code in SecurePoll allows tied rankings, allows omitting rankings (treating them as ties for last place), and doesn't care if any values are skipped.

  The complication is all in explaining how it's determined who wins, as the procedures and formulas can get pretty complex. But I wonder how many people even care about the current S/(S+O) formula, rather than just trusting that the votes will be counted in some fair manner. Anomie⚔ 15:09, 1 September 2019 (UTC)

Statement #3 by Anomie

No change to the voting method for 2019, as there's not time to implement a change. An RFC should be held in February 2020 to choose a new method, giving time to ask the developers to implement and test the new method in time for the December 2020 election. Anomie⚔ 19:18, 1 September 2019 (UTC)

Users who endorse statement #3 (Voting system):

1. The current system is not even an electoral system. Where have you seen an election where you can cast oppose votes? 4nn1l2 (talk) 19:28, 1 September 2019 (UTC)

Comments (Statement #3 by Anomie)

• Proposing this since some people above seem to say they'd support the change if there were more time for implementation. Picking February since it has been said "in January everyone is tired of ArbCom elections". Anomie⚔ 19:18, 1 September 2019 (UTC)
  • I do not think it is within the scope of this RfC to bind us to having a future RfC. If someone wants to propose it in February, go for it, but right now it’s just extra noise. TonyBallioni (talk) 19:24, 1 September 2019 (UTC)
  • I don't necessarily agree that it's out of scope for this RfC to decide to hold another RfC, but there are much more concrete actions that can be taken to make the February RfC more useful: workshopping proposals to be better explained and justified, doing (or finding and linking) analyses of the pros/cons, setting up a list of people who want to be notified when the RfC comes, etc. I'd be happy to work with anyone on setup for any February proposal on this. Best, Kevin (aka L235 · t · c) 19:28, 1 September 2019 (UTC)
• I don't think it's necessary to have an RfC reaching consensus to have an RfC in February. Let's just have the RfC in February. (Incidentally, it's what I proposed last year: having a discussion about electoral methods in the new year after the election.) isaacl (talk) 19:38, 1 September 2019 (UTC)
• @TonyBallioni and Isaacl: I'd like to avoid people trying to claim in February that there wasn't support for changing the voting method because Statement #2 here had consensus for the "no time" reason. Whether this statement is actually binding or just advisory doesn't make much difference IMO. Anomie⚔ 19:47, 1 September 2019 (UTC)
• I'm in favour of having a proper discussion about the topic (provided there is the option that the community could just say "we like it as it is now" rather than being expected to choose between alternatives that most of them don't want). Having said that, even on Meta when there has been an effort to have these discussions, it's wound up with extremely low participation, and I'd be surprised if the same isn't true on this project. This may be a question of "how much effort does the community want to invest in redesigning a process that is unlikely to result in a significant difference in outcome" - I'd lay odds most of the people participating on this page could have successfully identified 80-90% of the successful candidates during the "fallow" period before voting began going back to when SecurePoll was first used in 2009 or 2010. I don't think it's a good idea to plan an RFC of this type so far in advance, unless there is a dedicated group of interested users who will do the research and prepare excellent, reader-friendly summaries of various voting systems as part of the development of the RFC. I'd suggest adding a "minimum number of participants in the RfC" proviso to ensure that any change truly has broad support. Risker (talk) 21:57, 1 September 2019 (UTC)

Statement #N by USERNAME

Text of proposal ~~~~

Users who endorse statement #N (Voting system):

1. ~~~~

Scope of Election Commission

Statement #1 by Barkeep49

Proposal

Repeal "Commission should only intervene when there is a problem that needs resolving, and either discussion isn't working, the rules are unclear, or there isn't time" from the 2013 RfC

Rationale

This was enacted by only three users, so it never had a large mandate. When issues arose last year the commission had to decide if it was appropriate for them to intervene for one of the three allowed reasons. This would give them slightly more lattitude to decide when and why to intervene without changing the commission's charge to "solve disputes and problems during the election"

Users who endorse Statement #1 (Election Commission)

1. As proposer. Barkeep49 (talk) 16:55, 1 September 2019 (UTC)
2. There are a lot of other reasons that the commission may need to intervene, such as (just an example) some sort of software problem. Risker (talk) 17:29, 1 September 2019 (UTC)
3. TonyBallioni (talk) 17:33, 1 September 2019 (UTC)
4. sometimes a bit of ambiguity is necessary. - Frood (talk!) 21:06, 1 September 2019 (UTC)

Comments (scope of electoral commission)

In the original 2012 request for comments, the mandate of the commission was defined as to deal with unforeseen problems, adjudicate disputes, and as Jimbo continues to shift his role, to ceremonially announce the final results. While I agree the commission should have flexibility to step in as seems best, I think it would also be desirable for the commission to avoid curtailing community decision-making unnecessarily. isaacl (talk) 18:04, 1 September 2019 (UTC)

Statement #2 Barkeep49

No change be made to the scope of the Election Commission

Mass message

Statement #1 by KTC

In additional to current rule of eligible voters who have edited within the last 12 months before nominations, mass message not to go out to accounts with blocks that extend beyond the end of the election, globally locked accounts, bot flagged accounts, accounts in Category:All Wikipedia bots, Category:Wikipedia alternative accounts, Category:Wikipedia doppelganger accounts, and Category:Deceased Wikipedians. KTC (talk) 18:34, 1 September 2019 (UTC)

Users who endorse statement #1 (Mass message):

1. This is a good change. I don't feel super strongly about time-limited blocks or not, so I'd also support statement #2 as my second choice, but I figure that it's fairly confusing for users who aren't able to vote (and won't be unless unblocked) to get the invitation to vote. Kevin (aka L235 · t · c) 19:32, 1 September 2019 (UTC)
2. * Pppery * _{it has begun...} 23:29, 1 September 2019 (UTC)
3. –Dave | Davey2010^Talk 23:43, 1 September 2019 (UTC)

Statement #2 by KTC

Per statement #1 but DO send mass message to those with time limited blocks that extend beyond the end of the election. KTC (talk) 18:34, 1 September 2019 (UTC)

Comments (mass message, statement #2)

As blocks can be appealed and lifted, it may be preferable to continue to deliver notifications to blocked users who are not site banned. isaacl (talk) 18:40, 1 September 2019 (UTC)

Statement #3 by KTC

No change to current rule. KTC (talk) 18:34, 1 September 2019 (UTC)

Users who endorse statement #3 (Mass message):

1. Don’t mind the first proposal, but I’m not a fan of adding complications to the list generation. TonyBallioni (talk) 20:01, 1 September 2019 (UTC)

Statement #4 by KTC

Send message out to every accounts that are technically eligible. KTC (talk) 18:34, 1 September 2019 (UTC)

Election start date

Statement #1 by KTC

Extend nominations period by one day so election don't start on a weekend (US time) in case of technical problems. KTC (talk) 18:34, 1 September 2019 (UTC)

Statement #2 by KTC

Extend fallow period by one day so election don't start on a weekend (US time) in case of technical problems. KTC (talk) 18:34, 1 September 2019 (UTC)

Statement #3 by KTC

No change to current rule. KTC (talk) 18:34, 1 September 2019 (UTC)

Users who endorse statement #3 (Election timetable):

1. TonyBallioni (talk) 19:59, 1 September 2019 (UTC)

Comments (Election start date)

• The current schedule has the nominations running November 10–19, the fallow period being November 20–24, and the election beginning Monday, November 25. That's not a weekend. Anomie⚔ 18:42, 1 September 2019 (UTC)

  @Anomie: 00:01 GMT Monday is still weekend in the USA where most of the technical people + many WMF staffs are. The schedule you're quoting is the same as last year, where the election was delayed due to technical problems and fixes took longer than it could had done because the day of week the start of the election was on. -- KTC (talk) 18:46, 1 September 2019 (UTC)

  Do you have a link to discussion last year, e.g. the Phabricator tasks? Anomie⚔ 19:08, 1 September 2019 (UTC)

  phab:T209802. The election was postponed by the EC at about 01:00 GMT. -- KTC (talk) 19:15, 1 September 2019 (UTC)

  I suppose having the election start at 00:00 UTC on a Tuesday would give more time if the issue and its urgency were communicated only a day or two before the election. But in that case it looks like phab:T209656 was filed three days before (22:49 on Thursday, leaving all Friday for work), but the urgency was not communicated, while phab:T209802 communicating the urgency wasn't filed until after the election was supposed to have started. I see no way that having that happen at 00:06 UTC on a Tuesday rather than a Monday would have made a difference there. Anomie⚔ 20:03, 1 September 2019 (UTC)

Withdrawn/disqualified candidates

Statement #1 by KTC

List withdrawn or disqualified candidates in an appropriate section on the candidates page. KTC (talk) 18:34, 1 September 2019 (UTC)

Statement #2 by KTC

Do not show withdrawn or disqualified candidates on the candidates page. KTC (talk) 18:34, 1 September 2019 (UTC)

Users who endorse statement #2 (Withdrawn or disqualified candidates):

1. TonyBallioni (talk) 19:58, 1 September 2019 (UTC)