Wikipedia talk:Requests for comment/2010 ArbCom election voting procedure

Comment on the opinion from A Horse called Man

Before recommending that the Schulze system be used, please demonstrate confirmation from a suitably qualified developer that there is such an extension available, tested, and ready to use on the current configuration of this project. It is my understanding that such does not exist. It is not appropriate to expect that such be developed, tested, and ready to use a month before the election for which it is intended. If this was a priority consideration on the part of the community, it would have been discussed and actioned months ago; the RFCs for SecurePoll last year occurred several months before the election occurred, in plenty of time for discussion within the community. Risker (talk) 22:12, 24 October 2010 (UTC)

I'm confused—if the Schulze method has not been implemented in MediaWiki, how could we use it for the board election? Ucucha 22:15, 24 October 2010 (UTC)

The configuration that was used, as I recall, feeds to an outside source; there were a lot of peculiarities in order to ensure that there could be no tampering within the WMF family. It's my understanding that it is not trunk deployed. Risker (talk) 22:22, 24 October 2010 (UTC)

Thanks for the clarification. Ucucha 22:25, 24 October 2010 (UTC)

FWIW, most preferential voting schemes can be implemented by hand. Since this election is usually fairly small, it's not impractical to calculate the vote without a custom program, perhaps using a simple spreadsheet. However, even that would require creating protocols and procedures, and it's a bit late to develop those.   Will Beback  talk  22:28, 24 October 2010 (UTC)

Unfortunately, such problems will always be solved at the last moment. In July/August 2009, we had this RfC long before the next ArbCom elections. However, the result of that RfC didn't lead to any reaction of the developers. A Horse called Man 22:48, 24 October 2010 (UTC)

Erm, 8 people in that RFC explicitly supported Schulze method. Not anywhere near consensus. However, there were enough people in support of moving to a SecurePoll method that the developers certainly responded to that. Risker (talk) 22:59, 24 October 2010 (UTC)

I don't see any contradiction in the following set of facts

• I believe we should use a preference voting system
• No software is currently available to implement such a method

— Carl (CBM · talk) 12:54, 25 October 2010 (UTC)

I would just like to re-iterate Risker's point above about the difficulty in getting the necessary developer involvement for most of the changes proposed overleaf; a few weeks into inquiries and we still do not have a dev lined up to put us on the board, never mind to grant us the wishlist or repurpose the election system. Skomorokh 11:28, 26 October 2010 (UTC)

On the other hand, if we don't express a desire for the Schulze method now, while the issue is live, simply because it's not yet available, we'll be in exactly the same position next year. If there's clear support demonstrated in this RfC, then it can be used to show the devs that it's something the community wants next year, giving 12+ months for it to happen. --RexxS (talk) 03:37, 28 October 2010 (UTC)

Moved from main page

• I'm not really sure if YellowMonkey's comments are intended to say this is a good idea, or to say this is a bad idea but is similar to other, extant, bad ideas, like the UN Security Council. I unfortuntely acknowledge that there are comparatively few individuals who have the time or inclination to involve themselves in ArbCom matters which don't directly interest or affect them, and actually kind of like that. But I oppose this idea specifically because it would allow "diversity" from groups which could then register every person they know and all the animals at the local humane societies as editors to vote for them in an Arbitration election. Do not want. John Carter (talk) 17:16, 27 October 2010 (UTC)
  • Any system is vulnerable to that sort of thing, this is about minimising the worstcase scenario. With the system we usually use an organised lobby could potentially get all the Arbcom places up in one year, with STV you effectively guarantee that such a faction cannot sweep the board. I appreciate this is putting in a protection against something that has not yet happened here, but I think we all know that such groups exist, and I think it unwise to leave ourselves with a vulnerable system. ϢereSpielChequers 09:33, 3 November 2010 (UTC)

RFC question

I duplicated the "question" at the top to fix a formatting mess on the central RFC page. The bot picks up everything between the tag and the first date stamp, which in this case included the question, the background, the purpose, the reminder, and the first view. WhatamIdoing (talk) 23:05, 24 October 2010 (UTC)

Comment on Angus McLellan's statement

What responsibility are people freed from with a secret ballot? Surely plenty of people made uninformed votes with the open election system, so there's no responsibility to make informed decisions. If an arb isn't very good, will the people who voted for him be held responsible? As for block voting, is that actually prohibited in AC elections? I was under the impression that, since unlike every other vote-like-pseudo-discussion on Wikipedia, AC elections are an actual, purely numeric-based vote, nothing other than outright sockpuppetry/ballot stuffing was actually prohibited - any eligible voter is allowed to vote however they feel. For the transparency argument, I would ask the same question I do any time someone proposes something for transparency: "Why? What do we gain by transparency here?" Transparency is not always a good thing, so there should be reasons other than for the sake of transparency. Why do we need to know how other people voted? Mr.Z-man 03:21, 25 October 2010 (UTC)

Placing a high priority on transparency is a value judgement and reasonable people can disagree with that judgement. However, our processes in general tend to favour transparency. We very strongly discourage the use of undisclosed alternative accounts in process areas so as to maintain transparency as well as to avoid vote-stacking. We can only judge people on their record and how they vote forms part of that record.

The responsibility I refer to is "taking the election seriously". Even with the groupthink pressures found in an open vote, some editors may not do so. This should be distinguished from thinking outside the box, which is to be encouraged. This'll be another value judgement thing. Angus McLellan (Talk) 12:09, 25 October 2010 (UTC)

I strongly believe that we should not be using editors' votes on elections as part of a public record with which to make judgments about the editors. Votes are a matter of personal opinion and conscience, and the arbcom election is not about forming consensus. — Carl (CBM · talk) 12:57, 25 October 2010 (UTC)

There's a difference between putting a high priority on transparency and being transparent for no reason except the sake of transparency. I believe governments should be transparent, but I recognize that certain things have no good reason to be public and could be used for malicious purposes if they were. Mr.Z-man 21:39, 25 October 2010 (UTC)

Exactly: transparency in government is quite different from transparency in the voting booth. In the Soviet Union and East Germany, governance was opaque but the voting boxes were made of glass—they made sure they knew how you voted. We need the opposite in modern times: transparent governance (excepting, of course, the odd bit of day-to-day management where there are compelling reasons for privacy protection), and the ability to vote without suspicion of public obligation or fear. I'll keep my preferences private, thanks. Tony (talk) 13:42, 26 October 2010 (UTC)

Transparency is a good thing because the entire community knows the same thing and every editor can see why other editors voted the way they did (if they choose to share their reasons). The reality is that wikipedia is a largely unstructured place and editors here have only a secondary responsibility to the encyclopedia (work, family, school, presumably, are where our primary responsibilities lie). It is impossible for any editor to research even one candidate properly, let alone all of them. Instead we look for external cues and rely on the research that others have done (which is the collaborative wiki way of building the encyclopedia) and make our decisions based partly on our own research and partly on this 'others research'). This works very well, for example, in the selection of admins, where diffs provided in favor and against individual candidates are used as cues by others. All this disappears when there is a secret ballot. One could turn your question around and ask why should we not be transparent? The only reason I can think of is that arbcom members would then know who voted against them and, possibly, act in a vengeful way. However, a strength of the transparent approach is that everyone knows who voted against particular members and the community would take note of any bias against these no voters. Transparency ensures that no information is private and, as we should all know, only private information can be exploited for personal gain. Better to ensure that this never happens. --RegentsPark (talk) 13:42, 26 October 2010 (UTC)

Even in the open elections, we actively discouraged people from discussing on the voting pages. There's a fine line between people acting on others' research and people acting on what others' claim to be "research." One issue was that some people would just vote with the majority (either because they were too lazy to do any research, or didn't want to appear "different"), with the end result that the people who voted in the first few hours had more of an effect than the later voters. As for bias against voters, there's also the issue of the appearance of bias. If an arb votes for a sanction against someone who voted against them, is it because they're biased, or because that's actually the right decision? There's no way to know. With secret ballots, we don't have to take note of bias against voters, because there can't be any. Mr.Z-man 21:01, 26 October 2010 (UTC)

• I think we may be looking at this with the wrong framework. Most wikipedia processes do not (and should not) look like democratic processes where a large body politic votes for proxy decisions. When we talk about transparency in governance and transparency at the voting booth we are in some sense assuming that the two are mutually exclusive. ACE is probably the closest we will get to a voting arrangement, but editors who vote there are also non-trivially involved in project governance. They will participate in arbcom cases in a variety of ways, they will also shape policy outside of arbcom. I voted in support of secret ballots this year (and opposed last year), but we can't ignore the social expectation of transparency in wikipedia. To wit, our major optics problem with arbcom is not that votes for arbcom members are not protected and private; it is that arbcom is a secretive and non-democratic body. Adding more secrecy, even in the voting booth, adds to this. Protonk (talk) 19:07, 26 October 2010 (UTC)
  • I don't think that saying that ArbCom is "secretive" is anywhere near the truth. Much of what we do is "behind the scenes", as it were, but almost none of it is not transparently declared and discussed (the exception being, obviously, privacy matters and the small subset of matters which have serious legal liability potential). But Arbcom isn't a democratic process, nor should it be. Pretty much by definition, we're the last recourse when consensus-based processes failed or are not applicable; while we are responsive to the community we should not be directed by it (because that would just move the problem to a different venue).

    On the other hand, the method by which arbitrators are selected is exactly a democratic process. It may not be any more ideal than any real world democracy, but that's not an excuse to make it any worse either. — Coren ^(talk) 19:43, 26 October 2010 (UTC)

    • Declaring that you have an optics problem with respect to secrecy isn't the same thing as saying you are secretive. I think you guys do yeoman's work making private conversations as transparent as possible, but that doesn't mean people have forgotten the old arbcom or that people feel the current pressure for transparency counter-balances the fundamentally private nature of the committee. Protonk (talk) 19:46, 26 October 2010 (UTC)
      • Oh, I don't disagree with the principle; transparency is generally beneficial — I'm just very much convinced that the votes isn't a good place to apply it. The process of using SecretPoll is very transparent: in the end, the only thing that is not made public is who voted for what. And the argument that avoiding even the appearance of favoritism (or vindictiveness) of candidates towards people who voted one way or another is, to me, compelling. — Coren ^(talk) 20:06, 26 October 2010 (UTC)
        • I could extend the desire to avoid recrimination to a process where openness is key: XfD. I would love to be able to promise editors that their 'votes' in AfD would not be used against them in later discussions (RfA is only the most obvious and odious example), but I suspect that the community would reject a proposal to move AfDs to a secret ballot system for a variety of reasons. My point isn't to bring up a spurious comparison but to argue that mapping the wikipedia community to a demos is spotty. Protonk (talk) 20:24, 26 October 2010 (UTC)

(undent) It is (spotty), but XfD and RfX are bad examples in that they are pathological; they are attempting to be a discussion to reach consensus while using a vote-like structure but attempt to pretend they aren't polls. I've always though they should go one way or the other, but not try (with limited success) to sit on the fence. At least, nobody's trying to pretend ArbCom elections aren't a vote. — Coren ^(talk) 20:40, 26 October 2010 (UTC)

• Yeah, but I brought them up because we can actually show real interaction between XfD and RfA. Candidates have been judged on their "voting" record in AfD when they ask for advanced permissions, precisely the worry we have for open ACE elections. There are possibly dozens of reasons why this is acceptable to the community, not least of all the differences in scale, scope and nature. I think we understand each other here, and I suspect that if this election runs without a serious hitch we will transition to open voting secret balloting for ACE permanently. Protonk (talk) 20:47, 26 October 2010 (UTC)

Comment on the opinion from Fetchcomms

I don't understand it. Could somebody explain it/clarify as I'm not sure I could vote for something that has so many unanswered technical questions. Ta. GedUK  09:26, 25 October 2010 (UTC)

Are you referring to the fact that you don't know what the schulze method is? If so, basically it's a system where you rank the candidates instead of voting "yes" or "no". This allows you to put preference towards individuals and, furthermore, that the "more preferred" candidates always win. I won't try to explain all of it because the article does a much better job of it, but I think that's it in a nutshell. --Shirik (Questions or Comments?) 12:42, 25 October 2010 (UTC)

My ideal voting method is where we use the SecurePoll interface, but the interface uses the Schulze method rather than just "support" "oppose" "neutral". The interface would also publicly display which users have voted, and each user's votes/rankings. I think this would address the issue of vote stacking, election organization, and transparency. However, I am 99.9% sure the devs are not able to quickly modify the SecurePoll extension to do this (nor is there any discussion on such a change), so in the end, I think we should save some time and just use SecurePoll, which seems to have worked fine in the past. (One might disagree because of the CU/OS elections this year, but that's why I like the Schule method mixed in). /ƒETCHCOMMS/ 16:52, 25 October 2010 (UTC)

Thanks. GedUK  07:28, 26 October 2010 (UTC)

Suggestion

I endorsed the retention of the SecurePoll, but I have an additional suggestion or comment, and I am not sure whether to set it up as an additional comment on the RFC page (so people could endorse it), or just make it here, or what. The comment is this: Last year, at least one candidate "withdrew", and was removed from the list of candidates. I believe the same candidate also asked for his/her totals as of that time, which I don't think was done. I also recall that two or three years ago (when the voting was "open"), a candidate withdrew, but then a day or two later came back, so he/she was removed from the list of candidates and then restored. I think this is disruptive -- not in the punishable Wikipedia-way, but in the "bad election practices" way. Once the first vote is cast, there should be NO changes to the ballot. (Note that in "real" elections, even death does not get you off the ballot, and dead people have been elected.) If a person wants to stop "campaigning", fine. If they want to tell people not to vote for them, fine. But they should stay on the ballot. I got the impression in the past that the people who withdrew figured out that they were going to get a very poor approval percentage, and did not want that known publicly -- but I'm sorry, that's how elections work. Once the election is underway, the ballot should be stable. So, how and where do I make this suggestion so people can actually consider it? Neutron (talk) 15:17, 25 October 2010 (UTC)

Write it short (to avoid confusing people) and place it on the main RFC page - I suggest "To prevent balloting disruption, candidates should not be permitted to withdraw after voting begins. Candidates who withdraw will remain on the ballot." Hipocrite (talk) 15:20, 25 October 2010 (UTC)

Note that last year, the candidate who withdrew during the election wasn't actually removed from the ballot, their name was struck on the list with a note that they withdrew; people could still vote for them though. Mr.Z-man 21:49, 25 October 2010 (UTC)

Some countries may have a tradition of electing dead people, other countries allow candidates to withdraw, and restart elections if a candidate dies. I can see the argument in an old fashioned paper based election that you can't withdraw after the ballot paper has been finalised and sent to the printers - but if we are doing this electronically I don't think that argument applies. ϢereSpielChequers 22:18, 25 October 2010 (UTC)

Reply to Mr.Z-man: Are you sure about that? My recollection is the opposite. That is one of the reasons I made this suggestion. In any event, it is now on the RFC page so people can support it or not support it. Neutron (talk) 22:43, 25 October 2010 (UTC)

I'm certain it's written in this year that once voting starts, candidates can't withdraw. Tony (talk) 01:55, 26 October 2010 (UTC)

Where is it "written in"? I looked for discussion of the issue, but could not find any, much less a resolution. Can you point me to where it is resolved? Neutron (talk) 03:17, 26 October 2010 (UTC)

Why would we want to keep a volunteer in an election once they decide they want to withdraw? I can understand the argument that in a paper election you can't physically withdraw a candidate once the ballot papers have gone to print, and that in a political election allowing a candidate to withdraw disenfranchises their party's supporters. But I really don't see how either argument applies online or in this election. ϢereSpielChequers 10:51, 26 October 2010 (UTC)

I believe the issue is that the candidate can't be removed from secure poll once it starts DC T•C 11:26, 26 October 2010 (UTC)

As best as I remember, the candidate who withdrew last year remained on the ballot but the scrutineers did not include them in the published results. If this is correct, I'm not sure how this suggestion differs from the status quo... Skomorokh 11:31, 26 October 2010 (UTC)

It is becoming increasingly probable that my recollection of what happened last year was slightly off the mark, and I can't find the "evidence" either way. But I am leaving my proposal on the RfC page, as there seems to be support for clarifying this to the extent it requires clarification. No harm done. Neutron (talk) 16:04, 26 October 2010 (UTC)

That's fair enough. Anyway, should this RfC conclude with significant net support for your proposal, we'll incorporate it into the instructions for candidates, so at the very least next year's organisers will have a paper trail! Best, Skomorokh 18:59, 26 October 2010 (UTC)

The evidence you are looking for might be in the notes section right under the chart here. It mentions that secret withdrew, and the chart itself does not note statistics for him. Seeing as how in 2009 it worked out just fine, I don't think we really have to change anything, what you are suggesting became de facto policy already. Either way, as Skomorokh said, we should leave it there to gather support, and leave 2011 a paper trail. Sven Manguard _Talk 01:46, 27 October 2010 (UTC)

Comment on the opinion from Hipocrite

While I appreciate the reasoning in allowing voters the ability to more fully express their preferences, I am curious as to how proponents of a "none of the above" option would want it implemented, particularly using last year's setup.

Candidate 1 (S N O)
Candidate 2 (S N O)
Candidate 3 (S N O)
None of the above (?) or None of the above (S N O)

If there is just one option for "none of the above", it does not express anything more than opposing all candidates. If voters are able to support some candidates and "none of the above", it renders the results incoherent. In short, we already allow voters to express that they support all/some/none of the candidates, so what would this add? Skomorokh 11:39, 26 October 2010 (UTC)

Any candidate that fails to reach the level (either in raw amount of support or support/oppose ratio) that NOTA generates should not be eligible for seating. Hipocrite (talk) 13:16, 26 October 2010 (UTC)

Ah, very interesting, thanks. Skomorokh 13:24, 26 October 2010 (UTC)

This makes NOTA a useful bar indeed, in that the community, by its votes, sets the bar for taking the seat. It makes use of NOTA in a multicandiate election a very viable choice... I like this idea better and better. ++Lar: t/c 20:59, 26 October 2010 (UTC)

This is from Wikipedia:Arbitration Committee Elections December 2009 (footnote 1):

"Unlike previous years, all voters were required to register some form of vote for each candidate. The "neutral" column is simply the total votes, minus supports, minus opposes. There is no effective difference between "neutral" and "abstain" as far as this election is concerned."

Theoretically, if someone voted straight neutral, that would serve the same function, without requiring additional technical features be built in. Would that satisfy you Hipocrite? Sven Manguard _Talk 01:37, 27 October 2010 (UTC)

Can't speak for H but to me it's not the same thing at all. NOTA is a candidate like the others. Not an abstention. ++Lar: t/c 03:50, 27 October 2010 (UTC)

So I'm still a bit fuzzy - would clicking NOTA set all candidates to Oppose? Or would it just leave them neutral? UltraExactZZ ^Said ~ Did 12:13, 27 October 2010 (UTC)

Given candidate A with support/oppose ratio SA(a) and raw support S(a), and candidate NOTA with SA(o), S(o), candidate A is only eligible for seating if SA(a)>SA(o) and S(a)>S(o). IE, any candidate who loses to "None of the Above" in any way is not seated. NOTA is treated as is any other candidate, except losing to NOTA makes you ineligible to sit for that year. You are permitted to S/O/N any and all candidates, including NOTA. The effect of various voting scenarios is left as an exercise for the reader. An alternative is to provide a statement that candidates with a S/O ratio under X or with less than Y total opinionated votes will not be seated regardless of any other factors. Hipocrite (talk) 16:00, 27 October 2010 (UTC)

Then one likely scenario could be that almost everybody would support their (around 10) preferred candidates and NOTA. Given a reasonable spread of voter preferences for 20+ candidates, the obvious result of that would be that NOTA would have greater support than any candidate, and nobody would be seated. The problem is that NOTA is unlike a normal candidate, because support for NOTA is not analogous to support for a candidate, and the purpose of opposing NOTA is completely undefined. It needs to be thought through again. --RexxS (talk) 04:18, 28 October 2010 (UTC)

• "None of the above" is a backdoor way to eliminate the Arbitration Committee, and should not be permitted. If the community, in its collective wisdom, wishes to deactivate the Arbitration Committee, it should be done in a straightforward community discussion specific to this issue. If individuals wish to oppose all candidates, they should simply vote "oppose" for all candidates. Risker (talk) 15:31, 27 October 2010 (UTC)

No, it's not. It's a way of ensuring that candidates who are less favorable than seating no one, and candidates who are little known are not seated. Because the selection rules currently permit seating people with ratios under 50%, or only one person supporting, voting oppose is not enough without the requirement of a quorum of votes and a minimum support principal.Hipocrite (talk) 16:00, 27 October 2010 (UTC)

That's not what you are proposing, though. Instead of this, which actually does have the effect of voting to demob the committee, perhaps you should propose minimum criteria for appointment. I have added a statement myself that identifies the two points I consider bare minimum for appointment - majority support and actual candidacy - and you may wish to identify other minimum criteria, such as minimum number of voters commenting on a candidate. Risker (talk) 16:10, 27 October 2010 (UTC)

You assume a specific outcome, which was not my intent, nor my expectation - do you actually think ArbCom is so unpopular that if a secret ballot were held to disband or retain you, you would be disbanded? Further, I note your initial statement - "If individuals wish to oppose all candidates, they should simply vote "oppose" for all candidates," would not work without enforced minimum requirements. A vote to oppose-all in all previous years was identical to a vote to support all, or a vote to neutral all, or just not voting. The only previous option to say "All of these are not acceptable," was to fork. Hipocrite (talk) 16:40, 27 October 2010 (UTC)

Risker: It is simple: If you can't beat NOTA's totals, perhaps you shouldn't sit. That's the point of NOTA. Claiming that it would "back-door eliminate arbcom"" is not really a fair assessment of NOTA at all. It front-door eliminates candidates who don't have widespread support and don't have a good percentage. It's exactly the same as a fixed percentage favorable and a fixed total vote favorable requirement, except that it ebbs and flows with election participation and candidate quality, and thus is more dynamic. If we set that you have to get 500 positive votes to get seated, what happens in an election where 400 people vote total? no one gets seated. what about in an election where 20,000 vote? everyone easily passes that test. So fixed numbers are not a good and flexible test. ++Lar: t/c 16:53, 27 October 2010 (UTC)

(PS afer ec) Hip has it exactly right. ++Lar: t/c 16:54, 27 October 2010 (UTC)

(e-c) If done correctly, I think that this, if implemented, might be a good idea. The problems lie in implementation. I think historically we have always had at least a few editors who have voted exclusively for their friends and people they like individually, without necessarily reviewing and considering the candidates with whom they are less familiar. That was, is, and I think always will be a bad thing to do. However, there is another kind of potential voter, one who, having reviewed all the candidates, finds many or most of them individually actively unacceptable, and votes only for those comparatively few who they do find acceptable.

My own choice, such as it might be, might be something like this: Allow NOTA to be used as an "early cut-off" by voters who are in the second group, who find most of the candidates unacceptable. So, as an example, I might possibly vote for Risker, Kirill, Roger, NYB, Carcharoth, and NOTA. (They're the first names that came up, don't assume anything from their use, OK?) By doing so, I'm saying that those individuals named are the only ones I find acceptable. In a situation with eight seats to be filled, one vote for each of the named candidates would be recorded, and one vote for NOTA, regardless of how many positions aren't filled in that ballot. The only problems I can really foresee arise if there are so many NOTA votes, and so few regular votes, that the NOTA votes bring the number of candidates who get appear on over 50% of the votes cast is not sufficient to fill all the open seats. In that event, maybe the seat could be left empty for that election cycle, and a "special election" to fill the remainder of the term take place during the next election.

As I remember, last year, we basically let Jimbo review the votes and ultimately announce the winners, based on the votes cast. I could see, maybe, if this were implemented, allowing that to continue for the first election or two using this system, until all the bugs are worked out. But I do think that, in some instances, it might be an acceptable idea. John Carter (talk) 16:05, 27 October 2010 (UTC)

That doesn't actually work, John. There is no method by which the system could determine which "NOTA"s an individual voter's NOTA vote applied to. Risker (talk) 16:10, 27 October 2010 (UTC)

I think it might work, if it were used in the way indicated above. What would be counted would be the number of votes for an individual candidate, with the stipultion that, for instance, they get support on 50% or more of the ballots cast; NOTA is not an individual candidate.

This is a fine differentiation from "I voted only for my two buddies - I don't care about the rest." This is saying, in effect, "I looked at everybody, and actively think these are the only acceptable candidates." I acknowledge the number of instances in which that differentiation would come into play probably isn't that many, but I can envision a few, unlikely, circumstances when for whatever reason a "party" here wants to get candidates to vote their way, and many other editors don't want that "party" getting in power. I don't necessarily see it happening anytime soon, but I've been wrong about that sort of thing before. John Carter (talk) 16:23, 27 October 2010 (UTC)

In response to Hipocrite, Lar and John: What may be intended and what the actual effect is are two different things. Essentially, what Hipocrite intends is a flexible level of support below which no candidate could be appointed. A "None of the Above" vote is a vote for leaving ALL seats unfilled. These are not the same thing. Indeed, it seems you now want to support some candidates and also vote "none of the above", which doesn't make a lot of sense. Oppose the candidates you don't want to be arbitrators. It is not reasonable, however, to have a statute-bound bar to appointment that has the potential to change with every single voter. Risker (talk) 18:04, 27 October 2010 (UTC)

I was going to support this, but in practice, having a moving bar with NOTA seems to be basically the same thing as having a set bar with people just opposing every candidate if they don't like any. The latter would just bring the support% down rather than raising the passing%. Mr.Z-man 15:57, 30 October 2010 (UTC)

Moved from main page. "None of the above" just means no supports or opposes. It is, technically, precisely the same as voting Neutral for all candidates. Please, it would be embarrassing to ask our hard-pressed tech at WMF to muck around doing the programming (and the testing) for such a change, when there are so many better things we could get them to do but feel we can't ask at this late stage, such as creating a show-your-previous-vote display if you want to change your vote. I really find the idea of allowing nihilists an official platform in this way, when it will have no effect on the tally, just untenable. Tony (talk) 13:48, 30 October 2010 (UTC)

Problem with the reordering

The reordering helps quite a bit, but I'm confused why the "Statement by CBM" ended up in the "hybrid" part since it advocates exactly for SecretPoll, only with a different rationale? — Coren ^(talk) 00:49, 27 October 2010 (UTC)

I'll look it over and fix it if it needs fixing. Thanks for pointing it out. Sven Manguard _Talk 01:02, 27 October 2010 (UTC)

While it does say "some sort of secret ballot" rather than explicitly supporting Secure Poll, the statement doesn't make any specific criticism of Secure Poll or give an alternative method of secret balloting. I suppose supporters of this statement are rather agreeing with the stated criticisms of open polling. So I also think it belongs in the support section. Diderot's dreams (talk) 01:06, 27 October 2010 (UTC)

(edit conflict) Okay, I see what I did. CBM's comment could have gone in support, but I left it there because it was also a direct response to roux's comment. While it is in favor of secret ballots, I believe based on the timing and placement of the statement, that it was primarily a counter argument to releasing the votes after the election. We can always move it, but for the moment I think it makes the most sense where it is. Sven Manguard _Talk 01:08, 27 October 2010 (UTC)

I agree, CBM's comment is entirely for secret ballots (i.e. SecurePoll). Od Mishehu's statement is also essentially for secret ballots, just with a technical modification to the extension that doesn't really affect the secrecy; its not really a hybrid solution as there's no aspects of public voting involved in it. Mr.Z-man 01:27, 27 October 2010 (UTC)

I've fixed it. Tony (talk) 01:35, 27 October 2010 (UTC)

Right, there goes that idea. It's alright, I do not have the power to read people's minds over the internet, so if consensus is that CBM's comment is not a response to roux's comment, then by all means, move it. I did add an explicit oppose section to roux though, just in case I was right and people want to double back. Sven Manguard _Talk 01:40, 27 October 2010 (UTC)

history

Should probably noted that open voting was originaly only used because we couldn't get acess to the closed voting softwear that year.©Geni 01:05, 27 October 2010 (UTC)

That's actually worth noting; I always wondered why the switch took place, and trying to dig that far back in the archives for a discussion you have no idea where it took place is an exercise best left to someone with Copious Free Time. You don't happen to remember where that was decided, do you? :-) — Coren ^(talk) 01:10, 27 October 2010 (UTC)

I saw the other day that Tim Starling, the developer, helped with the very first one back in 2004. Tony (talk) 01:21, 27 October 2010 (UTC)

Exactly the where the switch took place is messy. There is also a bunch of arcane wikipoliticing going on which means that the archives can look a little odd if you don't know what is going on. It started here. Fortunately jimbo proved rather apathetic which meant that other than his latter attempt to take control here he didn't gain much traction. The vote was then delayed as a result of debate stating here which lead to this change which is basicaly what we went with. A few tweaks were made here. If you want to get a more complete feel of events you probably need to read the whole of that talk page.©Geni 02:35, 27 October 2010 (UTC)

Comment on statement by roux

moved from main page

Users who oppose this summary

1. Although Piotrus's idea I could live with, an immediate release of the information defeats half the purpose of the secret ballot. Sven Manguard _Talk 01:31, 27 October 2010 (UTC)

Should just have explicit parliamentary factionalism

We may as well have open factionalism, eg, political parties to support certain racial/religious POV pushing groups, mates' parties, social clubs, parties for writers, deletionists, hat-collectors etc.... The arbcases that get attention are the ones that involve the famous warlords and factional chiefs of wikipedia, and in those cases the side with more political power wins, or more likely, it is a draw as both sides have some power and would punish any arb that votes against them if they want re-election. When the dispute involves two nobodies, it proceeds quickly without anyone getting too fussed. Let's face it, people only get interested in cases, from both sides of the fence, if they have a horse in the race, except for the usual chunk of admins who just like joining in any drama. So we may as well have interest-group style representation openly, with no hypocrisy, fake "outsider" interest, recusals etc... people are only interested if they are stakeholders. The current system favours people who pretend to be independent when they are not and don't recuse and only leads to situations like UN Security Council vetoes from involved warlords/parties' main ally....otherwise they are perpetually snoring YellowMonkey (new photo poll) 04:21, 27 October 2010 (UTC)

Again, I'm really confused as to what you are saying. I think you're being either sarcastic or are reflecting a deep loss of faith/trust in ArbCom. No one is forcing you to participate in the election, indeed no one is forcing you to participate in Wikipedia at all.

I'm not saying you should leave Wikipedia, but I would go as far as to ask you not to use the discussion space for the election procedure as a soapbox for an Anti-Arbcom essay. You are bordering on being pointy. Sven Manguard _Talk 04:45, 27 October 2010 (UTC)

I'm not sure I understand YellowMonkey, but I think he makes an argument for open voting. With open voting, any factions are obvious. --SmokeyJoe (talk) 05:46, 27 October 2010 (UTC)

Are they obvious? Only if they're all stupid enough to vote exactly the same way.  Roger ^talk 06:05, 27 October 2010 (UTC)

When a faction behaves as a faction, and everything is open, it is obvious. I don't believe there is organised factionalism going on now. Secret voting for arbs opens a door for hidden factionalism. --SmokeyJoe (talk) 06:26, 27 October 2010 (UTC)

Yes, except only the most incomptent faction would do that. So, perhaps more accurately, open voting facilitates accusations of factionalism. And, contrariwise, SecurePoll robs factions of the ability to strategically vote to tip the balance (because the results aren't known until after the poll closes). In fact, the solution to both these concerns is to get more voters out. The more votes that are cast, the less the impact of a faction.  Roger ^talk 06:37, 27 October 2010 (UTC)

Don't much agree with the first part. Accusations stem most easily from a lack of information. But I think you miss my main point. A faction may include a very large sock farm. Many accounts, named to look like real people, with faked personal details, all old, intermittent, very capable editors, all kinda similar in style, perfectly behaved, working in their own areas of mainspace, almost no interest in wikipolitics, nothing to link them except they all voted the same way. They know how to fool checkusers. Maybe they are a checkuser. No one would ever join the dots if the voting is secret. No senior wikipedian and arb candidate could ever do such a thing, surely not! --SmokeyJoe (talk) 07:05, 27 October 2010 (UTC)

SecurePoll has sock detection features built into it, which the scrutineers can see.  Roger ^talk 07:09, 27 October 2010 (UTC)

Is this confidence supported by technically competent and independent people? --SmokeyJoe (talk) 09:46, 27 October 2010 (UTC)

What confidence? It's a statement of fact.  Roger ^talk 10:23, 27 October 2010 (UTC)

Sorry. I read an implied confidence that the sock detection features were effective. Are there independent and technically competent people who will say with confidence that the sock detection features cannot be avoided by technically advance sock farm. --SmokeyJoe (talk) 10:48, 27 October 2010 (UTC)

I'm not sure what your point is here. The number of socks required to significantly change an election result is large. And the greater the turnout, the higher the number of socks needed to influence it. However, as the tipping point is unknown, because the votes are unknown until after the election, the sockmaster has no idea whether to throw ten, twenty, fifty or a hundred socks into the ballot. It is not easy to create this many, and build up the requisite number of qualifying edit etc without detection. In any case, voters lists are available throughout and they can be scrutinised by whoever is interested in looking for behavioural similitaries.  Roger ^talk 11:32, 27 October 2010 (UTC)

The second part. Absolutely, get more people to vote. Last time I found it hard to get abreast of the candidates. The process included a lot of assumed knowledge as to what was going on. The nomination period becoming the voting period was confusing. As I thinkI said somewhere, I think it would be very helpful if every candidate were introduced by a senior wikipedian (not necessarily all by the same person). Self introductions can be very stilted and incomplete. --SmokeyJoe (talk) 07:05, 27 October 2010 (UTC)

Sven Manguard: I don't know who you are, never seen you before. It's a big wiki... it happens. I wonder if you know who Yellow Monkey is? Sometimes it's worth looking into things first before rushing to judgment. Given that he's an ex-arbitrator, who also happens to have a long and distinguished career of service to the project in other areas, I'd tend to let him have his say rather than bandying about accusations of pointy-ness. I may not agree totally with his view, (if he actually meant it (Swift didn't mean A Modest Proposal either), which I doubt) but factionalism IS a big problem on this project. And it's not easily spotted. ++Lar: t/c 11:25, 27 October 2010 (UTC)

Honestly, I've been around in Wikipedia for a few years as an IP, but only as a passive content editor. If I saw a comma missing, I put in the comma. The whole administrator notice board and ArbCom and WikiProjects and AfDs, and all that stuff that most people only get into if they have accounts, I was only peripherally aware of before I created my account a month ago. If what you say is true, than I thank him for his service. However I still think that he could have simply said "XXX voting system would curb factionalism, which would reduce the drama in ArbCom." Instead, we a long, confusing, paragraph (I honestly can't tell if it's serious) which was cut and pasted into multiple different sections, then posted here. Maybe wp:pointy was a bit of an overstatement, but the comment itself was mildly disruptive, IMO. I in no way meant to indicate that Monkey shouldn't participate in Wikipedia, or even at this RfC, just that his posting an essay in the RfC multiple times isn't constructive or helpful. Sven Manguard _Talk 16:36, 27 October 2010 (UTC)

• YellowMonkey is correct that factionalism is a problem, albeit one that exists whether the balloting is public or private. I'm not convinced by the argument that it is, somehow, easier to detect it when the votes are visible to the public — most people supported or opposed with little or no written rationale and it no more proper to declare that X voted like Y because they are part of one "faction" than it is to guess at it from the privacy of the (virtual) voting booth.

  One problem that can be caused by open ballots is when one early voter makes a completely false allegation as oppose rationale and that allegation cascades into a number of spurious oppositions because the allegation was public and in direct view of anyone who then voted (and people were unable to verify it, or did not bother; or were influenced even if they did). This gives a disproportionately large influence on the first couple of voters who will lead the bandwagon effect and can completely fuck up results. Even if voter privacy and the attendant appearance of COI weren't issues, public ballots are a platform for campaigning of the worst kind and that would be reason enough to avoid them. — Coren ^(talk) 17:19, 27 October 2010 (UTC)

• Just to add a bit to what Coren said, that statement might be particularly true if one of the first few voters may have been somewhat recruited by a candidate a few months before the election, or even a direct sockpuppet of possibly even another candidate, for the specific purpose of slandering another candidate in the election. This could be a particular problem if there is some reason to believe that there might be an ArbCom case coming up that deals with a clearly defined group, like, maybe, some of the previous Scientology arbitrations. John Carter (talk) 20:28, 27 October 2010 (UTC)

Statement by SmokeyJoe

moved from main page

This comment is almost identical to "Statement by Angus McLellan."
SmokeyJoe: please consider merging your vote into "Statement by Angus McLellan."

No, not the same. Angus does not make the point that securepoll creates a vulnerability to carefully sockpuppetry. Only the checkusers can know, and if unconvinced their privacy policy prevents them from letting the community know about their suspicions. --SmokeyJoe (talk) 03:34, 27 October 2010 (UTC)

Suit yourself. I added the tag not because I disagree, but because it makes determining consensus later easier. For the record, there were only just over 900 voters in the 2009 ArbCom election, so a good sock could have an impact, if they managed to qualify their socks, which might be a bit difficult, and they knew to vote. In other words, I doubt the casual vandal would find this and I doubt the hardcore sockmaster could string an account along long enough to qualify. Sven Manguard _Talk 03:43, 27 October 2010 (UTC)

Secret voting is so commonplace that nobody usually even bothers to question it. They forget that secret voting requires positive identification guaranteeing one vote per eligible person. A number of further comments are precluded by WP:BEANS. --SmokeyJoe (talk) 05:26, 27 October 2010 (UTC)

Could one person voting in favour of secret ballot please address this point? Anthony (talk) 08:18, 1 November 2010 (UTC)

It's already been addressed in the section immediately above.  Roger ^talk 08:30, 1 November 2010 (UTC)

Thanks, Roger. So, a sock farm of 400 or 500 geographically disparate registered users with different but fixed IP addresses would be just as effective in either a secret or open ballot? Anthony (talk) 08:59, 1 November 2010 (UTC)

If socks avoid detection from CheckUsers (those who have access to the ballot and are engaging in the scrutinizing), the only way that is possibly different from having them detected on an open ballot (i.e. an RFA, but CUs as far as I know don't readily scrutinize that) is if they successfully avoid detection and another insightful user (probably a non-CU) sees something going on. I personally think, with the 150-edit mainspace minimum requirement, pulling something off like this (i.e. socking on this scale) would be monumentally difficult; even if you lowered that minimum to, say, 50, that would still be difficult. Besides, in either case, if there was socking on such a large scale, open or closed ballot, you got a tainted election – you might as well throw all the results out at that point and start all over. –MuZemike 09:21, 1 November 2010 (UTC)

Are you saying that, unlike a secret ballot, an open ballot allows for insightful users to notice something fishy? I think that was SmokeyJoe's point, and why I support an open ballot. Or have I misunderstood you? Anthony (talk) 09:36, 1 November 2010 (UTC)

This is duplicating the thread above. With SecurePoll, people can peruse the contributions of public voting lists (updated in real time) for oddities, plus there are built-in security features.  Roger ^talk 11:13, 1 November 2010 (UTC)

I would appreciate your patience. I had read the above thread before I commented in this one. I read it again after you pointed me to it, but it doesn't seem to refute the proposition that in an open ballot anybody can see who voted for whom, and examine the edit histories of the voters, and possibly sniff out suspect behaviour that evades CU. I couldn't grasp MuseMike's response to my request for clarification. He seems to be saying that in an open ballot insightful readers may detect funny business that CU missed - but since I didn't really understand what he wrote I've asked him for clarification. I realise this must be tedius, but I am genuinely just groping for understanding here.

What I don't understand about your response, Roger, is

With SecurePoll, people can peruse the contributions of public voting lists (updated in real time) for oddities ...there are built-in security features.

What people? What are "public voting lists"? What are the "built-in security features"? Anthony (talk) 16:31, 1 November 2010 (UTC)

there are built-in security features?

This page, which is the pages linked from the RFC and which I would assume is tha authoritative page, says (of substance, in its entirety!):

Special page extension for elections, polls and surveys. Currently incomplete, mainly due to the lack of an election setup interface. However it has been used for Wikimedia Foundation Board elections, arbitration committee elections and the Wikimedia license transition vote.

and this leaves me unimpressed. The best answer (mostly from Coren) against open voting seems to be that the lack of drama that comes from a lack of information is a good thing. I expect that a dedicated and skilled Wikipedian could feed a large sock farm though the process. Other checks on the arb setup would be these:

• There needs to be an unsuitable recipient arb candidate, and these are presumably vetted well;
• Arb com is a majority rule several-membered body and an aberrant arb would have little effect;
• the powers of the arbs are really quite limited, limited to cases and editors. They would have little effect in forcing something inappropriate into mainspace. A single arb would have none. --SmokeyJoe (talk) 22:13, 1 November 2010 (UTC)

In the 2009 election, there were 996 voters. To bring the 10th place candidate into 9th place would have required at least 13 sockpuppets, which would take nearly 2000 mainspace edits. That was an extremely close split, 2.3%. To do the same with 11th and 9th would have required 18 socks; in each case, the candidate would have still had the support of the majority of the community. To elect the 15th place candidate (the highest placed candidate to get <50% support) would have required 78 sockpuppets (nearly 12000 edits). Of course, there's no way to know until the end how many socks would be necessary. This is complicated further by the requirement that the edits be made by November 1, while candidates cannot nominate themselves until November 14, so you won't actually know who the competition is when you need to prepare the socks. Mr.Z-man 15:25, 2 November 2010 (UTC)

That's very comforting. Thank you for that Mr.Z-man. Anthony (talk) 18:38, 2 November 2010 (UTC)

As Mr Z-man points out, a serious effort prepared ahead would be required. However, we have seen an editor make hundreds of edits per minute without using a bot. There are web sites advertising software for facilitating socking, hiding your IP, location, computer. How many old user accounts are there? It still looks possible for a sock farm to push, or sink, a candidate. I would feel better, if, contrary to what Coren says "even the scrutineers are unable to see what the votes are", if scrutineers were able to see how individual voters voted. A few hundred apolitical, relatively inactive, old vandal fighting accounts voting may not raise suspicions, but if they all voted similarly, a closer look might be a good idea. --SmokeyJoe (talk) 12:22, 3 November 2010 (UTC)

Websites that allow you to change your IP are generally banned here. DC T•C 12:30, 3 November 2010 (UTC)

You mean Wikipedia:Open proxies? --SmokeyJoe (talk) 12:50, 3 November 2010 (UTC)

"hundreds of edits per minute" without a bot seems unlikely - that's several edits per second. As for inactive editors, it should be possible to prevent them from voting as well. We did that in the WMF board elections; users were required to have at least 50 edits in the 6 months prior to the elections, as well as 600 edits total. Raising the edit count threshold even slightly would also have the effect of significantly increasing the amount of work required to prepare socks. Raising the threshold by 50 edits would not be a major hurdle for a single editor, but its an extra 1000 edits for someone making 20 socks. Mr.Z-man 22:04, 3 November 2010 (UTC)

Such an edit rate would imply an automated method, and official bots aren't the only way.

Ambivalent about excluding inactive accounts. I think it is probably desirable for once very active Wikipedians to maintain some involvement in the project, even if they do now have a full time job and a new family. (above I was imagining sock accounts created long ago)

I think 150 mainspace edits is reasonable and I don't mean to suggest altering this. Raise it too high and the process is elitist. I understand that people want to have secret votes. What do you think of the suggestion that scrutineers should be able to see and discuss, in-confidence, how people voted? As the system stands, it looks like if socks were detected, it would be unknowable who they were voting for. --SmokeyJoe (talk) 01:53, 4 November 2010 (UTC)

Someone making a hundred edits in a minute is going to attract attention. 20 people doing it is probably well beyond the point where WP:AGF stops and WP:DUCK begins, and if a user is indef blocked for running an unapproved bot, they can't vote. I'm somewhat hesitant to support allowing the scrutineers to see actual votes. On one hand it reduces the chances of ballot stuffing, but it also means we're putting even more trust in the scrutineers. There's also the question of whether some people would not vote as a result of slightly decreased secrecy. Perhaps if there were some additional security measures, such as a requirement that multiple scrutineers must agree to see a vote, with a publicly available log of who viewed which votes. Mr.Z-man 02:54, 4 November 2010 (UTC)

Statement by CS55cp

Sven Manguard wrote: "This comment is almost identical to Statement by Angus McLellan. CS55cp: please consider merging your vote into Statement by Angus McLellan." My statement is not about the open voting (statement by Angus McLellan) vs. secret ballots (statement by Will Beback) controversy. My statement is about the Support/Neutral/Oppose voting vs. preferential voting controversy. Support/Neutral/Oppose voting can be used both with open voting and with secret ballots. CS55cp (talk) 05:27, 27 October 2010 (UTC)

Okay, apparently I misunderstood you. I see you removed the comment, rephrased the statement, and moved the section. I approve of your boldly fixing my mess. In my defense, I was trying to streamline what I saw as four very similar opinions. I suppose that as there are now only two nearly identical opinions, and two clear branches, I succeeded , all be it not in the way I had envisioned. Thanks for making me aware of my mistake and taking the necessary steps to remedy it. Sven Manguard _Talk 05:40, 27 October 2010 (UTC)

Response to Sven Manguard's comment

I wanted to start a separate RfC (ASAP) to determine that part as I intended here. I think doing this at this specific RfC may make things too much confusing, but I left it open as to where we should discuss such issues. I would like to start that RfC within the next 24 hours so that as many people can get comments in before 14 November, when nominations open (so we're not sandbagging the nominees). –MuZemike 05:43, 27 October 2010 (UTC)

Separate, or here, either way. But those are good questions. ++Lar: t/c 11:26, 27 October 2010 (UTC)

Ambiguity and Re: Statement by CS55cp

I think that if someone with more knowledge could make a clear and concise statement as to what system is currently (barring the current discussions) set to be the method of voting, it would clear things up.

It looks to me like last year each voter voted Support, Neutral, or Oppose for each candidate, rather than voting to support one candidate. Is that correct? Is that style what is being planned on for this upcoming election? It seems like some people think it is going to be a "pick one from the list" type vote.

I think that if we cleared this up now, it would clarify a lot of things in the RfC itself.

Sven Manguard _Talk 06:46, 27 October 2010 (UTC)

You get a list of all candidates and you are offered tick boxes for Support / Neutral / Oppose for each of them. You vote for as many or as few as you like. Is this really an issue?  Roger ^talk 06:59, 27 October 2010 (UTC)

I'm a very orderly person, and I might be seeing confusion and ambiguity that might not exist, but as I said above, it seems like some people think it is going to be a "pick one from the list" type vote. Either way, we should always strive for clarity, especially in areas this important. Sven Manguard _Talk 16:01, 27 October 2010 (UTC)

What would be useful (if it doesn't already exist) is a screenshot of a typical voting screen, so people know what to expect. With additions like red circles indicating what to click and so on. Maybe even a video clip showing someone voting to show what needs to be clicked, though that might be tending a bit too far towards being patronising! It might be needed, though, if the idea to massively ramp up participation with user talk page notices gets consensus. And those voting for the first time may be confused by what they see on the voting screen. Carcharoth (talk) 06:11, 30 October 2010 (UTC)

An annotated screenshot is an excellent idea, preferably with a clear explanation that you can vote for as many or as few candidates as you wish, and that you can revisit the voting page while voting is open to amend votes. Since the majority of currently active editors did not vote last year, it ought to go some way towards dispelling some of the reluctance of first-time voters to participate. Any ideas on how to get a typical screenshot ahead of time? --RexxS (talk) 12:21, 30 October 2010 (UTC)

I believe that there are many users who don't know that Support/Neutral/Oppose voting is currently used for the ArbCom elections. Otherwise, there were more users who supported my statement. I doubt that Schulze is preferred to Support/Neutral/Oppose voting by a near consensus of 11:1. CS55cp (talk) 14:02, 30 October 2010 (UTC)

Whenever a largish number of seats are to be filled from a large pool of candidates by vote, support/neutral/oppose voting does not discriminate strongly. A more sensitive method such as a rank order (with equal ranks possible) allows the voter more flexibility in their decisions, and can produce a more nuanced result. Schulze is only one of several methods of analysing rank voting, but has several advantages over competing methods. I don't find it odd that the editors interested enough to comment here might strongly prefer it. --RexxS (talk) 19:07, 30 October 2010 (UTC)

(←) I will screenshot my voting screens during the election (without the votes though) and upload it for this purpose. We can annotate from there. An excellent idea. Sven Manguard _Talk 20:43, 30 October 2010 (UTC)

This is a vote

While the header says that the purpose is "to determine is there is a rough consensus," past experience says this really means "this is a binding up-or-down vote" and that even a narrow margin will be interpreted as "consensus." Let's be honest here. Short Brigade Harvester Boris (talk) 13:00, 27 October 2010 (UTC)

As a reminder, a consensus you disagree with — however strongly — is still a consensus. And qualifying 4:1 support as a "narrow margin" is disingenuous at best. — Coren ^(talk) 14:54, 27 October 2010 (UTC)

I did not say that I would dispute the "consensus", however arrived at. I am only seeking clarification that this is, in fact, a vote. But you've deflected the issue and slipped in a borderline personal attack while doing so. One might hope that as an arbitrator you would model the restraint and good faith that you demand from others. One would apparently be mistaken. Short Brigade Harvester Boris (talk) 16:09, 27 October 2010 (UTC)

There is no attack intended, explicit or implicit. I do believe that you are letting your own personal distaste for private ballots skew your evaluation of the process to the point of attacking its legitimacy rather than acknowledge that you are holding a position not shared by as many editors as you think should support it. This is understandable, if erroneous, and no attack on your person to note it. — Coren ^(talk) 17:09, 27 October 2010 (UTC)

Sigh. I'm perfectly willing to let the issue be decided by an up-or-down vote, and even on a narrow margin. All I'm asking is that we be clear up front what we're really doing. I think I'm a fairly good judge of what I mean to say, but you seem to feel you know better than me what my intent and motivations are ("you are letting your own personal distaste for private ballots skew your evaluation of the process" etc). Short Brigade Harvester Boris (talk) 20:43, 27 October 2010 (UTC)

Some, including myself, use "consensus" to mean unanimous accord (or nearly unanimous accord). Perhaps, if we used the less ambiguous "majority" (or "plurality") instead of "consensus" then it would be clearer how this decision process works. —Quantling (talk) 15:24, 27 October 2010 (UTC)

Consensus is the ideal we are shooting for, but it's no more obtainable here than it is at any particularly bad AfD nomination, or at ANI, ect. We should keep the word consensus, because when we combine the votes and the comments at the various RfCs, we will reach something that approxomates consensus. At the very least, a narrow one. Sven Manguard _Talk 15:55, 27 October 2010 (UTC)

This is kind of hinging on some a priori assumptions, but votes are fine where the scale demands them, the choices are clear and discrete, and the outcomes are not expected to hinge on strong arguments which experience clash across positions. This poll meets those criteria pretty neatly. If the poll grinds down to <60% support for one overall option (private versus public voting), we can talk about scrapping the poll and moving toward a more directed discussion of the pros and cons (unlike another infamous poll taken and re-taken this year). If current trends continue, we will not have to do that in order to get a rough gauge of support for private voting. Protonk (talk) 18:17, 27 October 2010 (UTC)

Precisely. The question is well suited to an up-or-down vote, and I think it should be a vote, so that the outcome will be clearly defined. But we're calling it something else. Short Brigade Harvester Boris (talk) 20:45, 27 October 2010 (UTC)

In my experience asking for community discussions to call a spade a spade is a mugs game. A little cognitive dissonance is unavoidable. Protonk (talk) 20:56, 27 October 2010 (UTC)

I don't see why any more than a majority vote would be required, either way. It is not like either an open ballot or a secret ballot are the long-established way that all votes are taken on Wikipedia. From what I have read, for a couple of years the ArbCom ballot was secret, then for a couple of years it was open, last year it was secret. Votes for other positions (including WMF board) have sometimes been by open ballot, sometimes by secret ballot. In other words, there is really no "default" position. There are, effectively, two choices: Open or secret. Whichever gets more votes wins. Why make it more complicated than that? Requiring "consensus" (a term that Quantling correctly pointed out, is not really used correctly on Wikipedia) requires that one position be the "default", so an overwhelming vote to the contrary is required to change it. That isn't the case here. Neutron (talk) 20:53, 27 October 2010 (UTC)

Definitely. So why do we insist on using the fuzzy and ambiguous language about "to determine is there is a rough consensus," etc? Let's just say "This is a vote to determine whether there will be open or secret ballots. The option gathering 50%+1 votes will be adopted." Nice, clean and clear. Short Brigade Harvester Boris (talk) 02:59, 28 October 2010 (UTC)

I don't know who is saying what on this issue, because I try to stay out of the internal politics of Wikipedia. Indeed, I have done my best to remain ignorant of it. I just edit here and there when I can. I wish I had time to do more, but since I don't, I mostly leave the "governance issues" to you more active folks. I have gotten a little bit involved in commenting on the election procedures, because it is something that I know something about and have real-world experience in, and I'm trying to be helpful. On the issue of "consensus" generally, while I do not think the word is used correctly in Wikipedia's policies and practices, it seems to be so ingrained that it probably isn't going anywhere. But I get the sense that there is a deeper political and/or ideological struggle going on here, in which you are involved and I am not, so I will leave you to your struggle. Neutron (talk) 14:56, 28 October 2010 (UTC)

It's just that I'm a stickler for clear and precise language. There's no underlying political/ideological agenda. We're going to treat it as a vote; there's nothing wrong with a vote (that's how important decisions are made in the real world); so why use vague language implying that it's something else? I'll drop the issue since no matter what I say, people insist on imputing some hidden motive. Sometimes a cigar is just a cigar. Short Brigade Harvester Boris (talk) 15:24, 28 October 2010 (UTC)

"Narrow consensus" is an oxymoron according to (my reading of) the Wikipedia page for consensus. —Quantling (talk | contribs) 20:16, 2 November 2010 (UTC)

More imagination needed

Some ArbCom members should be chosen one way, and some ArbCom members should be chosen another way. That will result in more diverse viewpoints, and if one of those selection methods turns out to work spectacularly better than the other, then the other can be changed.

Personally, I like a method similar to jury selection. That is, make it random, perhaps with the ability of a two-thirds majority to overturn the random selection. A minimum number of article edits could be required, and editors though perhaps not admins would be eligible. Any randomly chosen person could decline.

To those who say two selection methods would be too complicated, here's a reminder: every bicameral legislature in the United States uses two different methods.166.137.138.153 (talk) 18:53, 27 October 2010 (UTC)

I have serious doubts that that would work. The people that support open voting would claim that the people who were voted in by closed voting got their positions through a corrupt system and sockpuppetry and yada yada yada. The people that support closed voting would claim that the people who were voted in by open voting got their positions through groupthink and biased commentary and yada yada yada. At its best, this will make everyone unhappy instead of just the vocal minority. At its worst, it will create a split in ArbCom between those elected in open voting, and those elected in closed voting. Don't worry so much about people being unhappy, some people thrive on being unhappy. That being said, both sides have legitimate concerns, but I don't think your proposal will solve any of them. Thanks for the suggestion though. Sven Manguard _Talk 19:37, 27 October 2010 (UTC)

Well thanks for thanking me.  :-). But I suggested a change far more fundamental than that: random selection.166.137.137.161 (talk) 20:02, 27 October 2010 (UTC)

That one will never happen. A high level knowledge of policy, a track record of sound judgment, the trust of the community, and approval by the cabal are all requirements to becoming an ArbCom member. Plus, unlike jury duty, which is mandatory (ha! but I digress) we can't force people to be on ArbCom, which is a stressful and high pressure job, unattractive to most, which does not have the power that most people ascribe to it (like the courts, they can only make decisions when cases are brought to them). In short, only a few people (in 2009 22 out of several million) even volunteer to become part of ArbCom. Even then some are not clearly not viewed as qualified (at least 3 candidates of the 22 received less than 20% support, with one candidate having just under 500 more oppose votes than supports.)

In short, random selection will never be accepted, and will never work. Sven Manguard _Talk 20:14, 27 October 2010 (UTC)

As I said above, "Any randomly chosen person could decline." Moreover, I seriously doubt that ArbCom policies are more complex than all the statutes that juries must typically deal with. If ArbCom policies are that complicated, they need to be scrapped. In any event, I specifically said that not ALL of ArbCom should be randomly chosen. The wise and wizened choices of the cabal would still be there to guide and steer the randomly-chosen plebian morons.166.137.137.161 (talk) 20:24, 27 October 2010 (UTC)

One major difference is, a jury deals with one case. Then a new jury is selected to deal with the next case. The jury is instructed as to the law, but only the law relevant to that case. They are not expected to have knowledge of all law that could potentially come up on every case. The ArbCom is different; they deal with (potentially) all the cases that come up during their two-year term, and therefore need to have at least general familiarity with Wikipedia policies and guidelines. For that reason, a random selection would not work. The ArbCom is really more like a panel of judges than a jury, though there are differences between the ArbCom and a court as well. Neutron (talk) 21:00, 27 October 2010 (UTC)

Why couldn't the elected members instruct the random members about WP policy on a case-by-case basis? More importantly, even if you reject random appointment, the fundamental point of my initial comment was to use more than one selection method. For instance, you could divide the ArbCom seats equally between admins elected secretly by admins, and non-admins elected openly by non-admins.166.137.137.161 (talk) 21:09, 27 October 2010 (UTC)

Two comments, if I may:

• Fundamentally, I have never seen ArbCom to be much assembled as a "jury" as much as a "judiciary", and as such, we want to make sure we have the best and the most clueful serving the Committee.
• I know it's unavoidable and "everybody does it", but especially with regards to the ArbCom election which the standards are intentionally much lower to run, I don't think it's a good idea to further separate admins from non-admins; this goes back to the "factionalism" issue YellowMonkey brought up above.

–MuZemike 21:21, 27 October 2010 (UTC)

Comments on Statement by Risker

Per our tradition of not having threaded discussion on an RFC, I am moving these discusssions over here:

(vote)Agree with Risker's original proposal, less sure about the additions of endorsers. John Carter (talk) 16:14, 27 October 2010 (UTC)

John Carter - would you accept the seating of a candidate who received 1 total vote - a support? If not, you support a quorum call. What number of voters (either raw, or expressed as a formula of something) would you support? Hipocrite (talk) 16:32, 27 October 2010 (UTC)

You seem to be calling for a quorum call yourself, with the 33.33% figure. I like Risker's original proposal of more support than opposition votes. In all honesty, I cannot see it ever arising that the hypothetical you put forward is ever likely to happen short of nuclear devastation of the planet or something similar. John Carter (talk) 18:38, 27 October 2010 (UTC)

(vote) It is important that we clarify this issue, as there is a potential conflict between the number of successful candidacies desired and the minimum threshold for success. Skomorokh 16:40, 27 October 2010 (UTC)

So someone who gets 10 supports, 9 opposes and 981 neutrals qualifies over someone who gets 499 supports, 501 opposes and no neutrals? When people strategically vote oppose (rather than neutral) for all but the candidates they support, it will automatically suppress the overall ranking percentages. The only solution is to drop the 50% threshold, which was developed not by the community but by Jimmy in completely different circumstances. Tony (talk) 16:47, 27 October 2010 (UTC)

I'm proposing two specific criteria, which I believe have the support of the majority of the community. That doesn't mean others can't propose other criteria, such as a minimum number or percentage of voters indicating a non-neutral position. I'd thought about doing that, but don't have time to run the numbers on what percentage of candidates in the last 2 years would have met whatever criteria I may have come up with. In particular I can't easily find the number of community members who voted in the 2008 election. Risker (talk) 16:51, 27 October 2010 (UTC)

I'm going to expand on my response here. Just under 1000 voters participated in the 2009 elections, and all candidates who were appointed received either support or opposition from at least 35% of the voters; this would meet Hipocrite's 33.3% threshold. Indeed, every single candidate in that election met that ratio. However, about 900 voters participated in the 2008 election (I don't have exact numbers, unfortunately)...and only 7 of the 10 appointed candidates got 300 or more combined support/oppose votes. Of the candidates not appointed, only one other candidate reached the 300 or more combined support/oppose votes: Kmweber, who had 294 opposes and a support ratio of 10.4%. Given the change in voting practices between the two elections, I suggest that Hipocrite's additional criterion be rethought.

As to the 50% threshold, I cannot understand any kind of justification wherein someone who was rejected by a majority of the voting community should be appointed to the committee. Not a single candidate last year had fewer than 300 combined support/oppose votes, so there was plenty of feedback to every candidate, even those whom some might consider "fringe".

I am, however, concerned that an obsession with filling seats may lead to a situation where candidates who are opposed by more people than they are supported, or who did not participate as a candidate in the election, wind up being appointed to the committee. I don't understand how that would be considered a "good thing". Risker (talk) 19:55, 27 October 2010 (UTC)

IIRC, in the 2008 election it was far easier to not-vote on a candidate, while in the 2009 and 2010 elections, one needs to choose to click a box. As such, I do not believe that the 2008 election is precedential. However, there is a number or % - if only 1 of voter voted on a specific candidate, I think we all agree they should not be seated. As such, it's a matter of degree. Would you be comfortable with 20%? Hipocrite (talk) 19:59, 27 October 2010 (UTC)

Oh, I agree with you that 2008 is probably not precedential, for the reason you've identified. I've just pointed out that the 33.3% combined support/oppose vote level is meaningless, because every candidate met that last year. In fact, the lowest number last year was 40% combined support/oppose vote, every other candidate had a higher ratio. We will see in about a month and a half if that holds true for the second year in a row, but it seems this is a solution to a problem that doesn't really exist. I would, however, really like to hear your thoughts about appointments of candidates with less than 50% support and/or who did not actually stand as candidates in the election. Risker (talk) 20:20, 27 October 2010 (UTC)

I don't think either of those things are good ideas. ++Lar: t/c 04:19, 28 October 2010 (UTC)

My name appears in the support section of your well-thought out alternative to NOTA, as I support both an absolute lower bound and a subjective lower bound. Hipocrite (talk) 12:29, 28 October 2010 (UTC)

I'd rather have interim elections than arbitrators whom a plurality of voters rejects; the latter situation is simply not justifiable for an elected representative body. The number of neutrals is something of a red herring, as it is a meaningless metric even under the status quo. Skomorokh 18:27, 27 October 2010 (UTC)

(moved as of now: Risker (talk) 19:04, 27 October 2010 (UTC))

• The assumption seems to be that an Oppose is a rejection of a candidate, rather than simply a way of strengthening the position of the candidates one Supports (rather than leaving the others on the default Neutral). Tony (talk) 04:27, 28 October 2010 (UTC)

• Yeah, that does seem to be an assumption being made here. I hope that it is, in general, a reasonable one, although I do know that there are going to be at least a few people who do act in the way you describe, probably in every election. The question, I suppose, is what would happen if there were fairly clear evidence which led a substantive number of people to believe that the "tactical" opposes of the kind you describe led to a form of "hijacking" of the election. I guess in that instance we could either have a special election, as a form of "run-off", maybe. Alternately, maybe something like Australian weighted ballots might work best. So, for instance, in what I hope are most instances, if there are 8 seats to be filled, someone votes for the eight they think best qualify, and then list in descending order of preference the others they would accept if their first choices don't make the cut. Those they would not think basically acceptable need not be "numbered as alternates" at all. A system of that sort, I think in most cases, might help make it easier to see if their are a lot of the "tactical" opposes Tony postulates. John Carter (talk) 14:23, 29 October 2010 (UTC)

Jimbo Wales' response

After being notified of these issues, here is what Jimbo had to say:

I suppose the core question here is what I am supposed to do in the event that there are fewer than 18 arbitrators available with at least 50% support? The options would include, at least in theory, that I appoint someone who ran but gained less than 50% support, or that I appoint fewer than 18, or that I extend terms of existing arbs who weren't running for election, or that I appoint people on the advice of the arbcom and others based on past experience, or that I call for a second round of elections. Risker's proposal would suggest that I remove some of those options - precisely the ones that I wouldn't consider in the first place, though.

We traditionally have annual elections, but I have the right to call additional interim elections in case of a shortfall in staffing. What I think would make sense is that if, at the end of the election, we have a shortfall of a seat or two, I would likely treat it the same way I treat resignations throughout the year... basically, I would leave the seats vacant. If we had a bizarre situation in which there is a serious difficulty in finding candidates with at least 50% approval at all, I think it would be wise to consider that a serious signal that something has gone wrong with the entire process, and I would call for a wide-ranging discussion about ArbCom composition, function, etc.

One of the core advantages of our traditional "constitutional monarchy" system is precisely that in case of breakdown of process in some way - which is bound to happen although hopefully less and less frequently over the years as we gain experience and deal with various issues, we have an "answer", which is that I am theoretically free to dismiss ArbCom and even dismantle the entire system in favor of something else.

Imagine if the ArbCom angers 51% of the community, and a poll is held which involves 51% of the community demanding that the entire ArbCom resign immediately. Or imagine your own favorite meltdown scenario. There is no rule or policy which would give the general community the right to do something like that, and there are good arguments against it. (One thing we want from our judges is a certain amount of political independence and the ability to take unpopular decisions that are right for the encyclopedia, within bounds of course.) We have the choice of either trying to a priori figure out every possible thing relating to such scenarios, or we have the choice of what we do now: don't worry about it and try to do something sensible based on whatever the conditions are at that time. I think that's a good thing to continue. :)

I am happy when there are processes (like Risker's RfC in this case) that give me sensible guidance, not specific to a particular possibly inflamed situation) as to what to do in weird circumstances. It is my strong preference to do nothing at all, rather acting as a conduit and insurance that thoughtful and deliberate moves aimed at broad consensus in support of our encyclopedic mission is always the guiding principle.--Jimbo Wales (talk) 22:29, 28 October 2010 (UTC)

Skomorokh 11:33, 29 October 2010 (UTC)

Comment on MuZemike's statement

moved from main page

I think this would be better addressed by moving it to 'Statement by Risker' at the talk page. Jimbo Wales might also have addressed this in that section. Sven Manguard _Talk 16:11, 29 October 2010 (UTC)

Recommended reading

Anyone interested in the discussion of the merits of competing voting systems may be interested in reading Gaming the Vote: Why Elections Aren't Fair (and What We Can Do About It) by William Poundstone, or the last two chapters of Archimedes' Revenge: The Joys and Perils of Mathematics by Paul Hoffman, both of which are excellently written and very accessible. Newyorkbrad (talk) 02:53, 28 October 2010 (UTC)

Here are book reviews from the NYT: Poundstone, Hoffman.166.137.139.187 (talk) 14:43, 28 October 2010 (UTC)

Hooray for ruling class self-justication. In any case, hooray for POV heaven in eastern Europe and all of Asia articles. Hooray for massive WMF investment in the subcontinent, where a large minority/most editors are POV pushers, and where this now a big pool of non-editing "leaders" trying to cash in and promote themselves. Corrupt/racist in RL -> corrupt/racist on WP. All prepare for a D-Day style flood of subcontinental POV YellowMonkey (new photo poll) 05:09, 29 October 2010 (UTC)

• Can someone contact YellowMonkey, the real one, and have him check to see if the account posting here, YellowAssessmentMonkey has been compromised. Without knowing YellowMonkey, I am going to assume that it has to be, because I have a hard time believing that someone elected to serve on ArbCom would go around trash-talking WMF and ArbCom, especially like this. Also, can someone explain to me why someone needs an alternate account just to do assessments? Sven Manguard _Talk 05:48, 29 October 2010 (UTC)
  • Sven, YM has far more experience than you, and I know others share some of his frustrations about several issues that have been raised. While it is possible that you find it distressing that you cannot quite understand what YM is actually saying, or why it is being said, your comment is, like some of your other contributions here, remarkably unhelpful (some examples: [1] [2] [3] [4]). You are clearly capable of making useful contributions here (example), and everyone would like to see more of those, but the silly assumptions really do need to stop. Ncmvocalist (talk) 11:15, 29 October 2010 (UTC)
    • As to the first two edits you list, I was doing those becase there were (and still are) several statments that all are very similar to each other. With the amount of statements we have, currently 26, removing duplicates goes a great way in making the whole thing more readable. I can tell you now that there really are only four or five statements that are going to get any traction here. Neutron and Risker's statements are likely to become official policy. Hipocrite's statement is likely, if not now, in 2011 to be policy, and at the very least, will merit a continued discussion after the election. Will Beback and Angus McLellan's statements will form the basis of the consensus on the voting system. CBM's statement has wide support as well. That's 6 out of 26. Most of the rest are legitimate, if not fringe, ideas. A full 14 have less than 5 supports, many with one or two. That's more than half. The number with 10 votes is 9, roughly a third of the proposals. The point being, a lot of these are just taking up space. Some are outright duplications. I'd love to have just up and removed about five of the ones on the page. (Smallman12q's for example, which is one of the ones I tagged as being too similar to Angus' statement, has five votes, all of which are duplicates of votes at Angus' statment, and the statement itself is very similar to Angus' statement, and would, if it gained consensus, result in the exact same outcome as Angus' statement.) Two edits you listed were not mine, but I can see which comments you meant to list. I am standing by the comment on negitivity, becasue as you can see from the comments here, a lot of doomsaying is taking place, and most of it isn't helpful. Finally, as to the issue of YellowMonkey, I acknowledge that he has more experience, but the comments he is making are not coherent, and not very constructive (although, all things considered, with everything else I've seen, calling him out on being unconstructive is like calling one lemming out on falling off a cliff.) I would hope that even when I fail, most of my contributions can be seen in the light of that I am trying very hard to bring order and organization to a chaotic process. If I step on a few toes, that is regrettable, and I mean no ill to the people themselves, but this is a horribly chaotic process, and I'm only trying to improve it. I'm sorry that it isn't working. I really am. I intend on staying away from this page until the seven days are up and it gets closed. It's too chaotic for me to help with, so I need to step back and stop flailing, so to speak. Sorry Monkey, I'm sure you're a great guy and you don't deserve getting singled out. Sven Manguard _Talk 17:35, 29 October 2010 (UTC) P.S. If we do get Carter to oversee the votes, it would be the single most awesome thing on the internet, ever. I admit the chances of it happening are infinitesimal though.

Let's stop being so negative...

I would like to turn everyone's attention to WP:BEANS. Thus far, without any evidence presented that these things will or are likely to occur, we have had mentions of possible future sockpuppet voting, collusion, slander, bad faith mass opposes, and a half dozen other things too complicated to easily summarize that all are based on the as of yet unsubstantiated assumption that others will be acting dishonest and in bad faith as part of the voting procedures.

Seriously people, enough of this. How about this. Everyone just assume good faith, and stop listing all the ways things can go wrong. If something does go wrong, there's nothing we can do until after the election anyways, so this entire listing of scenarios thing is pointless.

Please? Sven Manguard _Talk 05:50, 28 October 2010 (UTC)

Problems which can be anticipated can in most cases be solved or at least mitigated; security through obscurity is a discredited superstition. Skomorokh 15:27, 28 October 2010 (UTC)

I'm not looking to the future. I'm looking to the past. Despite its popularity, I feel the closed voting in 2009 was a failure. It was not the most efficient way to get the best arbitrators for the job. Vodello (talk) 15:43, 28 October 2010 (UTC)

You are basing that feeling on two unsupported presumptions: that the result would have been significantly different with a different voting method, and that this difference would have been an improvement if it occurred. I see no indication of either. What I did observe, however, is that the 2009 elections were the most drama-free elections since those elections exist. — Coren ^(talk) 16:19, 28 October 2010 (UTC)

To be fair, I should note three unsupported presumptions of the secure poll group as well. Two are the same as your list above. That the results would have been different with an open voting method, that the difference would not have been an improvement. The third, that drama is detrimental to the selection process. One could easily argue that drama is essential to an open selection process because drama reveals a lot of what a secure poll aims to hide. While not a parallel, I can think of several RfAs, for example, where drama was essential to ensuring a desirable outcome. --RegentsPark (talk) 14:33, 29 October 2010 (UTC)

Actually, not really. I don't expect that the results of a secret ballot are significantly different that those of an open vote. I make no presumption either way because, as far as I'm concerned, the result of the vote isn't the objective. The objective is to remove (a) some very nasty side effects of raised hand voting that are detrimental by being there even if they do not actually end up affecting the result, and (b) the possibility that arbitrator impartiality is affected by who voted for or against them (or, at least, the appearance that it can be affected which is exactly as damaging).

Now, it's entirely possible that the secret ballot also has an effect on the result (by affecting tactical voting and groupthink). That would be a side effect on which I have no expectation; but if you read up on voting systems and with a few thousands of years of experience and literature in democratic elections, the expectation is that if there is a difference it is invariably towards fairness and "better" results. So "secret ballots give better results" might be presumptive, but it's a very well supported presumption. — Coren ^(talk) 19:32, 29 October 2010 (UTC)

I'm not so sure if the analogy with democratic elections is correct. Wikipedia is not meant to be a democracy and arbitrators are not there as representatives of the community (as is the case, at least ostensibly, in elections of governments for example). Rather, the idea is to choose people well suited for the job (i.e., they have the time and energy, they have shown that they can be neutral, and have demonstrated adherence to the principles of the encyclopedia). To see whether a particular candidate has these qualities (amongst others of course), we need more information rather than less. A secret ballot is an information reducing mechanism rather than an information increasing one and, in that sense, the result of a secret ballot is likely to be less 'good' than an open ballot, drama and all. --RegentsPark (talk) 20:47, 29 October 2010 (UTC)

Even if one granted the point that it's a job selection more than an election, you seem to be under the impression that open voting provides valuable information. I don't believe that is true. It contains plenty information-free rhetoric, and misinformation aplenty; but any valuable information is to be found on the extensive Q&A and on those voters' pages who took the time to make and make public a careful evaluation of candidates. Those are just as available under a secret ballot as they ever were. — Coren ^(talk) 23:26, 29 October 2010 (UTC)

I guess we just disagree on that. I find the open election format full of information. For one, it is clear who is voting for and against a particular candidate and that itself tells a lot. Then, there are the comments that most voters add to their votes. These tend to be very informative, often more informative than the tl;dr opinions of those voters who took the time to make and make public a careful evaluation of candidates. It would be great if we had the time to read everything and research everything but we cannot and any and every piece of information on a candidate that is readily available is, in my opinion anyway, useful. --RegentsPark (talk) 20:40, 30 October 2010 (UTC)

You traded for a drama-free election, and in return received a drama-loaded arbcom. It's not too late to veto that deal. Vodello (talk) 18:41, 28 October 2010 (UTC)

I don't believe this is an accurate description of reality. — Coren ^(talk) 19:50, 29 October 2010 (UTC)

Even if I were to agree with your viewpoint on the current ArbCom, and I don't, I would suggest that the voting method is not the reason but the candidates; we can only vote for and against those whose names are in the hat, and those chosen from that pool, by whatever means, are the basis on how an ArbCom is perceived. LessHeard vanU (talk) 14:57, 30 October 2010 (UTC)

None of those are hypotheticals about the future, Sven, they are observations about the past. All of your examples have occurred in past elections, to varying degrees, so the assumption that those things would happen again are not unsubstantiated. — Coren ^(talk) 16:25, 28 October 2010 (UTC)

Comments on statement by Od Mishehu

Moved from main page

• Comment. The WMF developers don't owe us anything in relation to this election, and we are very thankful that they are willing to devote one of their very busy techs to assisting in the running of SP again. (Could I remind users that there is a gigantic backlog of technical changes for en.WP?) It would be stretching the friendship to ask for major changes to be made to the system at this late stage. I'm sorry to pour cold water on a change that I want too; by all means, ask for it, but please don't expect it. As FloNight says, next year is a better time to push for major changes. It would need to be done earlier in the year. Tony (talk) 13:55, 30 October 2010 (UTC)
  • Additional Comment It is likely, based on what I have seen, that this will be affirmed as community consensus, but it is going to be impossible to do for 2010. After the RfC is closed and the election is over, we should bring this to the developers in January. That will give them the full year to do it. Mind you, there is a backlog, and it might not be given a high priority, but that's why we're giving them a year. What is important is the follow up. Once a month, no more, no less, you should check in on it and make sure that it isn't forgotten. The follow up seems to be what the 2009 resolutions were lacking. It will, of course, be up to you, Od Mishehu and the signatories, to do the follow up. Sven Manguard _Talk 20:12, 30 October 2010 (UTC)

Reminder: discussion due to close in three days

This is just a memory jogger for editors who may be meaning to comment further here but have been distracted by events around the wiki in the past few days: the closing time for this RfC is listed as 22:01, 3 November 2010 (UTC), 10 days in. I will be posting a request for closing by an uninvolved admin over the next few days, making it clear that this RfC needs definite outcomes and distinctions between reforms that are presently achievable and those that would require outside assistance. Given the difficulty in having a proper discussion and gaining momentum behind a proposal over a day or two, editors who want to make a statement intended to effect this year's elections would do well to do so soon. Last year's discussion saw a lot of late proposals that failed to gain sufficient support for adoption, and the closing statement is worth taking into consideration for the threshold of support required. Thank you for your attention, Skomorokh 21:13, 31 October 2010 (UTC)

Page stalkers, this is your 24 hour (or close enough) warning on the deadline. If you have something to say, now is the time. Sven Manguard _Talk 23:27, 2 November 2010 (UTC)

Well, let's clarify that the part of the RfC that needs to be closed tomorrow is the voting method that we are going to be using, which looks like, barring any Act of God or anything, we will use SecurePoll again. I'd argue that we can keep the other parts of the RfC (i.e. levels of support needed to be chosen for the ArbCom, Schulze method, factionalism, etc.) going for a bit longer, say another 10 days; we would definitely need to close everything up before 14 November, when nominations open. –MuZemike 23:42, 2 November 2010 (UTC)

It has been widely publicised. There is a waterfall of commentary. Contributions has tapered off markedly (although there might be a few more today in view of the deadline). I'd prefer the whole thing was closed soon. We have to get the elections up and running without having an uncertain environment. Tony (talk) 02:44, 3 November 2010 (UTC)

I concur with Tony on this; trying to close off portions of the RfC while leaving the rest of it open for further comment (including new statements on the same issues) would have potential for much confusion. Commentary has slowed to a trickle, so I think we can safely have it closed and invite further discussion on open issues at the election talkpage. I'll be posting at WP:AN to request an uninvolved admin later today. Skomorokh 09:24, 3 November 2010 (UTC)

I third the motion. Let's get the RfA closed as soon as possible. In all likelihood, we're going to have disagreements and another long thread of comments once this is closed and the consensus is announced, so we want to have time for that before the election starts up. Sven Manguard _Talk 20:11, 3 November 2010 (UTC)

Interpretation of results

I have a comment relevant to what the closing admin is going to have to sort out here: Although the statement by Risker (appointees must have at least 50 percent support, 27 supports) and the statements by A Horse Called Man (Shulze method and single-transferable vote, 11 and 5 supports respectively) are in different sections, it seems to me that they are opposing proposals. In other words, if a candidate needs at least 50 percent support, that assumes support-oppose(-neutral) voting, and that would exclude a ranking-based system such as the Shulze method or STV. As I understand it, those systems do not tell you what percentage of voters supported each candidate. Risker's statement received more support than either of the other two statements, and more than both combined. Although the negative comments regarding the Shulze/STV proposals on this talk page mostly concerned the fact that they are technically infeasible for this year, a larger issue is that they did not gain as much support as the current support-oppose system. At the same time, I think the results of this RfC should apply only to this coming election, with a lengthier process taking place next year that will hopefully result in a "permanent" framework for these elections. Personally I would like to see some kind of "proportional representation" if it could be worked out in all its details, but clearly, it hasn't been, yet. Neutron (talk) 01:18, 1 November 2010 (UTC)

In one sense, the proposals don't conflict. Shulze is a method of analysing ranked votes, not a method of voting. Voting Support/Neutral/Oppose is equivalent to voting all your Supports as preference 1, Neutrals as preference 2, and Opposes as preference 3. That's a good reason in itself for using rank voting, as it allows finer levels of support/oppose if the voter wishes – so you could emulate 'strong support'/'support'/'weak support' as well equivalent opposes by using 7 ranks of preference (or more if desired, right up to giving a preference between all candidates). Of course, Shulze doesn't show level of support (>50% for example), because that's a crude metric, somewhat unsuited to the job of filling 8–10 seats from a field of 20+ candidates, since there's only three levels of preference allowed. I guess this is really a debate for a later time, but it doesn't hurt to raise awareness while the topic is still 'hot'. --RexxS (talk) 02:11, 1 November 2010 (UTC)

I think Neutron's point was to query how the Risker criterion (successful candidates require more support than opposition) could be reconciled with a system of ordinal ranking. There are conceivable ways of doing it (e.g. set the quota to half the electorate), but I don't think anyone wants the closing admin of this RfC to come up with them. Skomorokh 02:19, 1 November 2010 (UTC)

That's right. What I was trying to avoid was a closing admin thinking, "Well, there's no statement directly opposing preference voting so I am going to declare that there is a consensus to implement it in the 2011 election if it cannot be implemented this year." I do not think there is such a consensus. I would be interested in seeing whether there is a way to blend the two systems, as Skomorokh suggests, and then the community can discuss whether the benefits outweigh the drawbacks (including increased complexity and possible voter confusion over the consequences of their vote) and a decision made in time for next year's election. Neutron (talk) 02:49, 1 November 2010 (UTC)

It's not easy because of the insistence that one 'support' vote means precisely the same as another 'support' vote, and that they are exactly cancelled by each 'oppose' vote. At present, when one editor strongly supports a candidate, that vote is nullified by another editor who only weakly opposes the candidate. If we're happy with that, then >50% support is a reasonable hurdle. Rank order presupposes that all the seats are to be filled, so that criterion doesn't work in such a method. Oddly enough, thinking about it, rank order voting is actually amenable to 'None of the above' (although it would be none of the below), by including a NOTA candidate in the rank order when voting. Schulze (or similar) could then order the candidates (including NOTA) and we could agree that any candidates ranked below NOTA were unelectable. How's that for a guarantee of "sufficient support"? --RexxS (talk) 02:59, 1 November 2010 (UTC)

But that's not what "None of the above" means. Logically, someone who voted for the NOTA "candidate" COULD NOT vote for any other candidate, even in a ranked system, or they are (in effect) double voting. The technical people would have to figure out a way of "locking out" any votes by someone who votes for NOTA, and that's not something a developer should be asked to do on a rush basis. And I don't think my "lockout" idea is what any of the supporters of having an "NOTA candidate" are proposing, which is one more reason I don't think the "NOTA" idea is ready for prime time. And that holds true even if you make NOTA (or "NOTB") a ranked candidate in a preference voting system. It seems to me that the possibilities for unintended consequences in such a system are almost endless. In effect you are combining the vote for candidates with a vote on how big the arbitration committee should be, and I don't see the benefit of that. Why can't we just decide on the size of the committee (which has already been decided), and elect enough people to fill those seats? Neutron (talk) 03:17, 1 November 2010 (UTC)

Those who oppose preferential voting should do so by endorsing the statement by CS55cp. CS55cp (talk) 06:01, 1 November 2010 (UTC)

Thank you CS55. Moving on, I would like to emphasize that Risker's proposal would be slated to go into effect for the 2010 elections, while any alternate voting method that does not already exist (i.e. not secure poll or RfA style open voting) cannot go into effect until 2011 for technical reasons. The conflict is not immediate, so we can reconsile this later if we run out of time here. I suggested somewhere that we hold another RfA immediately after the elections to deal with things that we want for the 2011 election, as it will give us plenty of time to decide and actually implement proposed changes. Sven Manguard _Talk 07:35, 1 November 2010 (UTC)

The Wikimedia Foundation adopted the Schulze method on 19 May 2008 for a poll that started on 1 June 2008. I question the claim that the Schulze method is "technically infeasible" for this year. A Horse called Man 08:07, 1 November 2010 (UTC)

How was it implemented? Sven Manguard _Talk 15:19, 1 November 2010 (UTC)

It was implemented mainly by Kwan Ting Chan and Tim Starling. For more information, contact the 2008 and the 2009 Wikimedia election committees. A Horse called Man 15:41, 1 November 2010 (UTC)

Was there a discussion after those elections to obtain feedback on how the voters think those elections worked out? It would be interesting to read such a discussion as a starting point to a discussion on whether to implement this system for ArbCom elections in the future. One thing to consider, and this is just a supposition on my part but I am pretty sure it's true, is that a majority of people in the English-speaking world have never participated in a "real" election using any sort of preferential voting, and are therefore unfamiliar with it and perhaps uncomfortable with the idea. I know it is used in Australia and New Zealand, but except for a very small number of places here and there, it is not used in the U.S. and I know it is not used in parliamentary elections in the UK. Perhaps it is used at other levels of government in the UK, but I don't know. I am not saying we should be U.S.- or UK-centric or anything, just explaining why a lot of people might be leery of such a proposal. An election for the WMF Board of Trustees is a little different, in that it is a "worldwide" election, and these types of systems may be in greater use in other places in the world (I don't know.) For elections to the English Wikipedia Arbitration Committee, some people (including me) seem to think that the elections should look something like other elections we've seen in our lives, unless there is some proven advantage to doing otherwise. I myself was familiar with STV but not with the Shulze method. I think our articles on both are rather confusing. As I said elsewhere, in theory I would favor some form of proportional representation, because 51 percent of the voters should not be able to elect 100 percent of the winners (as they can do in theory now), but there is also some virtue in simplicity or at least understandability on the part of the electorate. I just don't think we're "there yet" as far as implementing one of these systems, even if it is technically feasible. Neutron (talk) 15:59, 1 November 2010 (UTC)

Here is the feedback for the Wikimedia Board of Trustees elections. A Horse called Man 16:10, 1 November 2010 (UTC)

Neutron's comments show great wisdom. I have used ranked preference voting in the past, though the method of tallying the results was different, and generally I felt satisfied at the results. That being said, ranked preference voting with more than five or six candidates and more than one or two winners becomes immensely complicated to calculate by hand, and makes verification of the results difficult.

Here is the exceedingly simple method we used in the elections with ranked preference.

1. Rank candidates from 5 to 1 in order of preferance, with 5 being the most prefered. Leave all unranked candidates' slots empty.
2. Collect the ballots. A ranking of 5 is worth five points, a ranking of four is worth 4 points, ect.
3. Tally points, the winner is the person with the most points.

Scary simple stuff, and we were able to rig a Scantron machine to do the counting for us. In all those elections, few people were ever unhappy, because while the winner might not be everyone's first choice, it was usually in everyone's top three, and at the very least, was someone almost everyone saw as a worthy candidate.

Now I'm posting this here as an illustration of how we can keep things simple. Right now I am not proposing that we implement it for 2010. Sven Manguard _Talk 16:26, 1 November 2010 (UTC)

That's the so-called Borda count. The Borda count is very simple, but it has very severe problems. A Horse called Man 16:37, 1 November 2010 (UTC)

In the UK, many elections select a single winner from several candidates – elections for the British Parliament being an obvious example. It is also common for local government elections to select two or three councillors from a field of candidates, in which case the number of votes cast per elector is limited to the number of seats available. In European elections, however, a single constituency will return multiple MEPs, so a system of proportional representation is used. All of these are likely to be familiar to a UK voter, but I doubt that anyone in the UK will have seen a governmental election where 10 seats are selected from 20+ candidates by a voting process of unlimited Support/Opposes. I don't think the analogy is apt, and it is generally recognised that classifying multiple candidates from a broad field is difficult to do fairly by support/oppose voting. The simplest analogy that occurs to me is that of awarding grades in public examinations, where borderline candidates may be classified by their rank order of predicted grades. Nevertheless, I'd pose the question to all who oppose a preferential voting method for ArbCom elections: "How do I vote if I want to support candidate A, but not at the expense of candidate B?" Whenever the number of seats elected is greater than one, and particularly when the number of candidates is large, Support/Oppose voting simply does not allow the voter to express those sort of nuanced preferences, and the results become more liable to being unsatisfactory to a larger number of voters. --RexxS (talk) 16:38, 1 November 2010 (UTC)

(←) Borda, huh. Thanks for the link, I didn't realize that there were that many single-winner preferential voting election methods. Ah well. It worked in the small semi-homogeneous groups that I dealt with, but I suppose junior high and high school elections are not quire the same as ArbCom elections. Sven Manguard _Talk 20:10, 1 November 2010 (UTC)

The result of this RfC has also been discussed here. A Horse called Man 14:57, 2 November 2010 (UTC)

An important point

I was hoping someone would pick it up, but I don't see it. The NOTA option — as a "candidate" — makes no sense and really should not be implemented. It will not have any useful result unless people who select it select it to the exclusion of anything else; it certainly does not have the effect of setting a threshold if combined with any vote (consider the possibility that every voter picks a number of favored candidate and NOTA: NOTA would have 100% support!) It also has bizarre significance if people are allowed to oppose that pseudocandidate; with or without votes for real candidates. For that matter, if nobody opposes the candidate it's even more ridiculous: it would also get 100% support if even one person "votes" for it.

At best, it'll be a meaningless option that'll confuse voters. At worse, it's going to pop some random number which will be misinterpreted (inevitably, since it has no semantic value) into a dispute about the election result. It only makes sense if it's a single "support" option to the exclusion of any real candidate, in which case it becomes exactly the same as voting oppose on every candidate. — Coren ^(talk) 02:47, 3 November 2010 (UTC)

I think I sort of did touch on this point in my comment in the preceding section, although you cover it more comprehensively. Perhaps more to the point, there is a comment ("Statement 2 by Neutron") on the RfC page in which I state that NOTA is both unnecessary and confusing, which I think wraps it up fairly simply. So people can choose Hipocrite's statement favoring NOTA or mine, opposing it. Of course, the way this RfC is structured, you wouldn't necessarily know there are competing statements on this subject. I am wondering whether I am the only one who has concluded that an RfC is not a good way to make final decisions on issues like this. Neutron (talk) 04:49, 3 November 2010 (UTC)

You're right — I hadn't even noticed there was now a statement in opposition! — Coren ^(talk) 11:13, 3 November 2010 (UTC)

That's the point I was trying to make with this comment above. I don't agree with NOTA under any circumstances, and it would only make any sense if the voting method were changed to rank each candidate, with NOTA being one of the candidates. That could allow an analysis to show those candidates who were pair-wise defeated by NOTA. That would a terrible kick-in-the-teeth for anybody who submitted a candidature in good faith, and I suspect it would have a big chilling effect on recruitment of a good pool of candidates. --RexxS (talk) 19:25, 3 November 2010 (UTC)

For next year's election

If not now, then perhaps next year sometime, could we make a note to have a discussion about moving the dates of the election to a different time of year? It takes a fair bit of work to research and vote on candidates, and having the election in the middle of the holiday season puts unnecessary time pressures on both candidates and voters. --Elonka 18:30, 28 November 2010 (UTC)

I don't necessarily disagree with the above, but would be interested in knowing which if any periods would be preferable. Early January (after the holidays) might work, as might any number of other times, but there is also, possibly, something to be said for the fact that around the holidays a lot of people get more time off, which they might, potentially, if they're masochistic enough, want to spend on reviewing candidates. Maybe. Does anyone know what times of year there tend to be the most frequent number of edits by the largest number of editors? John Carter (talk) 20:55, 28 November 2010 (UTC)

Also, holidays are at different times in different countries, so it would probably be impossible to find a period in which everyone has sufficient time. Regards SoWhy 21:36, 28 November 2010 (UTC)