Wikipedia talk:Requests for comment/Arbitration Committee 2/Archive 1

This page is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

We dont have much time. I think we need to spend a day or two quickly planning and collecting our thoughts. Can we do a poll with only a few options?

If we cant manage a poll, I think that we have a few days where views are posted before the community starts endorsing each others views. This will allow for different approaches to emerge, and be merged, rather than having a few similar proposals causing unnecessary bulk and making it more difficult to interpret the consensus. John Vandenberg ^(chat) 14:16, 27 October 2009 (UTC)

Thanks for putting this up, John. Personally, I'd like to see the issue of two-year terms proposed and endorsed, so that the community sends a clear message to Jimbo in time for the election. What's important this year, as I see it, is that we know going in that we're electing X number of people for Y length of time, without any of the tranche confusion of previous years, and the numbers being picked changing after the election. SlimVirgin ^{talk|contribs} 14:20, 27 October 2009 (UTC)

What Slim says. Pleased you've done this, John. IMO, there are compelling reasons that the community (1) adopt two-years terms as the norm for arbs' terms; (2) dispense with the notion that tranches have a functional role, beyond signifying the end of an arb's term; and (3) adopt secret ballots for ArbCom elections. I'd be pleased to discuss my reasons with anyone. Thank you. Tony (talk) 15:08, 27 October 2009 (UTC)

In regard to term lengths, I see three options which have been well defined:

• 18 months (strong support in the last RFC)
• two years (seems to be the prevailing opinion at the moment)
• three years and a tranche system with some successful nominees being given a shorter term. (i.e. no change)

In regards to number of seats, which interplays with the term lengths a little, there is less clarity. The trouble is that the number of seats is what we need to define before the election starts. We have 10 arbitrators who will retain their seat next year. I think the two main options are:

• 15 seats, meaning 5 appointments in this election
• 18 seats, meaning 8 appointments in this election

John Vandenberg ^(chat) 22:25, 27 October 2009 (UTC)

John, are you planning to launch a formal RfC? It looks like two (or three?) simple questions might be put, with options. Participants asked to respond to each. Needs to be done soon if at all. Tony (talk) 04:17, 28 October 2009 (UTC)

• It would be nice, if possible, to recieve clarification from
  *
  05:12, 28 October 2009 (UTC)

• It's not really a question of rights, Finn, because we're breaking new ground here somewhat. If the community's very clear about what it wants, there'd be no reason for Jimbo to do otherwise; it's only where there's a lack of clarity that he'd feel a need to step in. John, I agree with Tony that an RfC/poll could be drafted to deal with the few questions that remain. SlimVirgin ^{talk|contribs} 06:47, 28 October 2009 (UTC)

I've drafted it as a poll + RFC so that the community can vote on fine grained options that will help us determine if there are clear answers to the remaining questions, while also providing space for views/opinions/proposals to build consensus. John Vandenberg ^(chat) 09:46, 28 October 2009 (UTC)

Per

WT:ACE2009, should we have a yea/nay question regarding the secret ballot issue? Might be nice to put the issue to bed. UltraExactZZ ^Claims ~ Evidence

19:56, 28 October 2009 (UTC)

I am biased as I support using a secret ballot for this election based on the RFC that was held. The RFC results were clear, and people have been aware of the results for a long time. Next year we can hold another RFC about election methods, and follow wherever it leads us.

I'd also rather that this RFC is focused on questions about the structure of the next committee. I'm going to kick it off now as-is. John Vandenberg ^(chat) 09:00, 29 October 2009 (UTC)

Agreed, now that it's not this insurmountable techno-babble nightmare to implement, I agree (read as: Someone else is handling it). Thanks, UltraExactZZ ^Claims ~ Evidence 12:27, 29 October 2009 (UTC)

I'd like to second those who have commented that the format of this RfC is very well-organized and will help make it clear what the community's consensus is on these issues (if there is any). Thanks. Newyorkbrad (talk) 21:47, 31 October 2009 (UTC)

Seeing as there was significant support for moving on from

As there are no objections, I'll add a poll question shortly.

For simplicities sake, I've listed the options as

Why not just call them "Support/Oppose" and "Schultze"? And is there significant support for the latter? Not much at the RfC certainly ...  Roger Davies ^talk 13:14, 30 October 2009 (UTC)

If we aren't choosing a system for immediate use in this years election, it does not make a lot of sense to be overly specific when we have not even established the general principles on how to select winners. I would be good to be able to say next year "As the October 2009 RfC showed, there is strong support for retaining the approval method/changing to a preferential system/selecting arbitrators by lottery, so now we can happily make no change/decide between the different options for preferential systems/choose between picking names out of a hat or the largest prime number edit count". There are more than one Schulze methods for example, and enough interest in them to put the electoral method up for discussion here (I mentioned it in the introduction only to explain this). As for nomenclature, casting the options as approval and preferences cuts to the heart of the distinction I think.

barbarian 

13:34, 30 October 2009 (UTC)

I think we need more variety in there, so I have added cumulative voting. John Vandenberg ^(chat) 14:05, 30 October 2009 (UTC)

Good job.

barbarian 

14:16, 30 October 2009 (UTC)

• This section has just been drawn to my attention, and it ascribes my personal position to the Arbitration Committee as a whole. That is incorrect; it is my own personal opinion, but the Arbitration Committee as an entity has not taken a specific position on election method. Skomorokh, could you please amend? Risker (talk) 14:32, 30 October 2009 (UTC)

Risker, I am sorry if I have mischaracterised your comments, and I have removed the preface, but I am not ascribing your personal position to the Committee. In response to my comment about election reform, you wrote "After a lot of careful consideration, we have opted to go with straight support/oppose voting rather than preferential voting", which I phrased as "the Committee have indicated their preference for not changing the election method". By "we", surely you meant the Committee and not yourself? Could you or another arbitrator clarify whether or not the Committee has a position on the election method of the December 2009 ArbCom election or if it is up to the community to decide?

barbarian 

14:46, 30 October 2009 (UTC)

Ah, fair enough, Skomorokh; my intention was to speak specifically to the AUSC election, currently underway, which is held under the auspices of the Arbitration Committee. I don't think the Committee as an entity has a position on the appropriate method for electing new members to Arbcom. Risker (talk) 15:03, 30 October 2009 (UTC)

Oh that explains a lot, it did seem very peculiar for an elected group to set the terms of how new members are elected, but quite appropriate to do so for a delegation (AUSC). Again, I am very sorry for misinterpreting your remarks.

barbarian 

15:13, 30 October 2009 (UTC)

This raises the issue of whether or not the current RfC will determine the electoral method for the December 2009 election, and if so, whether we do in fact need to be specific about which exact method to use. Should we start over with support/oppose,

How would this work? Would we have elections every 6 months, or would we have fewer arbitrators for half the year? I signed under 2 years, but I would actually prefer less. Mr.Z-man 14:47, 29 October 2009 (UTC)

Logistically, I'd have to work up a chart - Best guess, I'd say three tranches with elections every 6 months. The group elected now would be up for re-election in June 2011, and so forth. There was a similar proposal from the RFC last year with 2 year terms, but 4 tranches elected every 6 months. With a simple and reliable election system (which might be easier with securepoll), two elections a year might be feasible. But there may be objections to the disruption to the community inherent with any election. That's part of why we do them in December, I think; it's a slower time of year generally, with all of the holiday shenanigans. UltraExactZZ ^Claims ~ Evidence 15:34, 29 October 2009 (UTC)

I'm not real keen on elections every six months, but that is probably due to having been through the harrowing experience. The election would be less of a deal if the appointments were for 18 months, and a secret ballot means there is less "buzz" during the election. Elections every 9 months would be more manageable; it isnt as "neat" as they would occur at different times of the year.

We could have elections every year, as we do currently, and require that half of the incoming arbs work on ban appeals for the first six months, and the other half work on ban appeals for the last six months. This would work better for 18 seats, with 14 voting members, however it also works for 16 members with 12 voting members. A quick ascii chart of both options:

               year
               2009      2010      2011      2012   
election 2008  2    2
         2009  8    2+6
         2010  4+4  8    8    4+4
         2011            4+4  8    8    4+4
         2012                      4+4  8    8   4+4
               -------------------------------------
         seats 12+4 12+6 12+4 12+4 12+4 12+4 ........ = 16 members, 12 voting.

               2009      2010      2011      2012   
election 2008  2    2
         2009  8    4+4
         2010  4+4  8     8   4+4
         2011             6+4 10   10   5+5  (one person misses out of basc)
         2012                      5+5  10   10  5+5
               -------------------------------------
         seats 14+4 14+4 14+4 14+4 15+5 15+5 ........ = 18 seats for the next two years, and then 20 seats thereafter.

John Vandenberg ^(chat) 08:54, 30 October 2009 (UTC)

Since the election draws near and this RFC's outcome could affect voting style and results for a lot of people, shouldn't it be advertised in more places with higher visibility (watchlist notice, noticeboards etc.)? Reagrds SoWhy 13:35, 30 October 2009 (UTC)

Definitely. Here's the list of relevant places the AUSC election was publicised:

• Wikipedia talk:Arbitration Committee
• Wikipedia talk:Arbitration Committee/Noticeboard
• Wikipedia talk:Bureaucrats
• Wikipedia talk:Elections
• Wikipedia:Administrators' noticeboard
• Wikipedia:Administrators' noticeboard/Incidents
• Wikipedia:Arbitration Committee/Noticeboard
• Wikipedia:Village pump (miscellaneous)

as the beginnings of a publicity list.  Roger Davies ^talk 13:39, 30 October 2009 (UTC)

I've just added it to {{

barbarian 

13:50, 30 October 2009 (UTC)

I added notices to

Wikipedia:Arbitration Committee Elections December 2009/Vote, both to highlight this RFC and to note that the voting process is subject to change (currently formatted for the open election). I ignored the voting process page, since - if we go secret - it'll probably be deleted entirely. UltraExactZZ ^Claims ~ Evidence

19:47, 30 October 2009 (UTC)

Also Posted at

Wikipedia talk:Arbitration. UltraExactZZ ^Claims ~ Evidence

20:39, 30 October 2009 (UTC)

First, I like the overlapping tranche system. It helps much more than it hinders. Institutional memory is a good thing.

If this is switched to 2 year terms, then I'd like to see 4 tranches of "around" 4 members, with 6 month elections. - jc37 14:16, 30 October 2009 (UTC)

Institutional memory is important, I agree. However, making significant changes to the composition of the committee every six months is destabilising. It takes a few months for a committee of this size (whether 15 or 18 members) to become familiar with each other, identify interests, strengths and weaknesses, and get "stuck in" to the work. Changing the personnel extensively on such a frequent basis will adversely affect the effectiveness of the committee. Risker (talk) 15:19, 30 October 2009 (UTC)

I don't disagree. Personally, I think it should stay to year elections with 3 year terms. - jc37 16:37, 30 October 2009 (UTC)

• A two-year term system with dual tranches being elected every year allows for overlapping. Half the Committee (barring resignations) will be continuing every year, with the other half being newly elected. This system seems to be the best choice for a direct and stable Committee. —
  *
  17:25, 30 October 2009 (UTC)

Is it necessary to have this discussion in this RfC? Do people feel that the consensus of the previous RfC was not sufficient? This is not a rhetorical question; I don't have a strong view either way, but many people, myself included, have been working (rushing through code updates for SecurePoll, updating WMF site configuration whilst in the middle of significant RL upheaval, etc) under the impression that the use of SecurePoll is expected for ACE2009. What has changed since 28 August in this arena to require this new RfC? Happy‑melon 18:11, 30 October 2009 (UTC)

Considering voters turn out in hundreds to vote on elections, I hardly think 40 or so people commenting is sufficient. Besides, there were several people who disagreed with the analysis and reasoning given for changing things. Majorly talk 18:22, 30 October 2009 (UTC)

@Happy-melon: There was indeed a previous RfC, but it was not very well advertised.[1] There was one small mention on a Village Pump page, it got a buried mention in a Signpost article after the RfC was closed, and I'm unclear whether it was on Centralized Discussions (it's not in the archive, though someone pointed out that they did see it there[2]). There have been strong concerns raised about secret balloting in several threads at the 2009 election talkpage, so I think it's worth bringing it up again in a more widely advertised RfC, to ensure we have a genuine consensus on the matter. --Elonka 18:35, 30 October 2009 (UTC)

It was on the centralised discussion template from 29 July to the 1 September. Davewild (talk) 18:59, 30 October 2009 (UTC)

How could you be unclear on whether it was on the centralised discussion template? You've mentioned that before, and you've been informed in detail that it was listed on CENT. Nathan ^T 22:05, 30 October 2009 (UTC)

I'm puzzled as well. I explicitly pointed this out, as Elonka as said. Either she is dissembling or (assuming good faith), she remembers looking at an outdated cached version of CENT and thus thinks it was "not there". It clearly was there on CENT. I would suggest that if Elonka is truly concerned that such discussion is publicised enough, that she notify those who took part in the previous discussion, so they can reiterate their opinions here. Carcharoth (talk) 08:00, 31 October 2009 (UTC)

Happy-Melon, I think the previous consensus to use a Secret Ballot was clear enough to justify using it for ACE2009. My objections were technical, and have been addressed - so I have no problem with the format. The caveat would be that this RFC is getting more publicity than the Secret Ballot RFC, so if a large consensus of editors turns out against a secret ballot, then perhaps we should rethink things. The open format of last year is in place as a default, ready to go, if that ends up being the case - was planning to rearrange those pages, but I'll hold off until it's certain. UltraExactZZ ^Claims ~ Evidence 19:38, 30 October 2009 (UTC)

That sounds like a reasonable approach: Continue with plans based on the last RfC, but be willing to change things if the new (and more widely advertised) RfC shows a different consensus. BTW, we should probably link this from

WP:AC/N as well. UltraExactZZ, would you like to handle that? --Elonka

19:52, 30 October 2009 (UTC)

I'd defer the announcement at WP:AC/N to an arbitrator, since it is for committee (and clerk) announcements. Any takers? As for RFAR and AE, I'll post it on the talk pages of both - the main pages should be kept just for business, and any notice would be lost in the shuffle at RFAR anyway. UltraExactZZ ^Claims ~ Evidence 20:30, 30 October 2009 (UTC)

 Done. Posted at

Wikipedia talk:Arbitration. UltraExactZZ ^Claims ~ Evidence

20:39, 30 October 2009 (UTC)

Looks good, thanks! --Elonka 22:14, 30 October 2009 (UTC)

The difference between the last RfC and this one is that this isn't an RfC, its a vote. Nowhere on the page does it discuss the meaning of "secret voting" - obviously its intended to sound dreadful, and

What else is this discussion page for if not discussion? Speaking for myself, I've read through this talk page, scanned/read through the last RfC, mulled it over, and lodged my oppose with all this in mind (and have continued to check in on further comments in case anything changes my mind). If people want to discuss and clear up misconceptions, they are welcome to use this talk page. -kotra (talk) 22:25, 30 October 2009 (UTC)

Another aspect to this is the bandwagon effect. We probably all know that the first few hours of an RfA are the most important. Candidates are more likely to get a fair shot if the tally is hidden until the end, which SecurePoll does. — Jake Wartenberg 04:34, 31 October 2009 (UTC)

Intimidation/harassment from public voting

Does this actually occur? Are people really that wrapped up in who gets to be an Arbitrator, a thankless job only an incredibly altruistic or unstable person would take? And if so, aren't we capable of dealing efficiently and effectively with such behavior? -kotra (talk) 22:40, 30 October 2009 (UTC)

Yes, some are that into it. Alas, blocking troublemakers is something that rarely happens, particularly if the troublemakers have powers in the community or influencial friends. Majorly talk 22:47, 30 October 2009 (UTC)

I don't doubt your honesty, but this seems like the sort of statement that merits investigation (not to dredge up old drama, but to make an informed choice). If diffs aren't available, can other editors corroborate this (particularly that people have felt harassed and/or intimidated)? -kotra (talk) 23:23, 30 October 2009 (UTC)

You can get harassed or stalked for anything, even opposing some powerful guy's FAC can get dozens of opposes at the reviewer's RfA (ostensibly for another reason of course, although usually transparent by the anomalous rationale of the opposer, although this is rarely acted upon) YellowMonkey (bananabucket) 01:27, 3 November 2009 (UTC)

As someone who helped run the 2007 and 2008 elections, I don't recall any incidents where votes were struck for intimidation. Some comments were moved for length to the talk page, and things may have gotten heated, but there were very few shenanigans during the vote itself. As far as stats go, I do have an analysis of the 2007 voting, and the number of votes indented for sockpuppetry, lack of suffrage, or double-voting. That analysis is here. It doesn't break out the reasons for the indents, just that they happened, and whose votes were indented (support/oppose, which candidate). Agree, it might be worth looking into, bearing in mind that there might be canvassing and intimidation without an on-wiki component that would be evidenced on the voting pages themselves. UltraExactZZ ^Claims ~ Evidence 01:55, 31 October 2009 (UTC)

• The arguments regarding intimidation and bloc voting could be made to a comparable degree with the RfA & RfB processes. Yet these have always been conducted in the open, without arguments that doing so allows approved administrators to exact revenge on those who opposed their RfA. I would be interested to hear opinions regarding why the Committee elections are so different and special that they require foregoing the traditional open consensus model for a model inspired by real-life politics. —
  *
  02:41, 31 October 2009 (UTC)
  • Three points:
    - First, variants of these arguments do crop up regularly, at
    WT:RFA;
    - Secondly, the sheer scale of the ArbCom election gives it great potential as a drama-fest and it is highly politicised;
    - Thirdly, in the ArbCom elections, consensus is determined by weight of votes rather than by strength of argument;
    The strongest arguments against public voting are that they create and encourage "follow my leader" pile-ons, and discourage the timid from participating. Few relish having their vote publicly challenged, by a strident editor with an axe to grind, and this probably effectively disenfranchises many many people.  Roger Davies ^talk
    03:22, 31 October 2009 (UTC)
    • Those are very reasonable justifications. It is especially persuasive to see sitting arbitrators argue against the arrangement that elected them - it speaks to their integrity. I am indeed loath to disagree with such an experienced editor.

-However, I have reservations about making this one area an exception to the

WP:AGF

atmosphere of openness that permeates the project.

-An example: Should we require secrecy at RfC's such as this one, because Mr. Davies might supposedly exact revenge upon me for expressing an opinion contrary to his - or because my "followers" might mindlessly "follow my lead"? That is obviously hypothetical, as Mr. Davies is of good character, and as I have no "followers" :) Such an objection appears to be the logical consequence when secrecy proposals are pushed to their ultimate conclusion.

I fear I shall always be a proponent of a genial environment at Wikipedia without the suspicion and distrust that are evident in the arguments for secret votes. That said, I really do respect my colleagues' point-of-view. —

*

04:02, 31 October 2009 (UTC)

• The key point here is that a great deal of public voting is very far from genial. I know many fine editors (who would make great admins) who simply will not go through an RfA because of the ruckus from cliques/editors they've fallen out with. One, in particular, knows it will be payback time for his past (very sensible and well reasoned) opposes. Secret voting doesn't remove this altogether, of course, but it does prevent the vociferous and strident from torpedoeing good candidates with snide comments attached to oppose votes.  Roger Davies ^talk 04:11, 31 October 2009 (UTC)

• <ahem> SandyGeorgia (Talk) 01:01, 1 November 2009 (UTC)

It doesn't have to be as blunt as harassment or coercion. Up until I quit I saw a few "We voted for ...x" comments here and there on pages here or there. Casliber (talk · contribs) 14:07, 1 November 2009 (UTC)

For myself, I'm increasingly concerned about intimidation and harassment from secret voting. I was one of the people who last year posted a "Voter guide" endorsements page.[3] There was minor risk from doing this, because my actual votes were already public as well, so anyone who really wanted to see how I was voting, could just check my contribs. But with a Secret ballot, it becomes a much riskier proposition for those of us making voter guides (see {{ACE 2008 guides}}), because we'll be the few people whose votes are not secret. This means that any candidate who wishes to hold a grudge is going to be much more focused on those with voter guides, than all those other "secret" opposers. --Elonka 16:27, 1 November 2009 (UTC)

No-one forced you to make a guide Elonka. Casliber (talk · contribs) 19:57, 3 November 2009 (UTC)

Or more generally, making a voter guide is akin to offering opinions in a newspaper, website or book about a candidate IRL. Your expectation of privacy diminishes when you are advancing a rhetorical claim about a candidate. Protonk (talk) 20:07, 3 November 2009 (UTC)

• The strongest argument to be made with regard to intimidation is not related to explicit harassment but tacit harassment. We should be worried about people who refuse to vote on the basis that their decision might be scrutinized in the future (say at a conduct RfC or at their RfA). Obviously this sort of harassment is impossible to measure, but it is concerning. Personally, I don't treat it as significant enough to merit the loss of transparency involved with moving to secret ballots, but it is likely more important than harassment in the form of chasing down voters to get them to change their stance. Protonk (talk) 18:55, 2 November 2009 (UTC)

• I don't share your notion of transparency. Why do we need to know how someone voted for an RfC or an RfC? Do we also need to know their name, address and phone number for transparency purposes? Do we also need to know about their medical history or sexual preferences? This isn't transparency, it's intrusion.

  Let's examine transparency a bit further. User:(a) orders User:(b) to vote for Candidate:(x) with threats of off-wiki outing. (And, yes, there really are people like this). What is transparent about User:(b)'s vote?  Roger Davies ^talk 19:33, 2 November 2009 (UTC)

• I'll step past your attempt to paint my position as nefarious by analogy. When I speak about transparency I refer mainly to the selection of candidates from among a pool. That mechanism needs to be trusted in some sense. If we had no past problems with less than transparent power structures (as it were), then I would agree with the general argument that trust in a vote counting mechanism is sufficient and would require no external observation. But we don't have a situation like that. We have a situation where a credible accusation could be leveled that election results have been tampered with in favor or in opposition to a particular candidate. In an open system, the election administrator can point to X votes tallied for candidate A and some agreement can be made that those votes are invalid for a particular reason. All (almost) information is shared among parties. In the securepoll system I can see who has voted (and when), but not for whom. So lets imagine that some block of votes is deemed 'invalid' by an election administrator. We then only have their word that the votes were for the candidate that is being credited w/ the changed tallies. The only assurances that can (or will) be made are similar to RFCU results--they will reveal nothing and piss everyone off (since revealing enough to satisfy everyone would violate the privacy of the voters).
• More importantly, I see plenty of analogies to advanced western democracies and what-not. Please keep in mind that wikipedia isn't really one of those democracies. We are on a transition from a small community where 'voting' like this would be burdensome and divisive (hence !vote and
  WP:NOT#DEM). Obviously we have discovered that discussion and collaboration do not scale. We also have a lot of imported norms about fairness and representation that would make it unacceptable to draw a group like Arbcom from the community through a discussion-like process. So we have this vote which is neither fish nor fowl. As such, arguments related to moving along the continuum from discussion to representation will appeal to a cluttered field of norms and principles. We have to be aware of this and make our decision based on the community we have (unfortunately invoking Rumsfeld). If our community has an image problem with transparency in decisionmaking then we better damn well be sensitive to that when making decisions. We can't just pick what seems to be the 'best' system for the ideal governance structure and apply it here. Protonk (talk
  ) 22:05, 2 November 2009 (UTC)

• I'm not sure that one's ArbCom election voting record being used against one in RfAs and RfCs is anything to be worried about, either. "As evidence of this user's poor judgment, they voted for User:BadFaithEditor in ArbCom election 2009"-style comments aren't particularly respected (or common) in my experience. -kotra (talk) 20:17, 2 November 2009 (UTC)

• That's the unfortunate nature of disruption like this. It neither needs to be common nor respected in order to provoke an outsized response or deter participation. Arguably it is more common than outright harassment. Protonk (talk) 22:09, 2 November 2009 (UTC)

• I don't know, if people are really that timid they probably have no business editing a public wiki, where everything, even something as innocuous as warning a vandal, has an equally likely chance of being "used against them" in an RfC or RfA. If people are really put off by potential comments such as these, I'd imagine editing a public wiki at all would be too intimidating. -kotra (talk) 17:58, 3 November 2009 (UTC)

• I mean, don't get me wrong, I think that the benefits of secrecy aren't sufficient to outweight the costs, but I'm not comfortable making the 'get a thicker skin' argument. Ideally we would want people to edit the wiki productively, anonymously and innocuously. I don't want people who edit like that to be disenfranchised because only the drama-immune (or drama-philic) will participate in governance. And I don't think we are in a healthy position to layer more shit on top of the already formidable collection of barriers to entry for Wikipedians. Protonk (talk) 19:27, 3 November 2009 (UTC)

• My point is that such "bad editor, they voted for X and s/he is bad"-style comments are just as rare and disrespected by the community as "bad editor, they warned X for vandalism and s/he is good"-style comments. If someone is intimidated by the potential for one sort of comment, they will probably also be intimidated by the potential for the other sort. The solution isn't to move potentially criticizable actions into secrecy (if we did so consistently, we would not have a wiki), it's to discourage and deal with the actual harassment. I don't think people should get a thicker skin, I just think people with that thin a skin probably aren't actively editing anyway. -kotra (talk) 20:55, 3 November 2009 (UTC)

Socks

More than a few people have opposed this because of fear that socks will be harder to spot. I really don't understand this. With SecurePoll the scrutineers can see everyone's IP and user agent info and compare them, finding socks that would never be with our current system. The list of voters is available as before for anyone to check up on. It seems that nothing is lost, only gained. — Jake Wartenberg 04:24, 31 October 2009 (UTC)

See, for instance, Special:SecurePoll/list/60. Everyone sees the username, timestamp and domain, election admins also see CheckUser-like data. The software also keeps track of cookie collisions. Happy‑melon 13:16, 31 October 2009 (UTC)

These technical additions are good, but secret voting still takes away one crucial way of detecting socks: comparing who votes for who, and when. For example, in the current system, one could see the following very obvious voting pattern:

1. kotra 11:01, 22 October 2009 (UTC)
2. Jimmy Smith 10:39, 27 October 2009 (UTC)
3. JamesSmith 10:41, 27 October 2009 (UTC)
4. JimS 10:42, 27 October 2009 (UTC)
5. Smith J 10:43, 27 October 2009 (UTC)
6. WikiJimmy 10:44, 27 October 2009 (UTC)
7. JS2009 10:46, 27 October 2009 (UTC)
8. Jake Wartenberg 19:39, 31 October 2009 (UTC)

This would certainly arouse suspicion and investigation. However, with secret voting, instead we would have a list of 300 or so users, not grouped by who they voted for or when (I assume), but randomly (or alphabetically perhaps). One might be able to identify a couple of the most obvious ones (like some of these) just by looking through an unordered list of names, but comparing who votes the same (and, to a lesser extent, when) is a crucial part of identifying socks. -kotra (talk) 19:43, 31 October 2009 (UTC)

As you could not possibly have failed to see on the page I linked to, the voting log is ordered by timestamp; it can also be sorted by user (or domain, not that that's useful here). I'm not sure how I can explain that better than by providing such an explicit example.

You see that list as a table almost exactly like your example, merely without the indication of which way they voted; the pattern is already obvious, as you say. I and the other election scrutineers see that page as a table of IP addresses and User Agents: the voting pattern you describe above would shine out like Time Square no matter how the accounts were named. No need to painstakingly reconstruct the data from a dozen separate checkuser queries; no need to see which way their votes were cast. You're saying that you would like election scrutineers to base their decision on whether to strike a vote on who the user voted for. Isn't that actually an incredibly dangerous approach to be taking? It is simply mistaken to think that with SecurePoll, only the tiny group of scrutineers has the ability to look for patterns of this kind; they merely have a more powerful 'X-ray vision'. Everyone is still involved in identifying sock trends like the example you provide, in either system. With SecurePoll, one piece of data has been removed from that search. I think some serious thought needs to be given as to whether it was ethical or objective to be using that information in the first place. Happy‑melon 00:09, 1 November 2009 (UTC)

Another important point is that this is better at ferreting out socks fairly. The election scrutineers can detect socking easier than with an open election but without knowing how they voted; removing any possibility that they "accidentally" overlook socking favoring a candidate they support. I doubt that's a likely scenario, but it still is additional fairness and control over and above what "open" voting could do. — Coren ^(talk) 00:38, 1 November 2009 (UTC)

I don't think so. In an open election, every side in a controversial elections will have motivated people to go through the vote lists for patterns and socks. For uncontroversial elections, on the other hand, a few socks do not matter. With a closed election, you rely on volunteers for a boring and unrewarding job. --Stephan Schulz (talk) 16:35, 1 November 2009 (UTC)

My apologies, Happy-melon, in my carelessness I skipped over the link you made note of, and didn't notice you mentioned "timestamp" in the list of criteria. Yes, you are correct then that the example I gave would probably still be easily identifiable since they were grouped within minutes of each other. However, what if the sock votes were more spaced out in time? Then what information could a normal user go by to identify the socks? In the current system, the grouping by what candidate one voted for is probably the most useful criteria for identifying socks. If one saw, for example, "FriendlyEditor" and "WikiFriend" vote for the same candidate, a day apart, one might be suspicious and investigate. If, however, they were separated in an alphabetical or chronological list of 300 or so (like the one you linked), one would probably not even notice these two similar names, and if one did, they would probably not raise any alarms. With a watered-down list like that linked above, I do not think normal editors would have enough information to successfully investigate any but the most incredibly blatant socks themselves, or as Stephan Schulz mentions above, would have any interest in trying when no actual impact on the election could be seen. For one reason or the other, we would find ourselves relying solely on a few election coordinators with checkuserlike access to identify socks, and we know from experience that the more eyes hunting out socks, the better. To be clear: the additional tools in SecurePoll you mention (IP address and User Agents) do certainly sound useful, and by themselves are an improvement over the current system. However, the rest of us can help with sock identifying too, and without groupings by vote, there is very little motivation or ability to do so. You wonder if it is an ethical criteria to search by. I can see why you ask: it allows people to "target" particular candidates for greater scrutiny. However, isn't every sock identified a good thing? If it means more people scrutinizing, and more socks identified (and I believe it does), I would certainly call it ethical. -kotra (talk) 09:06, 2 November 2009 (UTC)

But what is the impact of potential socking? If this year's turnout is comparable to last year's, you'd have to run dozens of socks to cause even a tiny blip. While how people voted remains secret, their editing contributions don't. You can still look at their editing history to see whether they ran up their 150 main space with trivial edits in one afternoon just before the qualifying period ended; whether they've ever contributed to WP:RFA, WP:AFD; whether they're SPAs etc. It takes time and effort to create a credible sock with 150 main space edits and an enormous amount of work to do this on a sufficient scale to make any real impact.

The other major point, which is being overlooked, is that ballot secrecy cuts both ways. Just as the community don't know who voted for whom, the sockmaster - because running tallies are not available to him - will have no idea how many socks to throw into the fray to influence the outcome. This creates the potential for huge amounts of work for zero gain and is hugely dissuasive.  Roger Davies ^talk 10:57, 2 November 2009 (UTC)

User:Scibaby is up at over 400 confirmed socks, plus 120 suspected ones. "Dozens" is not a large number. --Stephan Schulz (talk) 12:09, 2 November 2009 (UTC)

Most of whose socks have less than a dozen edits. It's not the number of socks that's important; it's the work involved in getting them up to 150 main space edits to meet franchise and keeping them clean while doing it.  Roger Davies ^talk 13:12, 2 November 2009 (UTC)

Remember that voters are voting on all the candidates simultaneously in a SecurePoll election; we talk of "3,000 votes" as 150 people voting for 20 candidates, but SecurePoll would see that solely as the 150 votes, 200 if you include corrections. The voter list will not be significantly longer than the open voting page would be in a wikipage election; if it is, it's because more people have got off their arses and voted, not because the voter lists have been combined. Socks identifiable on username or timestamp will, I am confident, be as easily identified in SecurePoll as in a wikipage election. Socks identifiable by IP, UA, cookie collision, etc, will be infinitely more easily caught. And, particularly at Stephen Schulz, it is completly disingenuous to say that the work of ferreting out socks rests solely with the scrutineers. The only data that the scrutineers have that every other user does not is the CheckUser-like information; in every other respect, every user is empowered to look and search for socks in exactly the same way. That is absolutely no different to a wikipage election.

There are quite a few improvements that need to be made to the voter list, such as linking the usernames, removing the useless "details" link for non-admins, and adding some more useful links. If anyone has any other suggestions, please let me know. Happy‑melon 09:33, 2 November 2009 (UTC)

You miss my point. I'm not saying that general users cannot go look for socks (although, of course, it becomes harder, as socks will typically vote en-bloc, and you don't see the votes anymore). I'm saying that it is unlikely that the general user population (i.e. volunteers) will go through random list of users trying to find socks (based on what kind of behaviour? Just the names and the times are available...) if they are not motivated by some tangible motivation ("that socking asshole will not get elected if I can stop him"). --Stephan Schulz (talk) 12:14, 2 November 2009 (UTC)

I'm not sure that "I'm going to go through that candidate's votes with a fine toothcomb until I find a reason he shouldn't be elected" is a motivation I like seeing in scrutineers... Happy‑melon 13:08, 2 November 2009 (UTC)

Why not? It's much easier to do with a single alpha-sorted voters list. And wouldn't you rather sockfinders were reasonably impartial rather than just looking for ways to improve their favourite candidate's chances?  Roger Davies ^talk 13:13, 2 November 2009 (UTC)

If the problem is herding cats, one doesn't solve the problem by making the herd larger. The experimental expansion of ArbCom last year was a failure. This 2009 Committee has had difficulty achieving quorum within a reasonable time frame (the Scientology case was the longest in site history) and has repeatedly attempted to extend its functions outside its core mandate while not handling that core mandate very well: the attempted RfC on content dispute resolution and the attempt to set up a broad-ranging advisory committee by fiat are two examples. A smaller and leaner ArbCom would have an easier time garnering enough members to read evidence and vote upon remedies in arbitration cases. For other accessory functions the Committee ought to help setting up independent bodies to handle matters, such as reviews of community bans under primarily community-based control. Durova³⁴⁹ 20:04, 30 October 2009 (UTC)

A smaller ArbCom would also mean that it's much prone to inactivity distributing too much work on too few people. A large ArbCom could for example decide to work as two semi-independent bodies, each in turn working on half the cases with all members only working together when deciding to accept a case or if further input is requested by one of the bodies. That would make ArbCom consist of two "lean and small" entities, thus sharing the workload among more people without those problems you mention. Regards SoWhy 20:15, 30 October 2009 (UTC)

That's exactly why they should be delegating. Instead they've been neglecting the core function while making aggressive moves toward becoming a governing body. A smaller Committee wouldn't be in a position to attempt to convert itself into an oligarchy. Durova³⁴⁹ 20:22, 30 October 2009 (UTC)

I don't really think 15 is small. 12 is still big, but it's the size of an Anglo-American jury. Deacon of Pndapetzim (Talk) 21:08, 30 October 2009 (UTC)

In this case, the herd size is beyond the control of the herders. A smaller and leaner ArbCom wouldnt have time to answer all the emails that they are sent, the number of which seems to keep growing. Delegation and devolution will help. A simple step would be for BASC to have its own email address, and a distinct team of volunteers, who present well researched appeals to the committee for decision. John Vandenberg ^(chat) 22:48, 30 October 2009 (UTC)

Delegation yes, but when on the occasions when the current ArbCom has actually delegated it tends to do so either within its own membership or to create bodies under its own direct appointment and control--even when no pressing need exists for that type of top-down setup. In a word, oligarchy. A smaller ArbCom couldn't hold all those strings, and would have to hand things over to community management. That'd be a healthy change of course from this year. BASC is a perfect example: why couldn't the community have set up that body and elected the membership? Durova³⁴⁹ 22:58, 30 October 2009 (UTC)

Work can be thinned out without similarly thinning accountability, and making elections so big the quality cannot be monitored by the community. I'm sure a large proportion of emails could be handled by trusted secretaries. Regarding all the cases, why do we need all those arbs reading all that evidence when others picked for their skill could be summarising it? As long as each elected arb is accountable for the mistakes of themselves and their delegates, everything should be fine democracy-wise. But simply increasing arb numbers commits us to the

law of diminishing returns

.

As many arbs take vacations from arbing, a better solution than spamming arbs is to have deputy arbs who'll fill in when one arb goes on break. These guys could be elected separately, elected with the arb (like the P/VP combos of US elections), or even just appointed ad hoc (though I don't much like that idea). Alternatively, a stand-by arb list could be taken from the most successful unsuccessful candidates, or perhaps in emergencies a "senate" of ex-arbs could appoint fill-ins (if an arb and a deputy were both on vacation). Deacon of Pndapetzim (Talk) 23:33, 30 October 2009 (UTC)

You make me smile, Deacon. Trusted secretaries? And how would we be able to appoint such individuals, who would all have access to exactly the same private information as arbitrators, without massive community uproar? We're still getting complaints about the Functionaries mailing list, and all of those people are Foundation-identified. Mind you, it would be fantastic to have someone fill exactly that role, ensuring that a response is sent out, the email is logged, it's addressed in the proper queue, linked in the right place on the arbwiki, etc. Risker (talk) 23:42, 30 October 2009 (UTC)

"Trust" can be community based, it doesn't have to come merely from the arbitrators. Arbitrators aren't that inherently trustworthy or untrustworthy anyway, as after all the only things that separate them from an IP is a history on wikipedia, popularity, and [now] rl-identity logged with the Foundation. Why's that so hard to appoint? Deacon of Pndapetzim (Talk) 23:53, 30 October 2009 (UTC)

I think the correct solution is to have a large ArbCom to ensure that there are enough Arbs at hand at all times, but have the Arbs divide the work between themselves, so that there will not be too many Arbs active on a single case. It seems that currently way too many Arbs have long periods of inactivity, probably because we are giving too much weight on how important real-life jobs the candidates have, and too little on how much time they are actually able to spend on WP. I think it might be best if we had 1-3 Arbs per case and having each Arb active only on one case at a time, so that they can concentrate fully on the task.

Adding members to a committee is an inefficient way to improve its throughput. I like the idea of two committees. It's analogous to two courts sitting in the same building. Two committees would also serve to reinforce the notion of ArbCom as integral part of the dispute resolution process, rather than a single body sitting at the apex of a power structure. MoreThings (talk) 00:51, 31 October 2009 (UTC)

One of my votes was deleted (likely an accidental edit conflict, error, or something), but in examining the page quickly, I found at least two other problems: under "Term lengths", Elonka and Hypocrite have "voted" twice. Is someone monitoring the page ? [4] SandyGeorgia (Talk) 01:24, 31 October 2009 (UTC)

That does look like an edit conflict. On the second matter, the instructions read "Sign multiple options if more than one is desirable to you." It might have been better phrased, but the leeway to support more than one alternative is there.

barbarian 

01:35, 31 October 2009 (UTC)

As Skomorokh says, the option to choose cast multiple support votes is available. A quick scan indicates a few others have done the same. Manning (talk) 01:59, 31 October 2009 (UTC)

OK, thanks (I missed that :) SandyGeorgia (Talk) 02:15, 31 October 2009 (UTC)

But, Hypocrites "vote" under "18 months" looks to be misplaced, since it actually mentions "Tactical voting". SandyGeorgia (Talk) 02:18, 31 October 2009 (UTC)

No it's in the right place. His actual statement says "Tactical voting because 2 years is too long and 1 year is optimal but 1 year is unlikely to have support". He is indicating he wants the "one year" option, but he is also voting for "18 months" because the "one year" option is unlikely to pass. It's a tactical vote against the "two year" option. There's no problem with any of it. Cheers Manning (talk) 02:31, 31 October 2009 (UTC)

Oh, my; ok, thanks again ! SandyGeorgia (Talk) 02:32, 31 October 2009 (UTC)

This is probably a dumb question...

But is there a type of voting which includes the cumulative method and yet also allows for opposing?

So basically: Support/Neutral/Oppose. But only can support a number of candidates up to the number of seats available.

That's what I would like best, I think. - jc37 02:05, 31 October 2009 (UTC)

Actually that sounds simple enough to explain. I'll just add it as an option. - jc37 03:55, 31 October 2009 (UTC)

An important element of cumulative voting is each voter only having a fixed number of votes that they can distribute across the options.

In my opinion, only allowing positive votes is another important element, because it means that there is no support/oppose percentage. A support/oppose percentage isnt very meaningful when each candidate has a different number of voters. In the current AUSC election, we have a neutral option for each candidate, which is in effect an option to not opine. As a result, when the AUSC election results are published, it is possible that a candidate who received many abstain results can rank quite highly when sorted by ratio of support:oppose.

Irrespective of what method we choose, I hope that the "cost" of negative votes is increased, as 2008 saw many voters switching from support to oppose at the end of the election period, because it was an easy way to radically affect the result. Then others would switch to offset this swing, and it all became a bit crazy for the closely tied candidates. In addition to this, the differing number of voters for each candidate played a big part. I had the most votes (for+against), so a switch on my vote page resulted in a +0.15 or -0.2% swing, whereas a switch for Coren resulted in a +/-0.4% swing.

Driving the election by percentages, and the resulting switches, is another reason why I don't like public voting; the votes towards the end are distorted by the running tally, and become a fight within the community.

That said, negative votes could be incorporated into cumulative voting if they are still considered to be a vote, reducing the number of positive votes that the voter can throw.

John Vandenberg ^(chat) 04:03, 31 October 2009 (UTC)

You're looking at it from the cumulative point of view.

I'm looking at it from the reverse.

I'm suggesting that we use the Support/neutral/oppose method, but limit the number of supports any individual can vote to the number of seats. No more, but also no less.

Among other things, this would help stop the "I oppose arbcom" votes who vote against everyone, and those of the reverse perspective who vote for everyone. It would also help stem the tide of "tactical voting".

And so this forces everyone who wants to vote to ACTUALLY vote for someone. And perhaps even nudge them in the direction of "Well if I have to vote for someone, then maybe I should learn more about the choices."

I would presume that making such a limitation in the software would be trivial? - jc37 09:30, 31 October 2009 (UTC)

Oh, and it occurs to me that this could also deal with the open/secret question too. Since you're limited in the number of supports you can give per person, then each person's oppose votes lose more than a bit of the potential stigma that others are concerned about. So having it "open" would be less of a problem. - jc37 09:37, 31 October 2009 (UTC)

(e/c)Thanks for clarifying; I had misunderstood. However I am still a little bit confused. You say it will stop the "I oppose arbcom" votes - will your election method also be limiting the number of oppose vote?

A lot of what I am saying is against any system that includes an oppose vote, as that introduces a ratio of supports:opposes. If the gods promised (before the election) to use the Net rather than the percentage, and the ballot was secret, I wouldn't mind "oppose" votes. John Vandenberg ^(chat) 10:11, 31 October 2009 (UTC)

Seriously, why change what worked perfectly well (with regards to how we vote)? Majorly talk 17:43, 30 October 2009 (UTC)

Because a serious argument can (and has) been made that "worked perfectly" doesn't describe reality. There was tactical voting, people who did not vote (or voted against their convictions) because of fear of retribution, people who did not oppose a candidate that would "clearly" win but that they felt shouldn't get a seat for fear that this would predispose that arb against them, people who voted against a candidate they felt would make a good arb but was unpopular for fear of being ridiculed for their position.

Those aren't hypotheticals: I think every sitting arb has at least a few cases like this they know of personally, not counting any that have not spoken up. — Coren ^(talk) 13:03, 31 October 2009 (UTC)

Endorse Coren's view. SandyGeorgia (Talk) 13:07, 31 October 2009 (UTC)

Secret voting brings on new problems too. Nothing is being solved here. Tactical voting will always occur, and it will be impossible to see if we don't know how people vote. If people are going to get back at other editors for voting in the "wrong" way, they should be removed immediately. Also, if people don't bother voting for whatever reason, that's their loss. Wikipedia is not the real world, and it's not a democracy. Whatever happened to openness and honesty? Perhaps you'll be saying next that RFAs should be held in secret - so that no one can see arguments brought up - or perhaps, FAC in secret. Because the same arguments you apply to secret polling for arbcom could work just as easily for both those things. There are arguments against though. But let's remember - it's a website encyclopedia, not a political game. Game players should be removed immediately. If admins/arbcom were tougher on such people, there would be no such thing as fearing retribution. Majorly talk 14:42, 31 October 2009 (UTC)

The main question here is how many editors are dissuaded from voting in ArbCom elections because of the combative atmosphere; the time-overhead of making loads of little edits to vote; and the risk of bellgerent challenge? Given the ridiculously low turn outs (less than 1000 voters at ACE2008), my guess is hundreds if not thousands. These people are effectively disenfranchised by a public voting system that positively discriminates in favour of the self-confident, the assertive and the vociferous.

In sharp contrast, voting via SecurePoll takes seconds; no-one screams at you for how you've voted, stridently demands a rationale, or expresses disbelief that you've voted the way you did.

Fear of cabalism is a lousy reason to stick with what we've got. There are loads of ways to game public votes and only the clumiest, most inept, cabals who are stupid enough to vote too narrowly or too pointedly get caught. More importantly, if we manage to significantly increase the number of people voting, we radically reduce the influence of cabals.

At its best, the way of the Wiki is wonderful, with collegiate discussions and consensus gradually emerging by genial give-and-take debate. Unfortunately, this doesn't scale so what we end up with is voting against an ever-present background of torches and pitchforks with much of the agenda set by wannabe rabble-rousers. Civilised society abandoned public voting a couple of hundred years ago for precisely these reasons. It is now our turn to gingerly step into the 21st century.  Roger Davies ^talk 15:08, 31 October 2009 (UTC)

I have always been seriously afraid of lodging Opposes in open polling, and I'm no shrinking violet; I suspect the Oppose tally will go up significantly with secret polling, and my concern is that the Support percentages will be low under secret balloting. Some of the "endorsement" pages may spring up because editors feel a need to justify opposes with strong rationale: mine did. I'll lodge a lot more opposes when I don't have to beat around the bush to avoid saying, "I oppose this candidate because he operates as part of a back-channel cabal, shows up at places he rarely frequents with sloppy commentary and assumptions of bad faith based on back-channel gossip and little research or understanding of the issues, is unlikely to diligently research and understand cases before ArbCom, and may divulge confidential ArbCom information", or "this candidate participates in canvassing and backchannel horse-trading". My Supports, however, are unlikely to change, because they were always based on candidate integrity. SandyGeorgia (Talk) 16:21, 31 October 2009 (UTC)

With secret voting, you can still make these points about the candidate, in the "Comments" section of their election sub-page. That is probably a better environment for reasoned discussion, especially as then you needn't explicitly state whether you'll be supporting or opposing. It also gives the candidate an opportunity to respond. Under the public vote system, of course, candidates are often perceived as aggressive if they comment on votes, even when the voting comment is way off mark or factually inaccurate.  Roger Davies ^talk 16:41, 31 October 2009 (UTC)

I can't and won't make the points anywhere, when raising these concerns might result in retaliation or when they're based on confidential and private e-mail; I hope my voting record stands for itself in terms of candidate integrity. I'm just saying that fear of retaliation does prevent editors from lodging (more) opposes. SandyGeorgia (Talk) 16:49, 31 October 2009 (UTC)

Agree with Roger here. Sometimes views on a candidate are unfairly coloured by remarks made that the candidate feels unable to properly respond to. I had that feeling in the election last year. There were several opposes lodged by people who I respect for their long experience and clue, but I felt unable to adequately respond because I felt that might be seen as haranguing someone for their view on my candidacy. If someone was merely giving opinions, without explicitly stating if they were opposing or supporting, I would have felt much happier debating issues with them. Carcharoth (talk) 17:12, 1 November 2009 (UTC)

Briefly, on Majorly's point about FAC, I have thought at times that those carrying out reviews should be able to do so anonymously. It avoids unnecessary personalisation of issues at FAC. If someone has a valid point about an article, it should be raised and discussed. What that would need, though, is a way of verifying whether anonymous reviewers had edited an article or not. As that isn't easy to do, the idea hasn't really got anywhere. And it's off-topic here anyway. Carcharoth (talk) 17:17, 1 November 2009 (UTC) In fact, so off-topic, please don't reply here. If anyone wants to comment on this, or thinks this is too off-topic, I'm happy to remove this and take it to my talk page or the FAC talk page. Carcharoth (talk) 17:18, 1 November 2009 (UTC)

• I don't wish to belabour this point. I do wish to note that I am in almost complete agreement with
  *
  18:36, 31 October 2009 (UTC)

• BTW, FAC is not a vote; one well-placed Oppose can counter 15 Supports. SandyGeorgia (Talk) 19:41, 31 October 2009 (UTC)

• Mr Casey: I suppose the key point is that the ArbCom election is a vote (which is why it's called an "election") and not a consensus-driven exercise. The notion that what (allegedly) works well at RfA scales to 30+ candidates and 1000 voters is, frankly, utopian (not, of course, that there's anything wrong with idealism).

  I don't know where this idea that SecurePoll will politicise the election comes from. ArbCom elections are already riven by politics and the effect of SecurePoll will, if anything, depoliticise the elections by weakening the demagogues. The Ordinary Joe, for the first time, will be empowered and enfranchised, able to vote according to his conscience, without fear of humiliation, challenge or reprisal.  Roger Davies ^talk 19:12, 31 October 2009 (UTC)

ArbCom elections are not a pseudo-vote like RFA/RFB. At RFA, a rationale is expected for opposes or the comment may be discounted during closing. ArbCom elections do not have this expectation, extended comments in the voting section are moved to another page. It isn't a pseudo-vote, its just a vote. Mr.Z-man 19:25, 31 October 2009 (UTC)

Then all the less reason for anyone to demand a rationale. Majorly talk 19:30, 31 October 2009 (UTC)

But many editors do. And become, um, vociferous if an explanation isn't provided. (See SandyGeorgia's comments above).  Roger Davies ^talk 19:35, 31 October 2009 (UTC)

If they demand a rationale, ignore them if you don't wish to give one. If they start engaging in harassment or other inappropriate behaviour, report it and they can be removed from the project. SandyGeorgia has not provided any examples, only her apparent fear of people retaliating because of her opposing people. Seriously, if people are that childish they feel the need to retaliate, they have no place on Wikipedia. Majorly talk 19:38, 31 October 2009 (UTC)

You are proving my point. Badgering of voters takes place: this is not in dispute and it is a major concern for people who have no desire to participate in a drama-fest. They vote with their feet and stay away. It is beyond dispute that the belligerent atmosphere that sometimes prevails therefore disenfranchises many editors: this is why the system must change.

What's more, you have yet to provide one single reason why you should know how I choose to cast my votes. This isn't transparency; it's an invasion of my privacy.  Roger Davies ^talk 19:47, 31 October 2009 (UTC)

Perhaps you are correct that my viewpoint is idealistic. I want to see Wikipedia remain free of politics and contention. Sometimes I wonder whether the arbitrators see so much strife in their assignment that they inadvertently become a bit cynical or jaded. I wonder whether "empowering and enfranchising" the humble masses is a bit... grandiose... for an encyclopedia-writing project. But on a more pragmatic note, I look forward to seeing how the secrecy program works out. Perhaps it will work very well... if so, I fear that there will be attempts to introduce voting in other areas of WP... but, in the end, the encyclopedia project will continue! —

*

19:51, 31 October 2009 (UTC)

Remain free of politics? LOL. I think our current ArbCom is well in touch with the community and the humble masses. SandyGeorgia (Talk) 19:55, 31 October 2009 (UTC)

I honestly find it difficult to accept the picture of this place painted above by Coren and Roger. Is it really that bad?

Your arguments seem predicated on the idea that WP is seething mass of deals, double-deals, coercion and cabalism. Sure that's the way it's painted in many a disgruntled talk-page edit, and it's a fun idea, but do you honestly believe that is the reality for most editors? I don't think it's credible to claim that hundreds or thousands of editors are dissuaded from voting for fear of someone penning a tough response to their vote. That's not the way editors are. And, anyway, looking back at the last election, I didn't see much in the way of screams or strident demands, and you didn't provide any examples.

Are most people not simply going to read the statements, do a bit of research, and take things at face-value? Also, there are people around here whose views I respect, and who seem to have their finger on the pulse; I would find it useful to know which way they are voting. Finally, there is a danger that secret voting will actually make it easier to install a puppet candidate. Suppose a large number of partisan editors from one of the hot-spots vote for one candidate. With public voting, that is immediately going to raise questions. With secret voting nobody will by any the wiser. And if we go for secret voting along with one of the systems which makes it possible to cast multiple votes for a single candidate, the problem will be compounded. MoreThings (talk) 21:32, 31 October 2009 (UTC)

@Roger: On a website like Wikipedia, there is no such thing as privacy (in this respect). Every edit is recorded, every log. When things start to go on behind closed doors unnecessarily, we have issues. When people's opinions are asked for, the community has the right to know what the opinions are. You have not provided a single reason why we shouldn't know how you vote. After all, we can see for RFAs, RFBs, AFDs, FACs etc etc. Everything except this one thing, where it suddenly becomes a problem. I also agree with MoreThings. Not a single example of such has been provided, other than speculation. People turned out in their hundreds to vote, so it's not like most people are afraid. It is often the case I see somebody's vote and change my mind precisely because of what they say, that I missed. Keeping it all hush hush for "privacy" purposes is nonsense to me, when we've been doing it this way for so long without any actual problems. Theoretical whatifs don't count for anything here. Majorly talk 22:35, 31 October 2009 (UTC)

• A clarification. I don't believe there has been any retribution done by arbs to editors who voted against them. I certainly never witnessed any taking place. But the fact remains that editors fear such retribution (or, more realistically perhaps, that arbs would then be unwittingly biased); and that those worries discourages frank voting — even if they turn out to be unwarranted. — Coren ^(talk) 23:22, 31 October 2009 (UTC)

No, but in the previous RfC you said "intimidation, retribution, vote buying, and tactical voting" have been observed. I have to say my immediate thought was: if there really is evidence of that, what has been done about it?. It sounds as though you are describing activities that would contravene policy.

Similarly here, there's much talk of people being afraid to vote, but not much evidence. If an editor visits the voting page, votes, and then moves on, it's difficult to see how anyone is going to "scream" at them without following them to their talk page. I really doubt that would happen, and I'm sure if it did it would be stamped on quickly and heavily.

And I don't really buy the argument that people are afraid to vote against an arb for fear of retribution. People understand that an arb is not going to think right I'm gonna get this guy because he voted against me, that's just silly. Same for unwittingly biased. Fine phrase, but I think most people know when they're feeling biased, and they know how to put that to one side, or to recuse.

I haven't really followed the mailing-list case, but it's not too difficult to imagine a determined group of editors coming together to support a particular non-neutral candidate. That kind of thing is lot easier to hide when the voting is secret. I think that's a much more serious concern than this business about being afraid to vote. MoreThings (talk) 00:56, 1 November 2009 (UTC)

Oh, I didn't mean intimidation by arbitrators. But we have received email from editors stating, quite explicitly, that they were surprised we handled some matter in such-or-such way despite they having voted against the arb handling it. That makes it quite clear that retribution or unfairness is expected or feared by some editors— but we'll never hear about those who downright didn't vote because of those apprehensions.

Likewise, I've received email from editors to outright they didn't vote, or changed their votes, because they had been browbeaten into it by prominent editors. I've personally heard from at least six such editors; and I know other arbs have received similar feedback. Those are only the ones that spoke up.

Then there are known case of harassment for specific votes; people who voted for or against some candidates were accused of things such as "hating Wikipedia", or being "trolls". It's not a stretch to expect that at least some editors would have refrained from casting their votes to avoid such treatment. — Coren ^(talk) 01:31, 1 November 2009 (UTC)

I appreciate that you didn't mean intimidation by arbs. My point is that I'd have thought evidence of "intimidation, retribution, vote buying, and tactical voting" would have resulted in some kind of action no matter who was doing the intimidating.

For the rest, it's anecdotal. I'm not privy to the e-mails. I've looked at the voting pages from last time, and I didn't see much evidence of editors being browbeaten. If it were as widespread as is suggested in the various posts above, I'd have thought it would be apparent in the voting pages.

My instinct is that the great majority of editors would not allow themselves to be browbeaten into anything, and would find the very thought preposterous. I can't think of too many who would be intimidated by being called a troll. That epithet is thrown around here so often it's almost a term of endearment.

All this just strikes me as too Machiavellian. For most people, using pseudonym to vote on a website is not activity that is likely to induce fear. But I can produce no more evidence for that pov, than you can for yours :) MoreThings (talk) 03:10, 1 November 2009 (UTC)

(undent) Yes, it's very much anecdotal. For all we know, those half dozen editors I know of may well be the only ones, and feedback received by my colleagues might be from the same set (we don't forward to each other mail addressed to one of us without an explicit request from the sender to do so, but we occasionally do discuss the general nature of the email when it's germane).

But my default position, in the absence of hard data, is to err on the side of safety: history has proven over and over again that open votes are gameable and easily manipulated whereas secret ballots are secure against most forms of abuse. They tend to be harder to handle fairly and cleanly, but we have the advantage of technical support in that area which the paper-in-person based forms of voting do not—offsetting that increased difficulty for the most part.

Now, it turns out that electronic voting overlaps neatly with my area of expertise (I do computer security and cryptography), and I can assure you that — on the technical front — SecurePoll's implementation is secure enough, and guaranteed fair, for our purposes (with the caveat that we have to trust the foundation sysadmins and the election admins to not collude to defraud the vote unless the cryptographic receipt feature is turned on; this may or may not be necessary). — Coren ^(talk) 03:52, 1 November 2009 (UTC)

Yes, but you're talking about the history of real world elections. If you have a guy with Kalashnikov "helping" you vote, or if knocks on the door at 3.00am are not unheard of, then secret voting is a very good idea.

This is not the real world. It's a web site. Despite the pitiful scene delineated above in which terrified editors engulfed by a climate of fear dared not venture out to the voting page, it all seems to boil down to a handful of incidences of "thanks for not holding my oppose vote against me" and "troll".

I have no concerns about the technology or the probity of the overseers. I'm sure both machine and man could do a great job. A concern I do have, one which you keep skipping, is the likelihood that secret voting will make genuine, concerted attempts at subversion easier rather than harder.

If there really is a problem, then creating an environment in which large groups of like-minded editors can club together, undetected, to vote for a single candidate does not seem like a good idea. If 30 Martians all vote for the same ostensibly neutral candidate then, with public voting, the Plutonians are immediately going raise concerns. At the very least they will ask that arb to recuse at the next Mars v Pluto case. With secret voting, they will have no idea. MoreThings (talk) 14:39, 1 November 2009 (UTC)

True, the magnitude and stakes are different, but the basic MO remains. While nobody here fear for their lives for voting the "wrong" way, the whole project relies on cooperation and reputation— attacks to either are proportionally just as destructive.

People on-wiki do act to perception, and to minimize the number of (real or imagined) "powerful enemies". They act—or avoid acting—trying to maintain the currency of the realm: reputation.

For instance, if one enjoys and works a great deal bringing articles to featured status and one of the most influent and visible editors working in the FA process was to run for a seat, most people will hesitate to oppose them publicly even if convinced they would make a terrible arb. (And I want to make clear here that I'm not pointing fingers; I just used FA as an example with nobody specific in mind). — Coren ^(talk) 15:13, 1 November 2009 (UTC)

I Dunno, I don't recognize this image that you keep painting of frightened, bullied, sheep-like editors, I guess we'll have to agree to disagree on this one. MoreThings (talk) 01:23, 2 November 2009 (UTC)

A couple of points responding to some of the above. Majorly said: "It is often the case I see somebody's vote and change my mind precisely because of what they say, that I missed" - in this system, you would still see what people said, and you could still change your mind after reading it (you can change your vote in SecurePoll any time before the election finishes), but that discussion would take place on the discussion page about the candidate. MoreThings said: "Suppose a large number of partisan editors from one of the hot-spots vote for one candidate. With public voting, that is immediately going to raise questions. With secret voting nobody will by any the wiser." With SecurePoll, you still see who has voted, just not who they voted for. So if a large number of votes from a partisan area of the project turn up, that can be identified by reviewing the voting list. You just won't know who they voted for. You can assume they voted against a particular candidate, and I'd actually be in support of limited data being revealed, such as how many votes each voter cast of each nature (voters that turn up and cast a single oppose vote, or a single support vote, or all oppose votes, or all support votes, would stick out like a sore thumb). Ditto for accounts that have been inactive for months and months, and turn up to vote, and then go inactive again. That could still all be shown while retaining anonymity. The other point is that it might be an idea to show (after the election) a timeline of the voting (I don't think the software can do it at the moment) showing trends. That way, it will be obvious if one candidate suddenly gets a massive upswing in support. Trouble is, if you publish voter timestamps, then you can match the voters to changes in the trends. If people think ditching when someone voted should be done in order to allow trends in voting to be shown after the election, that might be an idea for future elections, if the software can be tweaked to record that. Carcharoth (talk) 17:37, 1 November 2009 (UTC)

I don't really follow your refutation of my point. Take any of the hot-spots, or my silly Martian example above. Editors from both sides of the hot-spot will turn up and vote. Ordinarily, their votes would be distributed among candidates in in the same proportions as the rest of the community. Most editors from both sides of the hot-spot would vote for the favourites, a few would vote for the middle-runners and an odd one for an outsider. Now suppose that all of Martians voted for the same candidate—they vote for others, too, but there is one candidate whom they all support. In public voting, that would be very evident, and the Plutonians would notice it, but in secret voting nobody would know that they had voted that way.

I don't really see how your idea of revealing limited data would safeguard against that happening. They wouldn't vote only for their puppet candidate, they would cast other support and oppose votes too. And they wouldn't all turn up at exactly the same time to vote. It would be a fairly simple thing to hide. The only thing they would have in common is that they all voted for the same candidate, and there would be no way anyone could know that.

This does all seem a bit far-fetched, but I was quite surprised by the little I read of the mailing-list case. There probably are groups around here who would like a seat or two on ArbCom. And it certainly strikes me as less far-fetched than the picture of WP and its editors that has been painted in some of the posts above. If there is a danger, then I think it lies in opening the door to editors who collude off-wiki and will vote en-bloc in secret, not in editors being afraid to vote fore fear of someone posting a caustic riposte.MoreThings (talk) 01:23, 2 November 2009 (UTC)

You say above that "they would cast other support and oppose votes too" - but if you don't know who the favorites and front-runners are (there is no running total of the vote during the election, remember), then how would you do that? You would have to support your preferred candidate and oppose the rest, otherwise you risk unknowingly pushing another candidate in front of your preferred candidate. In other words, tactical voting (chosing your other votes to favour your preferred candidates) is only possibly in open ballots. Making it secret disposes of that element - you are forced to support only those you genuinely support. Carcharoth (talk) 01:43, 2 November 2009 (UTC)

You would make your best guess. Select the 6 or so candidates that you believe to be the outsiders. You all vote for your man and against those you have targeted. If you are a group of, say a couple of dozen, most of you have a sock or two, you could produce a significant swing. Even if the 6 turned out to be mid-table, there's a chance your man could leapfrog at least one of them. That kind of voting pattern would never be picked up by software, or overseers, but there's a reasonable chance that those on the other side of the hot-spot would notice it.

Many people seem to be comparing this vote to political elections. But in political elections candidates are avowedly partisan. They'll tell you: I'm going to lower taxes, cut City bonuses, devolve Cornwall, and bomb the French. If that set of prejudices matches your own, you support the candidate. Here, everyone is running on a platform of neutrality. These elections are much more akin to the citizenry electing a judiciary, and that doesn't happen anywhere in real world. In political elections, you are supposed to vote for the candidate who will argue your case; here we want the candidates to be entirely impartial, and we don't want allow any group to elect a candidate who would fight their corner.

I can understand the knee-jerk reaction to feel that secret is better, but I don't think it stands up to scrutiny. There are plenty of places where open voting is preferable. ArbCom votes, for example. RfA, for another. Think of parliaments, senates, and national assemblies, in those we all want to know how the representatives are voting. In a sense, we are all self-elected representatives. We set ourselves up as capable of writing this encyclopaedia and maintaining its integrity. I think we should be open with one another, and with the readership, and we should be prepared to stand by our views.

My fundamental point is that I feel voting should be public unless is there is a compelling reason for it to be secret. As far as I can see, the two main arguments for secrecy are ip logging and intimidation. I'd be in favour of using the software to count votes and log ip addresses, just not to hide the votes. I've not seen any evidence of intimidation, and not been convinced by the arguments above. I wonder if perhaps most of those making the intimidation argument worry that others might feel intimidated, while not feeling in the least intimidated themselves. That said, if there are those who genuinely do feel intimidated, then it's entirely understandable that they would vote for secrecy, and I wouldn't try to discourage them from doing so. MoreThings (talk) 23:23, 3 November 2009 (UTC)

@Carcharoth's point above: "in this system, you would still see what people said, and you could still change your mind after reading it (you can change your vote in SecurePoll any time before the election finishes)," followed by "With SecurePoll, you still see who has voted, just not who they voted for." These two statements seem contradictory to me: we can still see what people said, but not who they said it about? How does that help influence our decision? GlassCobra 20:37, 3 November 2009 (UTC)

Perhaps Carcharoth's point is that SecurePoll doesn't stop people commenting on the candidate's page. I imagine that they will be the seat of many lively debates, with the added advantage that the candidate can respond with risking antagonising those who believe that candidates shouldn't respond to comments attaching to votes, no matter how snide or inaccurate.

SecureVote also lets you switch votes, without drama. So if you change your mind, you can change your vote without it appearing pointy.  Roger Davies ^talk 20:52, 3 November 2009 (UTC)

As a project, we should embrace the transparency that comes with the wiki model as much as possible for the sake of dialogue and better understanding. Transparency is healthy for the wiki-culture, while secrecy undermines it—and we should consider the public or secret voting question in those terms. Additionally, sockpuppetry is a key problem we have to watch out for, and public voting facilitates the identification of socks. It's also worth noting that some of us don't have a whole lot of trust in the current ArbCom and Jimbo, and if the voting is done publicly then we can be sure that the results are accurate. Everyking (talk) 15:22, 31 October 2009 (UTC)

Well, I do trust the current Committee, but I am also strongly in favor of public voting. My rationale, as presented above, is that the decision to "secrify" this election will be the first step in divorcing Wikipedia from its traditional open atmosphere. —

*

18:38, 31 October 2009 (UTC)

It is incorrect, and totally disingenous, to say that a distrust of the current arbcom should manifest itself in a distrust of a SecurePoll election. The voting system is a component of MediaWiki, which is controlled by the WMF sysadmins. These people could alter the 'results' of a public election as easily as a SecurePoll one; they are placed above suspicion for our own sanity, because there is no way to protect the vote from them. The election admins for the AUSC election, comprise myself, as a technical observer, Roger Davies as the ArbCom observer, and three stewards as scrutineers. Your trust in the election's integrity rests, as far as it rests in any wiki user, in those stewards. They are trusted with the responsibility of striking votes that are deemed to be illegitimate, in the understanding that neither they, nor anyone else, know the direction of the votes they are striking. They have no ability to affect the election except by striking votes, no way to affect the election inappropriately except by striking votes inappropriately. For the ACE2009 votes, we can assume that a similarly-independent panel of scrutineers will be appointed, certainly it will not contain sitting ArbCom members. So within this framework, can you please clarify how, if we take an assumption that ArbCom is hopelessly partisan and is determined to commit complete electoral fraud, they are to have any chance of doing so? Happy‑melon 19:01, 31 October 2009 (UTC)

Transparency was a recurring theme in the first RFC. I don't think it is sufficient to say that a clutch of selected stewards or admins have visibility over the poll. Everyone needs to be able to satisfy themselves that the result of the poll was correct. In reply to these concerns first RFC I gave some thought to how transparency could be built into a method for secret balloting.

coṁrá

) 21:03, 1 November 2009 (UTC)

Transparency is a meme on Wikipedia - its repeated like a slogan, but hardly ever do you hear the actual, specific benefits of transparency expressed alongside the slogan. Transparency is often desirable, yes - common ideal in modern Western government, not at all new. So as a generic ideal, we can agree that transparency is good. That doesn't mean we should abandon the intellectual rigor involved in determining why it is good in specific context; transparency isn't good in covert operations, it isn't good in medical records, it isn't good in commercial product development or employment records or weapons development or advanced nuclear research. It's not good for checkuser data, the abuse filter, or oversighted revisions. An open, transparent revision history is good for collaboratively building an encyclopedia, but is it good in elections? That's a debate we can have, but let's keep in mind - Wikipedia didn't invent elections, and most election systems in the world seem to use a variation of secret balloting. The benefits of transparency don't automatically extend to all contexts; if you are going use it to argue against secret ballot elections, then let's have some specifics. For a "merits of transparency" section, so far such specifics are remarkably rare. Nathan ^T 17:27, 2 November 2009 (UTC)

"...most election systems in the world seem to use a variation of secret balloting." Most elections in the world are carried out though open voting. Very few elections elect presidents or members of parliament. Most appoint committee members to very mundane positions on local clubs and societies. Most have no need for a secret ballot. Those that do employ systems of running an elections that guard against the problem associated with the loss of transparency (particularly fraud). Once secrecy is introduced, the focus turns onto confidence in the electoral system. Real world elections (great and small) achieve this by issue voter lists and allowing (relatively open) access to supervise the casting and counting of ballots. Note how even in a "secret" ballot - transparency is regarded as important. --rannṗáirtí anaiṫnid (

coṁrá

) 23:37, 2 November 2009 (UTC)

In addition to being a meme, transparency is not policy. Interestingly, the bit of policy that does apply here -

assume good faith - is being massively ignored. So we're in the unusual position of seeing editors using a meme to (apparently) trump policy.  Roger Davies ^talk

21:09, 3 November 2009 (UTC)

How is AGF coming into this? If at all, it would lead us to not assume that people will use unfair pressure or elected Arbs will go through voter lists and keep track of who voted which way. --Stephan Schulz (talk) 22:02, 3 November 2009 (UTC)

One frequently given reason here for public polling is to prevent ArbCom and/or the devs fraudulently stuffing ballots. As there's no evidence whatsoever that this has happened in the past, it's an massive assumption of bad faith.

On the other hand, it's not in dispute that people are sometimes intimidated by public polling and that arbs can be influenced, even on a sub-conscious basis, by public votes for and against them: it's difficult to see how there is any assumption of bad faith involved in mentioning these things.  Roger Davies ^talk 22:29, 3 November 2009 (UTC)

I may have a different perspective of "frequent" than you have - it's certainly not an argument I have seen often, or that carries much weight with me. --Stephan Schulz (talk) 22:40, 3 November 2009 (UTC)

"it's not in dispute..." That is not entirely accurate. I do not believe that the concerns regarding an open ballot are as significant as some would present them, so their extent, if not their existence, is most certainly in what could be termed "dispute". I am persuaded that only a small number of editors are actually "intimidated by public polling". This is an encyclopedia - what is the source of this supposedly ubiquitous dread? Also, the idea that "arbs can be influenced" by apparently mystical and malignant "sub-conscious" forces is not something that is universally accepted. Secret voting is entirely unnecessary and is predicated on hypothetical evils not germane to the realities of a non-political project such as this one. —

*

02:21, 7 November 2009 (UTC)

You haven't addressed the issue. It is not an assumption of bad faith to suggest that people will be intimidated by the bear garden nature of public polling. In contrast, it is a massive assumption of bad faith to state that the voting needs to be public otherwise ArbCom and the developers will rig the results.  Roger Davies ^talk 22:51, 3 November 2009 (UTC)

It's not an expression of poor good faith to want evidence. One of our core policies is

WP:PROVEIT

.

If we are going to move to a new electoral system then we need to be sure of what we are doing - and we need to be able to discuss it freely. Confidence in a system is important. Expressing a desire for confidence in a system is not the same as casting aspersions on any real-life person.

We could simply "trust" everyone - but then why move to a secret ballot at all? (Inherent in the desire for a secret ballot is an expression of mistrust in others.) --rannṗáirtí anaiṫnid (

coṁrá

) 23:30, 3 November 2009 (UTC)

I'm very concerned over the option put up for "Approval voting (Support/Oppose)". As you would be able to tell from reading the wikilink to approval voting, the system proposed is explicitly not approval voting. The system proposed appears to be closer to the borda count with negative values, and has the same problems, most notably that it would be possible for someone with only a small minority of support being able to win. And that it disenfranchises those who do not vote tactically by making oppose votes for all other candidates even if they don't oppose them but want their preferred candidate to have a relative boost.

I might go as far to suggest that calling it "approval voting" may have mislead people into assuming it has the benefits and downsides of approval voting, instead of the benefits and downsides of borda count. And this would make the discussion on selection of the voting method tainted due to misidentification of what was being discussed and proposed.

In total, this system fails some pretty important

I also feel that there's been a grave flaw in that the suggested voting methods are all for Single-winner methods, and there doesn't appear to have been an attempt to use a

Good points. Can I suggest that whatever happens this year, that a continual discussion is held about this next year. Really. A year-long discussion starting in (say) February, that can really get to meat of some of these issues. It was proposed last year to have an election post-mortem in February 2009, but that never happened. By all means thrash out issues for these elections in the next few weeks, but anything long-term, such as detailed voting theory, should be deferred to a longer discussion. The system you disapprove of has been used in all the ArbCom elections so far. Carcharoth (talk) 17:41, 1 November 2009 (UTC)

As I responded on the page, our method is much farther from the Borda count than it is from approval. It is an example of

range voting; this is a 3-option range vote, while approval voting is a 2-option range vote. It is also strategically identical to approval voting. I don't see your reason for concern. rspεεr (talk

) 16:19, 9 November 2009 (UTC)

Two years and eighteen seats look rather safe bets right now; in the matter of the electoral method things are less clear (approval/preferential voting are not divided by much, and secret/public voting have been neck and neck). Given that we already have the technical capabilities to run with either of the potential election changes (secret/preferential voting from the Board vote), and the actual voting does not begin for another month, I don't see a major need for haste. I don't think it would be in our best interest to early close as no consensus, especially as if the proposed changes aren't defeated decisively, we will likely have to have the same discussion next year. Hopefully there will be greater distance between the options in a week's time.

barbarian 

23:43, 1 November 2009 (UTC)

Is it possible to use both voting methods, or is that too crazy to even contemplate? What would happen if the results were wildly different between the two methods? You could compare the voting lists from both methods. Those who want to use a secret ballot but not an open vote, would vote with their feet (if you know what I mean), and those who want to use an open ballot, but not the secret ballot could do likewise. You could even combine the two results to get a third result. Has there ever been any voting system anywhere crazy enough to use two voting methods simultaneously, or offer people the option of two voting methods to chose from? I suppose the only way this would work is if it is declared in advance which method will be used to decide the result (if you were combining them, you would eliminate from the open ballot those who voted using both methods, and then add the open ballot results to the secret ballot results; if you are not combining them, you pick one and the other is a "control"). Carcharoth (talk) 00:53, 2 November 2009 (UTC)

Jimbo actually suggested doing exactly this, here. This is one of those things that might just be crazy enough to work - worth examining, I think. UltraExactZZ ^Claims ~ Evidence 13:06, 2 November 2009 (UTC)

I don't know about combining open and secret ballots but on the topic of combining, cumulative voting could be used to combine approval voting with preferential voting - see my view on this on the RFC page with examples. Basically, you could use it to allow one group of voters to simply support who they want to support and allow the other one to support with different preferences. Regards SoWhy 14:12, 2 November 2009 (UTC)

I think implementing a hybrid system in such a large scale election with voters of very diverse levels of familiarity with the process without prior testing would be chaos. That said, although I don't have a position on public vs. secret voting and I'm sure it would not alleviate most of the concerns of the public proponents, I think implementing secret voting while encouraging those who wish to do so to add an RfA-like declaration of their vote and rationale on the candidate page would be very workable. Some variant of SoWhy's suggestion – such as allowing eight tokens to be distributed across the candidates at the voter chooses – would also work well for those with very different inclinations (sophisticated voting vs straight "these are the eight candidates I want as arbitrators"), but there are may be great drawbacks depending on the variation put forward.

barbarian 

20:32, 2 November 2009 (UTC)

I cannot help but notice that the AUSC elections, which are the "trial" for secret elections as proof-of-concept, has nearly 220 voters after barely three days of being open. There is, of course, no way to know whether there is a causal relationship with the fact that it is a secret ballot; but that extraordinarily high level of participation is certainly suggestive. — Coren ^(talk) 01:53, 2 November 2009 (UTC)

Does anyone have figures on number of voters for the past elections listed at

WP:ELECTION? I would note, Coren, that most of the votes in the ArbCom election arrive in the first few days, so you can't extrapolate the AUSC election numbers unless you assume that voting timings are different in a secret ballot (which they might well be). If anyone wants to go overboard on this, they could do a 'timings' analysis of past ArbCom elections and see how many votes were cast per x-hour period, and compare that to the current AUSC elections. Carcharoth (talk

) 02:07, 2 November 2009 (UTC)

Anecdotal sample: I would never have voted on that page without secret balloting. N=1 and all that. SandyGeorgia (Talk) 02:09, 2 November 2009 (UTC)

In order to properly compare it with turnout at previous elections, we would need to know how many of those votes include "neutral"/"abstain" votes which are comparable to neglecting to vote on a candidate in previous election methods. John Vandenberg ^(chat) 02:56, 2 November 2009 (UTC)

You can still get values for total number of voters, as distinct from the number of votes cast. Of course, for an open ballot, it is easier to count up the total number of votes cast. To get the number of voters, you need to do pesky things like cleanly extract voter names from the voting lists (eliminating those annoying comments and remembering to eliminate the struck votes) and then do some little magic trick (I use Excel) to get the number of distinct voters (you also end up with numbers for how many voted for one candidate, how many for two candidates, and so on). I tried it once for the December 2008 ArbCom elections. It's not that easy, but once you've done it for one set of ArbCom elections, you can do it for all of them. I do hope someone one day does a full statistical analysis of ArbCom elections. Carcharoth (talk) 03:16, 2 November 2009 (UTC)

What we can see quite clearly is that no discussion occurred, which makes more compelling the arguments for /debate subpages for the ArbCom election, deployed before actual voting begins. Cenarium (talk) 17:20, 6 November 2009 (UTC)

Originally landed on the RFC; moved to talk by John Vandenberg ^(chat) 12:35, 4 November 2009 (UTC)

Oh, what the heck. This doesn’t pertain to the voting at hand, but since there are so many influential people here on this issue who know there way around Wikipedia’s byzantine procedures and bureaucracy, I thought I’d mention it.

Far too much conflict and wikidrama is created by not having Admins’ autosignatures somehow indicate that they are an Admin. Different Admins have different styles of communicating. Sometimes, Admins will weigh in with sensible peace-making and it is instantly obvious to newcomers that whomever they are dealing with is fair and sensible, is someone of some import, and has influence with others. Still others can easily be mistaken for just another ill-tempered editor whose edit should be reverted. Whether it is an Admin reverting an edit, deleting text, or making some other kind of edit, many an ordinary contributing editor has needlessly gotten themselves crosswise with Admins simply because there is no automatic means by which they identify themselves as such. This has caused needless discord and wikidrama in the past and will continue to do so in the future.

I propose simply that Admins, Bureaucrats, and ArbCom committee members have an automatic and standard way of identifying themselves so that the rest of the Wikipedian community can quickly and easily identify with whom they are dealing. The best way, IMO, would be for there to be a standard parenthetical suffix appended to all editors who have special privileges and powers. That might be something like [Admin]. At the very least, there should be a standardized header tag at the top of such editors’ user pages. We shouldn’t have to hunt through user pages to find some little userbox entry stating that the editor they are tangling with is “not an Admin and does not wish to be” in order to get a clue who we are dealing with. And it is certainly asking too much of our new contributors to expect that they know how to navigate Wikipedia to figure out the privileges and powers of someone they are dealing with on a contentious issue. Greg L (talk) [I’m nothing; I’m navel lint.] 23:54, 1 November 2009 (UTC)

There are various tools that will allow you to quickly and easily identify such status (well, the ones associated with an actual switch of a bit in the database, that is - for arbitrators, you will just have to memorise the list). I'm sure others can point you to these tools, preferably on your talk page, as this is indeed very off-topic. Carcharoth (talk) 00:43, 2 November 2009 (UTC)

*I* know there are “tools” to identify Admins, etc. You know there are “tools”. I’m talking about something that is automatic and convenient for less experienced contributors; you know, that sea of editors out there who know how to edit pages but aren’t as familiar with the ins and outs of Wikipedia as are people who spend half their waking day here. It’s the same principle as having police officers wear a uniform and badge: sure you can probably go to the Web and look up whether they are employed by the city, but it saves needless confrontation when one instantly understands that the individual they are dealing with is one of those “regular editors who has ‘access to special tools’” (as some admins are fond of saying). More importantly, having Admins automatically declared in their autosignature as being Admins will save needless wikidrama and conflict due to editors not understanding precisely with whom they are dealing.

If Admins are just “regular editors who have access to some special tools,” I can hardly think of a significant downside to having their status declared in their autosignature. But I can certainly think of the needless conflict that is spared. There are some Admins who don’t strictly obey the rules and have an aggressive style. These Admins in particular might even behave better if it is advertised every time that they are an Admin and a high standard of conduct is expected of them every time they sign their work.

I’m also not so sure it is entirely off-topic here to be discussing this issue. The principle of declaring one’s powers on Wikipedia applies to Bureaucrats and Arbs too. Helpfully, your talk page discloses that you are an Arb. That’s a good practice; thank you very much for setting a leadership example. The very first other Arb’s page I jut checked prominently discloses at top only that he is an Admin. Not so good. Disclosing this information in one’s autosignature has no downside that I can think of that could offset the good it will do. Greg L (talk) 17:10, 2 November 2009 (UTC)

It might not be entirely relevant to the discussion here, but I agree with this proposal. Admins an arbitrators alike should definitely have to prominently identify themselves as such on their userpages (some of the current ones don't do this; or, only do so via an easily-missed category), and I like the idea of highlighting it in their signature as well. A simple 'Admin' tag like the one above would be fine. If necessary, it could be turned off when a user wishes to participate as an 'ordinary user' rather than an admin/arbitrator (e.g. to indicate that they're not acting as a representative of ArbCom), but that might be getting ahead of ourselves... the simple signature tag is a good idea, either way. Robofish (talk) 03:16, 3 November 2009 (UTC)

Over on Wikisource, admin user:pathoschild included "admin" in his signature.[5] I am a bit hazy about this, but iirc, the community objected because it gave the appearance that his voice was somehow more important in community discussions. If you are interested in seeing that discussion, ping his talk page here, as he may be able to find it quickly. John Vandenberg ^(chat) 12:43, 4 November 2009 (UTC)

I think I would generally feel the same way; in ordinary discussions being an admin doesn't carry any special authority. Repeatedly proclaiming one's status in signature form appears to make the claim that it is relevant, presumably granting greater authority or significance to one's views. Users shouldn't be treated differently in discussions on account of their status and I don't understand Greg L's claim that a tag in their signature will help hold administrators to a higher standard of behavior. Rather, I suspect, it will help them leverage their status to exert greater authority. Christopher Parham (talk) 18:14, 4 November 2009 (UTC)

Greg L, the new forum extension that Werdna has been developing, called mw:Extension:LiquidThreads, already has an autosig, so it would be quite simple to also display "admin" in the autosig. You can play with this extension over at strategy:Village pump.

A step in the right direction would be to make it mandatory for admins to have a userpage, and that it must clearly state that they are an admin. John Vandenberg ^(chat) 01:49, 7 November 2009 (UTC)

Maybe some contributors here would be interested to look at the recent Ireland-Republic of Ireland vote. That was a formal, public vote using a complex method (instant-runoff voting).

Some "innovations" we used included:

• A standard template for voting ({{stv-ballot|A=0|B=0|C=0|D=0|E=0|F=0|sign=~~~~}})
  • Made it easier to vote using a complex voting system
  • Facilitated
    bot counts and calculations
    of the result
• Banned all comments in the "balloting area"
  • Comments on procedure etc. were allowed allowd at the bottom of the page
  • Discussion etc. were allowed on the ballot page's talk
• A separate "intro" page was full-edit-protected and transculed into the balloting page
  • This an agreed neutral statement on the vote/procedure/options being altered during the vote
  • Notified users of voting rights, the period of the poll, punishments for breaches and the procedures for deciding the result (which were agreed before hand)

Not everything went smooth. There were (alleged and reported) incidents of voter intimidation. These were dealt with by a supervising admin. On-going tallies created tensions (of which I am not innocent). There were votes by socks. These were identified by users (something a closed poll would not have facilitated) and removed from the balloting area by a supervising admin.

Overall the process went well. Keeping comments out of the balloting area was a good idea. As were clearly stated "rules". The standardised template was a godsend in calculating the result (although that was not the initial reason for developing it). Running a public vote had it's own problem (this was a very high-tension issue) but ultimately (I think) was necessary to legitimise the vote.

Maybe it's an example that can be developed upon - particularly if a complex and formalised method of voting is to be carried out by a public vote. --rannṗáirtí anaiṫnid (

(Adds: after seeing this vote being carried out the way it was, despite all the

Candidates may begin submitting their nominating statements in 10 hours (less, now). Looking at the RFC, it seems that consensus strongly favors an 18-member Committee with 2 year terms. I'm not suggesting we close those sections of the RFC, but would it be acceptable to note the fact at ACE2009? I think it would help to inform some decisions from potential candidates, which is worthwhile. Or maybe the consensus here is that consensus exists on those issues, and they may be closed in favor of focusing debate on the remaining issues. (Of course, if a potential candidate hasn't been following this discussion, they're probably doing it wrong.) Obviously, more discussion is needed on the Secret Ballot/Open Ballot issue, and on the Support/Oppose and Preferential options. Whatever consensus exists about the consensus, if any, is fine - Just asking the question. Thanks. UltraExactZZ ^Claims ~ Evidence 14:27, 9 November 2009 (UTC)

We can or at least should assume that anyone considering to stand for ArbCom membership has been following the related pages for the last month(s) and thus is aware of the RFC, so I see no reasons to explicitly post at ACE2009 how the current consensus looks like. Since the questions do not really have an impact on being a candidate in itself (it only influences the chance of being elected and the term of office if elected), I think we should leave the completely RFC open until shortly before voting starts (maybe 28 November 2009) to be able to allow as many people as possible to express their opinions before consensus has to be assessed to allow the elections to take place. Regards SoWhy 14:43, 9 November 2009 (UTC)

It's ultimately for Wales to decide the number of appointments and term lengths, and even though he is likely to honour the outcome of the RfC and potential candidates should take heed of the state of play, I agree with SoWhy that we don't need an urgent final community decision. With three weeks or so until the voting begins, I do think it would be best to wrap the RfC sooner than SoWhy suggests, to give us time to make arrangements for changes (if any) in the election method and to have that side of things settled and everyone on the same page so that we can focus on the nominations/question phase. Ultraexactzz, given that you haven't participated in the RfC, how would you feel about setting a date for closing it?

barbarian 

15:06, 9 November 2009 (UTC)

That's fine. Procedurally, I would want at least a few days between the close of the RFC and voting to make sure the voting pages are set up, either as redirects to Securepoll or as traditional open support/oppose pages - so the 28th would work, except that the last week of November is the Thanksgiving holiday week in the US; what with travel and family dinners and (oblong) football and so forth, I'd rather have things set and done before the 25th. So, unless there are objections, I propose closing the RFC at 00:00 on 24 November. It gives us a full week before voting, and two full weeks from tomorrow to wind up discussion. UltraExactZZ ^Claims ~ Evidence 16:02, 9 November 2009 (UTC)

I think that sounds good Ultra - I'd also suggest that all open aspects of the voting method be chosen by simple vote this time, closing on the date / time you mention - which should give enough time for secure poll etc. to be set up if necessary. I've already poked MBisanz about trying to find someone appropriate to wrap this up, and I'm happy that you seem to be up for the task :-) cheers, Privatemusings (talk) 01:44, 10 November 2009 (UTC)

A few people are agreeing with Mr.Z-man's caveat that there should be some "minimum support" in case there are not many candidates. Can this be determined before the election begins? How many (non-joke) candidates would we need running before we are happy to let the election be deterministic (barring some unforeseen controversy)? Alternatively, should we add a poll option to determine the ideal threshold for "minimum support"? John Vandenberg ^(chat) 09:02, 30 October 2009 (UTC)

I'd say a minimum threshold (for example 75% support) is much better than a minimum number of candidates. I think we should definitely add it to the RFC asap to be decided before the election. Something like "Have a minimum threshold of 75% support and select less than 18 members if not enough candidates meet this threshold?" Regards SoWhy 13:22, 30 October 2009 (UTC)

Agreed. Though, being practical, with two proposed minima: 75% and 66%. With 75%, we'd only have seen three arbs elected last time.  Roger Davies ^talk 13:32, 30 October 2009 (UTC)

I think this might be ambitious. There may come a time when voters are even less impressed by the candidates than the sitting arbitrators, and may choose to direct their frustration accordingly. The last election was marked by an overwhelming desire to get the old crowd out and a new batch in; a reverse of that dynamic could leave candidates with two-thirds or more support looking rather lonely. As long as we're using a crude support/oppose ratio as the metric, and arbitrator attrition rates make constant influx a necessity, the natural place to draw the line would seem to me to be positive net support.

barbarian 

13:42, 30 October 2009 (UTC)

I agree that positive net support should be sufficient. I mainly just want to avoid scraping the bottom of the barrel in order to fill all the seats in case we don't have enough good candidates. Arbs tend to lose support quickly in their role; someone starting with 40% support may not last too long. Mr.Z-man 16:34, 30 October 2009 (UTC)

75% was only an example. I think the proposal should be 55% based on last time's numbers and even then we would only have 12 people to choose from. But then again, if people are aware that more candidates can be elected then they might be willing to support more (assuming that many opposed to have only those they thought best for the "job" elected. I think the number should be low enough to have a tolerance built in, i.e. to ensure that (based on last year's numbers) we have more candidates over the bar than seats to fill. Regards SoWhy 13:43, 30 October 2009 (UTC)

If we get secret ballots, I predict much lower numbers over all - if anything we may have seen an increase in tactical voting, and I'm guessing people felt free-er to oppose with secret ballots. In either case, I would suggest the existing consensus is that we already have is 50%+1vote for acceptance.--Tznkai (talk) 04:03, 11 November 2009 (UTC)

Moved from front page — Coren ^(talk) 00:26, 3 November 2009 (UTC)

One of the main reasons for opposing secret voting seems to be a lack of transparency. But what transparency are we actually losing?

• We know who the candidates are
• We know who voted
• We know who the election admins are

The only thing we no longer know is who voted for who. Why do we need to know this? If someone is eligible to vote, how is it our business to know how they voted? If they don't tell us why they voted, then its mostly useless information for any beneficial purposes. If knowing how they voted would create any more or any less of a reason to look at them with suspicion, then that's just another positive aspect of secret voting. Mr.Z-man 23:58, 1 November 2009 (UTC)

You also lose transparency into who people have voted for broadly (i.e. running count), so a burden is placed on later voters to analyze candidates who may already have overwhelming support or opposition. This is likely to result in a poorer allocation of voter effort, with voters wasting time thinking about irrelevant candidates that could be spent on marginal candidates. The presumed impact would be that the marginal candidates selected would be less likely to be the best of that group. This is one example of the way in which moving from a "consensus"-esque to purely democratic process is harmful: later voters do not build on the base of work done by earlier voters but instead have to start all over again. Christopher Parham (talk) 15:25, 2 November 2009 (UTC)

(Hoping Mr.Z does not mind the discussion here— this might be best moved to the talk page?) Then again, there is a strong argument that the phenomenon you describe (later voters voting according to how other people have already voted) is actually very much undesirable; it means that small differences early are amplified regardless of merit. People voting according to a trend rather than for who they believe will be able to do the best job are very damaging to the selection process. In fact, this is an easy strategy to sink any ArbCom candidate you want with open voting: make sure you have a group of about 10-12 friends ready to oppose the candidate within the first two or three hours of voting. Preferably, have one or two support their votes with a short but damaging comment that is very hard to disprove. I guarantee that this candidate will fail. — Coren ^(talk) 17:12, 2 November 2009 (UTC)

I agree that voting based on inertia is more of a problem than benefit of an open vote. Personally, I find it more of a problem than voter intimidation. What's the point of having a 2 week vote if the votes from the first day (or even the first few hours) determine the outcome? Why should candidates bother spending so much time making statements and answering questions if only the first 5% or so voters are actually going to bother to review them? A better way to minimize wasted voter time is to have some sort of minimum requirements for candidates to filter out candidates who really don't stand a chance. Mr.Z-man 21:40, 2 November 2009 (UTC)

I think this is by far the strongest argument in favour of secret voting, and am surprised it hasn't received more attention. The current system of public voting essentially disenfranchises any user who arrives to the election 'late' - say, more than 24 hours after it's started - as by then, the winners and loses have to a large extent already been decided, and their vote won't make any difference. I felt this in the last ArbCom election - I only came to it after about 3 days, and instantly realised my votes were virtually useless at that point. By contrast, those who vote in the first hour or two have immense power in an open vote, as they can start the basic trends of support that determine who stands a chance of being elected. A secret ballot would avoid all this; if the results were not revealed until voting was over, it would still be worth voting up to the very end. Robofish (talk) 03:44, 3 November 2009 (UTC)

The fact that your ability to impact a vote is limited does not mean you are being disenfranchised; you can't expect to be the pivotal decisionmaker on every issue. Why would it be "worth" voting in the secret ballot when the candidate is already losing 10-150; how does your ignorance of this fact change this action from useless to worthwhile? Frankly, I think people in this conversation are confusing causation and correlation. There's no evidence that the votes of early particpants determine the votes of latecomers or establish an "early trend of support" beyond the obvious fact that a small sample is quite representative of the whole. If you believe that voters are generally acting in this way (that is, voting in a manner unrelated to the qualification of the candidate) that's a garbage in-garbage out problem that no type of ballot can fix and suggests we shouldn't bother with an election in the first place. Christopher Parham (talk) 14:26, 3 November 2009 (UTC)

The issue isn't that people's votes won't affect the outcome, its that they are voting based on the current outcome. People who may have voted one way will vote the other way because they don't want to be seen going against a large majority. If enough people do that, it will have an effect on the outcome. One vote will almost never make a difference, but several dozen like that will. If people don't know the current results, then obviously they can't do that. They might vote randomly, but even that would be slightly better than just voting to reinforce the majority. Mr.Z-man 01:22, 4 November 2009 (UTC)

"What transparency are we actually losing?"

• We know who the candidates are - yes
• We know who voted - no, if we did it wouldn't be a secret ballot would it?
• We know who the election admins are - well, we know who got the job but we don't know if they are they ones we voted for

These are legitimate concerns and do not represent a "mistrust" in anyone. If you're going to do something then you might as well do it right. In real world elections there are systems in place that provide confidence that the result is genuine. That is important as much for the sense of legitimacy that it gives to the vote as much as much as to actually preventing fraud in a real sense. We should aim for the same if we are going to go down the line of a secret ballot. --rannṗáirtí anaiṫnid (

coṁrá

) 22:06, 2 November 2009 (UTC)

Real world elections appoint leaders with the power to make war, impose laws, command armies and levy taxes. A sense of perspective would not go amiss. Happy‑melon 23:01, 2 November 2009 (UTC)

Very few real-world elections elect leaders with the power to make war, impose laws, command armies and levy taxes. Most elections appoint committee members to write updates on a community clean-up project for a local newspaper, to act as treasurer for an underage football club, to mind the keys to a photo-club's dark room and to organise posters for a fundraising dance. Those elections follow best-practice. They have a sense of perspective. We should do the same. --rannṗáirtí anaiṫnid (

coṁrá

) 23:27, 2 November 2009 (UTC)

Have you actually examined the system? We do know who voted. We don't vote for the election administrators (just as we don't vote for poll workers in a RL election), so I'm not sure what you mean by your other objection. Mr.Z-man 00:15, 3 November 2009 (UTC)

I'm not aware of my local football club's triple-blind election system. My local council elections are less transparent than the WMF Board votes, and local councillors do have the power to levy taxes. I would rather use SecurePoll than postal votes in any national election. The system you've proposed is more suited to a world-leader election than for appointing ArbCom: the distinguishing feature is not the power of the office, but the lack of a trusted party to hold the keys. Those elections' "sense of perspective" is the correct evaluation that there is no available party who is totally ambivalent as to the result, who can be the cornerstone to build the election security around. You are discounting a perfectly valid cornerstone for on-wiki elections to chase after an ideal that is technically impossible to achieve. There is no way to "to it right". No one watches these watchers, that's a fact we all just have to live with. Except that's not true, because the WMF watches the sysadmins, and as said, if they're found to have committed election fraud, they don't get chastised, or desysopped, or banned; they lose their jobs. That is the sense of perspective I'm looking at. Happy‑melon 00:17, 3 November 2009 (UTC)

• HM - Your argument is full of
  straw men
  (e.g. my examples were the very opposite of government elections, local or national, and you say I am "discounting a perfectly valid cornerstone for on-wiki elections" when I mentioned never said any such thing) and is circular and contradiction ("there is no available party who is totally ambivalent as to the result ... There is no way to "to it right". No one watches these watchers ... Except that's not true, ... the WMF watches the sysadmins..."). I'm not going to reply to your comments here or on the talk page of my proposal because they don't make sense.
• Mr Z-Man - Did these people vote? How do I know? In real world elections we do not vote for poll workers. Instead interested parties can assign overseers and tally men to supervise all aspects of the electoral procedure (I've worked as a tally man on local and European elections in Ireland). In the real world there are practical limits on the number of tally men/overseers owing to the resrictions on physical space and the need to maintain order. In a virtual election there is no such need. That's the essence of my proposal - whether the curret details of the proposal are entirely satisfactory or not - to facilitate the greatest degree of supervision over the casting and counting of ballots by editors. Ultimately the WMF are responsible for running the election, but editors should be able to satisify themselves that the election was run in a satisfactory manner and not simply blindly "trust" that it was.

The proposal I am making does not guarantee 100% that fraud cannot ever take place. Once you introduce a sectet ballot, nothing can guarantee that. Instead, it is intended to give a greater degree of confidience among all editors that elections are being run correctly by allowing editors to supervise the balloting and counting process without interfering with the election. --rannṗáirtí anaiṫnid (

coṁrá

) 08:57, 3 November 2009 (UTC)

Well I don't really know how to get across thoughts like "you cannot protect a system from sysadmins" and "sysadmins will lose their jobs if they commit election fraud" any better than I have, so we'll just have to leave it at that I guess. As far as I can see, you are either trying to build confidence in an area where none is lacking, or chasing after an ideal that is both unattainable and unnecessary. Naturally if I now see a barrage of comments to the effect of "I don't have confidence in the sysadmins to host a fair election", then I'd have to reconsider that position. But I'm not expecting to. Happy‑melon 10:34, 3 November 2009 (UTC)

The way you wrote it just there is about as clear it could be said. Thank you. I think you've missed the point though (otherwise, if you mean it as a serious argument it's another example of straw men). Taking each one in turn:

• "you cannot protect a system from sysadmins" - Correct. We cannot stop sysadmins from committing fraud. There is no point in trying. Give up now. It will never happen. If a sysadmin wants to manipulate an election (or is instructed to do so) there is nothing we can do to stop them. What is possible in its stead is to devise a system whereby if they do manipulate a result, they risk being caught. Understand? The precept is not "I can't stop you”, it's “if you do it, I can catch you."
• "sysadmins will lose their jobs if they commit election fraud" - The motive for the risk of fraud doesn't matter. Or the consequence for the fraudster. Whether a sysadmin was the instigators of the fraud, whether the board instructed them to manipulate the vote, or whether else - it doesn't matter. Fraud happens. The advantage of an open vote is that the electoral process is so transparent that fraud is difficult to commit. When you move to a closed voting system, that safe-guard goes away and you need to think of other ones to take it's place. In many cases (e.g. local clubs, class rooms, etc.) simply saying, "trust me" is sufficient. Here, online, we don't know each other, there are no substantial societal bonds between us. Simply saying "trust me" is insufficient. Instead, there are means whereby I can have confidence in the result of a closed election without relying on "trust" alone.

It's not a slight against anyone's character that I would want something more substantial to base my trust upon. It doesn't mean that I am "paranoid" or lacking in "trust". It just means that I am not naïve enough to think that I can always trust in strangers. --rannṗáirtí anaiṫnid (

coṁrá

) 13:22, 3 November 2009 (UTC)

I'm certainly missing your point, because I don't see it as at all valid. We agree that sysadmins can do anything with, or to, an election server, and yet you continue to assume that you can construct a system to run on that server that users can confidently believe will be impossible for those sysadmins to game without being caught? Don't forget, of course, that in this quest you have sacrificed the holy grail of secret voting - the guarranteed anonymity of votes: having linked voters to votes with a voter id, the secrecy of the vote now rests entirely in the security of the hash function, which is mathematically impossible to secure. And don't forget that, since the sysadmins must not have the ability to put votes to voters (the impossibility of that notwithstanding), neither can the server, so there can be no striking of votes based on user characteristics (editcount, registration time, etc). Am I the only person here who concludes that these wheels within wheels Just Aren't Worth It? Given years of fraud-free elections, secret and not, for positions high and low, on-wiki and on external servers, including some where the WMF board has far more to gain from a particular outcome than the appointment of editor X instead of editor Y to two user rights, a private wiki and a mailing list, given that one of the board members has a veto over such appointments anyway, am I the only one to conclude that anyone within WMF who is in a position to commit electoral fraud on a SecurePoll election Just Doesn't Give A Shit?? That, to my mind, is the point. Happy‑melon 14:18, 3 November 2009 (UTC)

• "... you continue to assume that you can construct a system to run on that server that users can confidently believe will be impossible for those sysadmins to game without being caught..." No. That would be impossible. What I am proposing is a procedure of system assurance. Under that procedure, the chance of an inaccuracy in the system being being identified would rise at a rate geometric to the number of inaccuracies. (Note: "chance", it would make it more difficult not to get caught, not nearly impossible, only more difficult.)
• "...the secrecy of the vote now rests entirely in the security of the hash function..." You would have a better chance of cracking the secrecy of the ballot if you simply emailed Tim Starling and asked him for an SQL dump.
• "...since the sysadmins must not have the ability to put votes to voters ... there can be no striking of votes based on user characteristics" - Eh? Now you're just making stuff up. I never said that. Of course the sysadmin would know who voted and how. As you say, that would be impossible (and undesirable) to prevent.

No one gives a shit? Jesus, you're getting very worked up for a man who doesn't give a shit. If it doesn't make a difference then it doesn't hurt does it? --rannṗáirtí anaiṫnid (

coṁrá

) 19:48, 3 November 2009 (UTC)

Your points are flatly contradictory: you say that the secrecy of an existing secret ballot is most easily broken by getting the data from the sysadmins, which is true, but then you say that you would want the sysadmins to have the same records in the new system. Are they trusted or not? Clearly not since otherwise we wouldn't be here in the first place, but now you're building a system that has myriad checks against a sysadmin stuffing ballots, but absolutely no protection against voter blackmail, coercion, bribery, or any of the other things that secret ballots are supposed to guard against. If a sysadmin is prepared to ballot stuff, he'd certainly be prepared to buy votes. Which is the chicken and which is the egg?

In social terms, no I don't give a shit, I'd be perfectly happy to do my best to ignore all these obfuscated bells and whistles when casting my secret ballot. Why am I sufficiently motivated to keep at this conversation? Because in the event that you manage to persuade the community to run with this ridiculously overcomplicated system, I'm the poor sod who'll have to code it for you. You'll forgive me, therefore, if my not-giving-a-shit manifests itself slightly differently to most peoples'. Happy‑melon 22:13, 3 November 2009 (UTC)

Eh, my point was that an editor would be more likely to get an SQL dump of Tim Starling (quite unlikely) than to crack an SHA-1 hash (nearly impossible). Get it?

• "Clearly not since otherwise we wouldn't be here in the first place..." I don't know what you mean by this.
• "Are they trusted or not?" Trusting an individual is different from having confidence in a system. A system, particularly one based on anonymous online communication, should be able to stand on it's own merits and be independent of the individuals involved. It's not a matter of "trusting" the system admins, it's a matter of being "confident" in a system.
• "...absolutely no protection against voter blackmail, coercion, bribery, or any of the other things that secret ballots are supposed to guard against..." - Well. you said yourself, those are the "things that secret ballots are supposed to guard against".
• "If a sysadmin is prepared to ballot stuff, he'd certainly be prepared to buy votes." And one could do so now too if they liked. They would likely get caught since those sort of things are quite public (particularly coercion of a scale great enough to swing the outcome of an election where the difference is greater than a handful of votes).
• "Which is the chicken and which is the egg?" This doesn't make sense. You haven't given any chicken-or-egg scenarios (e.g. A begets B begets A)
• "I'm the poor sod who'll have to code it for you." LOL - well as another poor sod who has to code things you could have said that from the start! Your opposition makes much more sense now. I had good intentions when I first put the page together to code it myself, but I haven't even put requirements together for it or really dug around inside SecurePoll. On the surface it doesn't look too much - for heaven's sake what the hardest part? Most of it is there already in SecurePoll from what I can see (although I haven't even looked at what the Db behind SecurePoll looks like). --rannṗáirtí anaiṫnid (
  coṁrá
  ) 22:56, 3 November 2009 (UTC)

(undent) I think the point Rannpháirtí anaithnid is missing isn't on the technological side. The methods he proposes (most of which, incidentally, are already supported by SecurePoll) are sound, and would indeed defeat tampering even if the sysadmins colluded with the foundations and the scrutinizers into defrauding the electorate.

The point he misses is that the added overhead he proposes is entirely disproportionate to the actual calculated risk. The stakes are nowhere high enough and the probability of successful collusion is low enough that the actual risk (which is a product of the two) is completely neglectable.

• Demonstrably, none of the parties in a position to influence the counting have any incentive to affect it;
• even if they did, they could do little more than prop up or pull down one or two candidates without everyone noticing; and
• even if they did that it would be a minor influence on the committee as a whole, and thus not worth the risk of exposure and severe egg on face.

The point is, the proposed level of security is completely out of whack with the actual risk. Those elections aren't for people with a finger on a big red button. For that matter, they are not even for people with money to spend or people to hire. Confidentiality of ballots is desirable because it lets people vote freely without social pressures, not because there is a great security risk that needs to be mitigated at great cost. — Coren ^(talk) 14:59, 3 November 2009 (UTC)

Thanks, Coren, for the criticism. I'll reply first by asking, what are the additional overheads (aside from the initial development time on the SecurePoll extension)? Once that's done it's an automated and optional process with no more effort required of the voter or election organiser that already exists in current SecurePoll elections.

Second, the "it would never happen" arguments (whether that be because "it's not worth it" or because "no one would be motivated"), I say, misses the point. Whether it would or not happen doesn't matter. If we move from a open ballot to a closed ballot we place trust in strangers to conduct business on our behalf properly without our supervision. Whatever the stakes, it's silly to blindly trust anyone you don't know -

• Have you ever received a utility bill where you were incorrectly charged? Did you phone up the company and point out their mistake? It wasn't part of a conspiracy. They weren't out to defraud you. They had no motivation to do so. It was just a mistake - but it would have been foolish of you to trust them blindly.
• Have you ever had a
  take-away
  delivered to your home? Before the deliver guy left did you check that everything your ordered was there? The deliver guy wasn't going to cheat you. He has no motivation to do so. The restaurant owner isn't out to get you either. The costs in involved are minial - a bag or chips or a can of coke, nothing big - but before the deliver guy leaves, it's good to know that everything is there, isn't it? It would be foolish to trust them blindly.

That's the kind of confidence you lose when you move away from an open ballot. You place trust in strangers. And even if those strange have no motivation or reason to defraud you, it is unwise for you to trust them blindly, wouldn't it? --rannṗáirtí anaiṫnid (

coṁrá

) 19:00, 3 November 2009 (UTC)

It's all about proportionality. Yes, you'd want to check your phone bill; but what you are proposing is closer to making sure you time each of your calls with stopwatches made by two manufacturers just in case one of them had colluded with the phone company to defraud you. The stopwatch may only cost you a few dollars, but it's out of proportion with the actual risk to buy and use it. Likewise, your delivery analogy breaks down in that what you are checking for is error, something that the current SecurePoll system already covers very well; your proposal is closer to what has been said below tongue in cheek: you want to send samples of every delivered item to spectroscopy before you let the delivery guy leave. — Coren ^(talk) 12:14, 4 November 2009 (UTC)

So somehow we've moved from a system that's designed specifically around combatting determined and systematic electoral fraud on just about every level, to thoughts about "confidence" where we assume that no one is actually malign? Where are you pitching this? Happy‑melon 22:13, 3 November 2009 (UTC)

Happy, it's clear that you're not ("happy" that is). Just as the delivery guy doesn't take offense that a customer would look inside the deliver bag, or the phone company doesn't get worked-up that someone would request an itemised bill, don't take offense that someone would not blindly accept the word of SecurePoll without question. You are an actor in a

coṁrá

) 23:09, 3 November 2009 (UTC)

Now there's a statement that makes no sense. Who is this supposed to be "personal" against (or at least, who am I supposed to be claiming it's personal against)?? Of course it's not personal; a personal vendetta would be "I don't trust sysadmin X, I'm designing a system that makes it impossible for him personally to affect the vote". If anything, your system is totally impartial: it doesn't trust anyone. To extend your analogy, your system wouldn't just open the takeaway bag, it'd open up each individual food package and send it off for spectroscopy just to make sure that it hadn't been poisoned. Despite the fact that not only is the chef trusted to prepare pufferfish, if he's found to have stuck arsenic in the curry he gets fired at the very least. Happy‑melon 11:51, 4 November 2009 (UTC)

Better to embargo the submissions until the end, as some people will just pile on with opposes even if a disgruntled opposer writes nonsense, eg, that the candidate harbours socks or whatever. Afterwards, you can still see any traces of bloc voting or whatever. YellowMonkey (bananabucket) (help the Invincibles Featured topic drive) 01:45, 4 November 2009 (UTC)

This stuff about Diebold-like vote counting shenanigans is misplaced. Someone wanting to manipulate the vote count can most easily do it by enrolling sock accounts and voting with them. Public voting doesn't cure that, but it makes it a little bit less trivial, since the edit histories of dubious voters would be inspectable. With secret voting, I could enroll 100's of accounts over a several month period, make a few legitimate edits with each, and have a slow-moving bot make enough trivial, innocuous edits to fill out the required 150 for each account without anybody noticing. Public voting would catch out that pattern pretty fast. Secret voting makes no sense at all when the voters are completely virtual with no real-world authentication. Anonymity of low-activity editors, not the parts of the process run by the most visible people in the community, is where the real transparency problems are. We don't know who voted or even how many people voted. We only know how many accounts voted, and which ones they were.

Secret voting might be ok if it was done in person at Wikimania, and maybe some regional meetups held at the same time in other places, like real-world elections. That would ensure one vote per person, and also would ensure a higher level of commitment from each voter, since they'd have to actually show up at one of those places in order to vote (the 150 edit threshold is ridiculously low; any active editor has probably had binges where they made several times that many in one day). But it would still be

Could I remind editors that these RfCs were held to establish community opinion for the current election. Otherwise, they'd probably have been postponed until next year. It seems as clear as daylight that 18 seats and two-year terms win hands down; it is particularly telling that the preferences for these outcomes were each one of four choices, which is likely to spread voting more evenly. Community opinion, as of today, is:

Number of seats

• 12 seats: 5 supports (4.9%)
• 15 seats: 27 supports (26.5%)
• 18 seats: 63 supports (61.8%)
• "Flexible" (let Jimbo decide): 7 supports (6.9%)

Term lengths

• 6 months: 2 supports (1.8%)
• 12 months: 13 supports (12.7%)
• 18 months: 9 supports (8.8%)
• Two years: 78 supports (69.6%)
• Three years: 10 supports (8.9%)
• "Flexible" (let Jimbo decide): 1 support (0.9%)

I believe the community should adopt these outcomes forthwith. I am posting a note to this effect at the election talk page, and a query as to whether an uninvolved crat or admin should be asked to officially close these RfCs, which have been running for more than two weeks. Tony (talk) 02:45, 11 November 2009 (UTC)

I concur, for what its worth.--Tznkai (talk) 02:47, 11 November 2009 (UTC)

me too - I think these matters could usefully be considered closed at this point. Further, the issue of whether or not to use Secure Poll could be settled by the same person, perhaps to the schedule suggested above by User:Ultraexactzz ('I propose closing the RFC at 00:00 on 24 November') Privatemusings (talk) 02:56, 11 November 2009 (UTC)

The result of the election secrecy poll appears rather clear also. A clear consensus does not exist, and is unlikely to develop in the short time remaining. WIthout a clear consensus, the current method will not be changed, except by decree of Mr. Wales. —

*

07:00, 11 November 2009 (UTC)

Although my own opinion is that 18 arbitrators turned out to be too many, it does appear that there is a pretty strong weight of community opinion here (and I suppose 18 also builds in for some attrition). As for the term length, if we were to move to two years, can I assume that the arbitrators elected in previous elections are grandfathered (I'm finishing my second-year of a three year term)—or do I need to decide whether to start printing reelection posters? :) Newyorkbrad (talk) 02:58, 11 November 2009 (UTC)

NYB, you've pointed out an assumption that probably should have been clarified in the RfC text for the terms. I've always understood the RfC to involve new terms only. It would seem manifestly disruptive, not to mention unfair to existing arbs and those who voted for them in previous elections, for their terms to be changed mid-stream. Tony (talk) 03:02, 11 November 2009 (UTC)

It's my view that it would be reasonable to say that objections to existing terms should have been explicitly raised before this point. Bad luck brad. Privatemusings (talk) 03:04, 11 November 2009 (UTC)

I'd say grandfathered. The onus was on those voting to specify if they wanted to alter already extent terms.--Tznkai (talk) 03:08, 11 November 2009 (UTC)

(reindent) - Although the RFC proposal wasn't explicit, the preamble cast this entire RFC in terms of the upcoming election. As a result I certainly thought I was voting on the term for newly elected Arbs, and not sitting members. as I see it, NYB was elected to a three year term (not directly, but by virtue of being the higest vote getter in 2007, and it being known that a three year term was available). Also if the new arrangement WAS applied retroactively, then the arbs currently on a one year term would need to have their terms increased. Manning (talk) 03:15, 11 November 2009 (UTC)

Agree with Tznkai. The people elected in 2007 and 2008 for three year terms ought to be able to finish those terms; it would be unfair to force them to run for reelection when they, and I assume a large majority of the voters, would not expecting them to. It ought to have been explicitly stated in the RfC if this sort of thing were going to be changed retroactively. NW (Talk) 03:25, 11 November 2009 (UTC)

I took existing terms into account in my "tranches are obsolete" view, which seems to have been largely overlooked. Anyone want to comment on it? rspεεr (talk) 03:56, 11 November 2009 (UTC)

Any attempt to create an ex post facto diminshment of current Committee terms would be inappropriate. All current arbitrators will not have their terms adjusted. —

*

07:04, 11 November 2009 (UTC)

This will make tranches semi-obsolete. We'll end up shifting into a about 9 two year seats up every year, with adjustments for resignitations. --Tznkai (talk) 03:58, 11 November 2009 (UTC)

How about we apply two year terms beginning with those elected this year, and the currently sitting arbitrators (unless their terms expire this year) will all come up for election next year (even if the term was originally three years)? As Tznkai suggests above, this would enable us to still have one regular election per year and have some renewal. If we only had one election every two years, ArbCom lethargy would get so bad that they would get virtually nothing done. Everyking (talk) 04:14, 11 November 2009 (UTC)

We can burn that bridge when we get to it, but I expect that a good number of arbs will be persuaded to stand even if they are not forced. I will take embarrassing pictures at the next Cabal Christmas party to ensure it, in fact. But yes, 18 every two years would be a terrible idea.--Tznkai (talk) 04:20, 11 November 2009 (UTC)

Near as I can figure, based on {{

ArbitrationCommitteeChartRecent}}, Of the 10 arbitrators whose terms do not expire at the end of 2009, 6 will have 1 year remaining on their terms, and 4 will have two. So, to create two equal tranches of 9 arbs each, we should elect 5 arbitrators to 2 year terms and 3 to 1 year terms. Tranches would then alternate elections. Who gets what term might come down to preference (if a candidate asks for a shorter term, for example, or if a sitting arb asks to shift tranches), or Jimbo's discretion. No sitting arb should have their term shortened, except by their request. UltraExactZZ ^Claims ~ Evidence

06:20, 11 November 2009 (UTC)

Oh no, no, no, please. We've just had this wonderful poll that says 2-year terms are optimal. We also know that a percentage of arbitrators will resign every year. Please let everyone be appointed for 2 full years, period. One or two of those two-year appointees will resign before the next election regardless. Even with grandfathering in, we will probably be filling the majority of seats on the committee every year. One-year terms have been soundly rejected, and certainly should not be almost half of the appointments. Risker (talk) 07:09, 11 November 2009 (UTC)

Currently there are four sitting arbs with two years on the clock, and six arbs with one year on the clock. If we appoint all eight new arbs to two-year terms, we will have 12 arbs on two years when we really want 9. I'd say that losing three arbs in the first year of their term (or in some cases their second or third year) is not at all implausible. I'd say the opposite is more likely: this time next year we'd be having the same discussion the other way because the tables may well have turned completely. If there's one thing the tranche system has taught us, it's that fancy games with term lengths and staggering just do not work when there's such a high burnout rate. Happy‑melon 08:38, 11 November 2009 (UTC)

I concur with Risker. This RFC was to determine the parameters for the upcoming election.

I would like to see that we fall into a steady pattern where the majority of arbitrators endure their term to the end, and the number of seats available each year is stable, but I think we will need to take another look at it some time before the next election.

In regards to grandfathering, I tried to be clear about this from the beginning by prefacing each option with a note like "Based on 10 arbitrators being appointed to serve into 2010, 18 would result in 8 appointments in the upcoming election."

John Vandenberg ^(chat) 08:50, 11 November 2009 (UTC)

Why is the calculation I made about the need for "5 arbitrators to 2 year terms and 3 to 1 year terms" being repeated? I stated this explicitly in the view I posted on the RfC. The reason having 8 people appointed from this election to 2 year terms doesn't work is because you end up with 6 arb seats for election in December 2010, and 12 in December 2011. People have said that resignations will balance things out, but that depends entirely on who resigns and when. If someone resigns in their final year (a not uncommon event), it makes no difference whatsoever. i.e. Elect 8 arbs to 2-year terms in the December 2009 elections, and you end up with 12 on 2-year terms and 6 on 1-year terms. If no-one resigns during 2010, you have 6 seats available for the December 2010 election. If any of the 6 arbs with 1-year seats resigns in 2010, you still only have 6 seats for the December 2010 election. It's only if one of the arbs on a 2-year term resigns (that would either be one of the 8 newly-elected ones, or one of the 4 elected to 3-year terms in December 2008), that the number of seats for election increases. You would need three of the 12 to resign to bring the number of seats up for re-election in December 2010 to nine. Now consider if only two resign - leaving 8 seats to be filled in December 2010. That would mean that the following year (2011), you would have 10 seats expiring at the end of the year, and 8 not expiring until the year after. Again, if one of the 10 resign, it makes no difference, but if two of the 8 newly elected arbs resign, then you are back to the situation where 12 of the 18 seats are up for re-election in December 2011. In any system where you stagger the rotation, you need to make periodic adjustments to keep the numbers balanced, and natural attrition by resignation is not always neat or tidy enough to keep the numbers balanced, and sometimes in fact unbalances things further (e.g. no-one resigns in 2010, but two of the arbs elected in December 2010 resign the following year (2011), then you end up with 14 of the 18 arb seats being up for re-election in December 2011). The same unbalancing would decrease the number of seats available in the alternate year, in this case, from 6 in December 2010, to only 4 being up for re-election in December 2012. With a bit of bad luck, you end up with zero seats for re-election one year, and all 18 the next. Having said that, I can see that appointing people for 1-year seats can be a problem (it can make arbs more political if they intend to stand for re-election). Maybe the solution is to only have 5 seats available for 2-year terms, and to leave the three "1-year" seats empty. This would result in a committee of 15, with three empty seats, but would ensure that 9 seats could be contested in December 2010 (the 6 arbs with terms expiring then, plus the three empty seats). Carcharoth (talk) 00:35, 12 November 2009 (UTC)

I've posted a comment at the 'crat noticeboard inviting them to review this thread and (potentially) close the non-controversial parts of this RFC. See Wikipedia:Bureaucrats'_noticeboard#Partial_closure_of_ArbCom_RFC. Manning (talk) 05:18, 11 November 2009 (UTC)

Err...just asking: Why specifically crats? There are plenty of admins perfectly capable of closing any non-controversial parts. bibliomaniac15 05:28, 11 November 2009 (UTC)

Cos more people will accept crat X's word as impartial than will accept admin Y. Simple :D Happy‑melon 08:29, 11 November 2009 (UTC)

Agree with Risker and John Vandenberg. While it would be desirable for all arbs to see out their full term, there is strong evidence that this is not a practical expectation. Elections have functioned and will continue to function as a means of "topping up" those numbers, replenishing them back to what the community has determined is an optimal starting point each January.

Thus, there is no reason why the number of seats on offer at every December election should be the same. Theoretically, this number will trend towards nine as the terms of the three-year arbs gradually expire and given no attrition in the meantime, but if it is exactly nine for a particular election, that will be a coincidence and nothing to feel particularly satisfied about. What does matter is that there's sufficient overlap for experience and skills to be passed on to new arbs; I can't see that it will be an issue with a Committee membership of 18. Tony (talk) 10:33, 11 November 2009 (UTC)

What's up with "Theoretically, this number will trend towards nine as the terms of the three-year arbs gradually expire and given no attrition in the meantime"? You can calculate exactly what will happen if the three-year arb terms expire naturally and there is no attrition. It settles into a 6 seats for election one year, 12 seats for election the next year pattern. Those numbers only change if an arb resigns in their first year (or for the grandfathered three-year arbs, if they resign in their 2nd year). Arbs resigning in their final year have no effect whatsoever (as their seat would have been up for election that year anyway). If in one of the years where 6 will be up for election at the end of the year, a first-year arb resigns, then the balance is restored by one (7 will be up for election that year instead, and 11 the next). On the other hand, if a first-year arb resigns in the year where 12 are up for election at the end of the year, then that unbalances things even further (13 that year, and 5 the next year). There is probably a technical term for this sort of system where applying change in one half of the oscillation period increases the amplitude of the oscillation, and applying change in the other half damps the oscillation. I should in fact know that term, but it's been too long since I studied that sort of maths/physics. Carcharoth (talk) 00:51, 12 November 2009 (UTC)

See, the whole thing that makes this thought experiment so unrealistic is the "no attrition" part. We know there's tons of attrition. Carcharoth, your oscillating model is correct given the situation that Tony described, but that's a situation we're pretty sure won't happen -- it assumes all the three-year arbitrators will finish their term or at least resign with less than a year left.
Anyway, given a non-zero probability of arbitrators resigning before their election year, there are more chances for this to happen when there are more newly elected arbitrators. So this process will, in fact, trend toward a constant number of seats to be filled, and we can safely eliminate the archaic notion of tranches. rspεεr (talk) 03:21, 12 November 2009 (UTC)

You don't need to use probability. The essential point is that when a first-year arb resigns and you open up the seat for election that same year, you are calling the election for that seat early, and that introduces an imbalance. The next time that happens (the resignation of a first-year arb), the effect will depend on the timing. If it is the following year, the imbalance will be corrected, but if the next "first-year arb resignation" is two years after the last such resignation, then the imbalance will be increased. The general case is 2n+1 and 2n (odd and even years). Remember that second-year arbs resigning don't have any effect on all this. I'll try and explain this in more detail below. Carcharoth (talk) 04:04, 12 November 2009 (UTC)

I am aware of everything you have said, and what I added to it was probability. That was the entire point. It's of course true that sometimes it gets more balanced, and sometimes it gets less balanced, but on average it gets more balanced, and the probabilities come out like this because it's only first-year arbs that count.
Think of it this way: every year is either a "balance year" or an "imbalance year", depending on whether a newly-elected arbitrator would balance or imbalance the ArbCom election schedule by resigning that year. There are more newly-elected arbitrators in balance years than in imbalance years. That's why they're balance years -- because too many arbitrators just got elected. So the system tends toward balance. rspεεr (talk) 10:15, 12 November 2009 (UTC)

Agree with both Carcharoth (for unlikely no-attrition assumption) and Rspeer. The upshot is that we are very unlikely to see unmanageable oscillations. Tony (talk) 04:13, 12 November 2009 (UTC)

Poll on tranches

So people do have an opinion on tranches after all.

Would it be reasonable to make this a question on the project page? I'm thinking the topic should be "Tranches: keep or eliminate"? The two options would be:

• Keep tranches. The elected arbitrators will be placed into two "tranches" that determine when their term expires, so some arbitrators will be elected for one year instead of two. The tranches will be balanced so that at least half of the seats will be open at each election.
• Eliminate tranches. All arbitrators elected from now on get 2-year terms. The number of seats to be filled at each election will be determined only by the number of retiring arbitrators and the number whose terms naturally expire.

Comments? Objections to the wording? Objections to asking the question at all? rspεεr (talk) 04:14, 13 November 2009 (UTC)

Sounds good to me. Everyking (talk) 04:36, 13 November 2009 (UTC)

Either system would work fine, but I suppose an RfC section would be OK to discern general community opinion. The time of the election is fast approaching, so the RfC shouldn't be open for too much longer. —

*

04:42, 13 November 2009 (UTC)

I wouldn't object to this, but adding it to the RFC might slow things down. You are unlikely to get the same people coming back to give opinions on later questions, so you may get different answers. My view is that arbitration seats that are vacated over a year early should either be filled early at the next election, or left empty until the following year. Which of these two options is chosen should be done automatically depending on whether the number of seats up for election already that year is more or less than 9. This may sounds complicated, but in practice it would work. What it would mean is that with the current imbalance looking to be something like 6/12, then first-year arb resignations in 2010 (when there would be 6 seats, less than 9, up for election at the end of the year) would result in those seats being filled early (at the next election), reducing the imbalance to 7/11 if one first-year arb resigned. However, in 2011, when the number of seats up for re-election at the end of the year will be 12 (if there are no resignations), if any first-year arbs resign, you would leave their seats empty until the December 2012 elections. Second-year arbs resigning would, as always, have no effect on any of this. The only problem is what to do when a first-year arb resigns when the balance is at 9/9 - you would have a choice between leaving the seat empty and only electing 9 as normal at the end of that year, or you would have the choice of having 10 elected. My preference would be to keep it as close to 9 as possible each year, and accept that some seats will remain empty. Within reason of course - you would set a limit on the number of empty seats at any one time. On a committee of 18, I would say don't leave more than 3 empty after an election (i.e. let the number fluctuate naturally between 15 and 18, and don't insist on topping up to 18 every time). The reason, in case anyone thinks this is too complicated, is that I think electing to an 18-strong committee, 6 people one year and 12 the next, is not good. If in one year you have 9 good candidates (in terms of support in their votes), but only 6 seats available, and in another year you have 9 good candidates but 12 seats available, then you lose out on 3 well-supported candidates one year, and get 3 arbitrators in the other year with less community support than would be ideal. It's not important for this year though (as there are 8 seats for election), so I would hold all the arguments about tranches until after this election. Carcharoth (talk) 06:09, 13 November 2009 (UTC)

Interesting, but I don't think enough people are asking for a perfect balance of elections that it's worth overturning the fixed number of arbitrators. (Remember, we're coming off of a system where Jimbo put people in tranches on random whims and we didn't even know how many seats were up for election, and most people were still okay with that.) rspεεr (talk) 08:08, 13 November 2009 (UTC)

(Outdent) I don't see why they should be eliminated in the first place, though things would need to be rearranged if two year terms are implemented. When an arbitrator resigns, a special election should be held within a month or so to fill that person's term (or the person who finishes tenth in the December elections fills out the resigning arb's term in case of multiple resignations you just go down the line). Of course, with shorter terms less resignations are possible (I assume that's why people want shorter terms).

Interested editors are invited to participate in the SecurePoll feedback and workshop. SecurePoll was recently used in the Audit Subcommittee election, and has been proposed for use for the upcoming Arbitration Committee election at this current request for comment (RFC). Your comments, suggestions and observations are welcome.

For the Arbitration Committee,