Open main menu

Wikipedia talk:Requests for adminship

  (Redirected from Wikipedia talk:RfA)
RfA candidate S O N S% Ending (UTC) Time left Dups? Report
RfB candidate S O N S% Ending (UTC) Time left Dups? Report

No RfXs since 13:30, 11 June 2019 (UTC).—cyberbot ITalk to my owner:Online

Current time: 23:43:01, 18 June 2019 (UTC)
Purge this page

Recently closed RfXs (update)
Candidate Type Result Date of close Tally
S O N %
HickoryOughtShirt?4 RfA Successful 1 May 2019 182 19 9 91
RexxS RfA Successful 11 Apr 2019 164 92 15 64
Primefac RfB Successful 7 Apr 2019 151 7 5 96
DeltaQuad RfB Successful 13 Mar 2019 229 6 2 97
Evad37 RfA Successful 18 Feb 2019 212 15 6 93
Enterprisey2 RfA Successful 26 Jan 2019 253 2 2 99
De la Marck RfA WP:SNOW 4 Jan 2019 0 11 1 0

Contents

RfC on the nature of bureaucrats' discretionary range for closing Requests for AdminshipEdit

In future, should the discretionary range be understood primarily as A) a unit (bureaucrats should close Requests for Adminship based entirely or almost entirely on the strength of the arguments for supporting and opposing, and the raw percentage of supporters should not be a significant factor), or B) a spectrum (the strength of the arguments does matter, but the default expectation should be that RfAs with support near the upper end of the discretionary range are likely to pass and RfAs near the lower end are likely to fail)? Sideways713 (talk) 10:11, 12 April 2019 (UTC)

  • Extended initial comment Tempers have got pretty heated after the controversial closure of RexxS's RfA. I understand that many editors would like a longer cooldown period before any Requests for Comment even tangentially related to the RexxS case, and equally that many editors on the other side might like to see a much more aggressive RfC than this, one that proposes drastic changes to the RfA process or openly questions the appropriateness of the RexxS close.
    So I fully expect to be pilloried by everybody on both sides; but I think it makes sense to have this RfC first and have it relatively quickly, because it's something we ought to have a consensus on before other RfCs about the discretionary range can really work. Arguments that will be very strong in future discussions if the discretionary range is a unit can be much weaker if there's consensus to treat the range as a spectrum, and vice versa; so if we don't know which it is, we'll just be talking at cross-purposes. Moreover, 1) the answer will be relevant in any future RfA that lands in a 'crat chat, and there's no guarantee that's a long time away; 2) the answer is fairly unlikely to be rendered irrelevant by future RfCs (only happens if the discretionary range is abolished altogether); 3) it is, hopefully, if not completely noninflammatory then at least less inflammatory than some other RfCs might be; so I hope we can work together, concentrate on policy and avoid relitigating RexxS's RfA.
    I appreciate that the answer isn't completely binary and that the discretionary zone has both unit and spectrum properties, but I believe it's worth determining which nature, unit or spectrum, editors think is or should be more important. Sideways713 (talk) 10:12, 12 April 2019 (UTC)
Thank you for opening the conversation, but I might suggest brainstorming a little more to refine the question being asked. It seems that either option you’ve presented would allow for an RFA with 64.1% support to be deemed successful, yet many of those critical of the outcome felt that the 65% threshold should be a hard-and-fast breakpoint beyond which none shall pass. –xenotalk 10:19, 12 April 2019 (UTC)
I do realize there's an appetite for an RfC on the 65% threshold as a hard-and-fast breakpoint; but the strength of many arguments pro and con in that discussion will greatly depend on whether the range is seen as a spectrum or a unit, so I think that's something we should clarify first. Sideways713 (talk) 10:27, 12 April 2019 (UTC)
  • Thanks for bringing forward this RFC. I think we should strike while the iron is hot. I also agree that the scope of the RFC should be refined. Also, for at least the initial stage, the existing 'crats. should avoid participation, so that the community view as opposed to the 'crats view is clearly refined. They can contribute if asked and in any case later. It is my already clearly stated view that it is the 'crats themselves who created this mess and need to be accountable for it. Having said that I expect most who think the 'crats made perfectly the correct decision re Rexx will not agree and that is the difficulty - people will not agree to change something that suits their personal position. On the specific topic of the so called discretionary range - 65-75% is an 11% space for the exercise of discretion. That can be at the discretion of a 'Crat acting alone or as a collective. The introduction of a further zone of doubt below 65% is highly dubious alchemy. The community specifically rejected 60% - 65%. That embraces 64.1%. Simply put, the arithmetic boundary has to start somewhere. This is entirely consistent and since the 'crats have shown in successive cases to be incapable of consistent judgement it will help them at lot. Leaky caldron (talk) 10:56, 12 April 2019 (UTC)
    If the RfA had ended (164/88/15) [65.1%], would you still be contesting the result? –xenotalk 13:04, 12 April 2019 (UTC)
  • I think making some changes to Cratchat process rather than this more fuzzy 'how to look at it thing' might be better use of time, like majority of unrecused crats determining whether any !votes need to be stricken entirely; and a 75% of unrecused crats adopting a single reasoning statement when finding consensus. Throughout the pedia, editors working in small groups for consensus have to negotiate and find a a single way to say something and agreement on why they are saying it that way, all the time. Alanscottwalker (talk) 11:05, 12 April 2019 (UTC) (by stricken I generally mean given-0-weight, in case that's not clear - Alanscottwalker (talk) 11:49, 12 April 2019 (UTC))
  • Neither as you're presenting a false dichotomy between "the raw percentage of supporters should not be a significant factor" and "support near the upper end of the discretionary range are likely to pass and RfAs near the lower end are likely to fail". The reason we have crat discretion rather than straightforward votes is that not all voting is equal; in cases where the consensus isn't obvious, it's entirely legitimate that the closers consider whether those supporting/opposing are giving valid reasons for doing so, and that isn't immediately obvious. (There's a significant difference between "Support" without explanation and "Support per nom", for instance; while to take the specific RFA that prompted this, there's a world of difference between "Oppose. Don't like the guy" and the opposes who took the trouble to explain why they felt the candidate was unsuitable for adminship rather than just admitting they were expressing a personal grudge.)

    The raw numbers really aren't as significant as they're made out to be (personally I'd deprecate the ranges altogether and have the crats close purely on strength of argument); if we were to go down the mechanistic route of combining "all votes are considered equal" and "automatic pass/fail thresholds", it would be trivially easy to game any given RFA by flooding one section with sockpuppet votes two minutes before the RFA were due to close. ‑ Iridescent 11:16, 12 April 2019 (UTC)

    • Well put. I think the current misconceptions might be problem of the current (2015) wording "in general". The previous wording "historically" was more useful to highlight the fact that these are merely examples of the boundaries at which RFAs usually have succeeded or failed, not a rule that any given RFA has to fail/succeed if the support is larger/smaller than these numbers. I outlined some of the history on my talk page which shows that throughout RFA history, the wording was always clear that ultimately, crats should decide whether a certain RFX fails or succeeds. Regards SoWhy 11:36, 12 April 2019 (UTC)
      • I especially like the wording that had been in use in 2009: At the end of that period, a bureaucrat will review the discussion to see whether there is a consensus for promotion. This is sometimes difficult to ascertain, and is not a numerical measurement, but as a general descriptive rule of thumb most of those above ~80% approval pass, most of those below ~70% fail, and the area between is subject to bureaucratic discretion. Regards SoWhy 11:39, 12 April 2019 (UTC)
    • I know all that, and the reason for much of my extended initial comment was precisely to avoid giving the impression of any false dichotomy like that. Sideways713 (talk) 11:37, 12 April 2019 (UTC)
    • With all due respect, Iridescent (I was surprised which way you went on this one, but I was watching all through hoping to be persuaded that a change had taken place from when I became aware of this person and started avoiding them), that strikes me as a ridiculous fear. RfAs are scrutinised from all sides for signs of socking, and insofar as they are not votes it's because participants engage with each other, and often change sides as they're persuaded by others' points or evidence. How the percentages shift during an RfA is indicative. (I used to see the crats using such trends to discern consensus, before they started citing "the need of the community" for more admins as an overriding principle as one has here. In short: maybe I'm an idiot or demonstrating my unsuitability for a life of crime. Yngvadottir (talk) 19:15, 12 April 2019 (UTC)
  • RfA does what it is supposed to. Nothing needs changing at all at this stage - RfAs have become such a rare phenomenon that they can all be taken on an individual basis without having to re-politicise the whole thing. Discretionary range or not, a closing 'crat is not obliged to call for a 'crat chat - that is also part of the bureaucrats' discretion. Like this extremely close call which might well have gone the other way if there had been a 'crat chat - which it probably should have, but nobody was prepared to make a hooha of the result so why should they be suddenly doing so now? Is it just for the want of yet more drama? IMO, people should be looking at what's wrong with the voter mentality instead of down the wrong end of the telescope, and try to make RfA a more inviting environment for any potential candidates of the right calibre who might be left. Kudpung กุดผึ้ง (talk) 12:05, 12 April 2019 (UTC)
 
Explanatory figure
  • I'm not surprised to see people piling on to complain about how bad this RfC is; but much of the opposition seems to be to something that the RfC was never intended to be about, which is unfortunate. If that is my fault for not being sufficiently clear in opening it, I apologize for that.
    I hope the explanatory figure helps make the point clearer. This RfC was not intended to be about eliminating or limiting bureaucratic discretion; it's to determine whether people think Fig. 1 or Fig. 2 more closely approximates the way the discretion range works or ought to work. I hope the difference between Fig. 3 and Fig. 4 makes clear why I think we need to be clear on that before any RfC about 65% as a hard limit. Sideways713 (talk) 12:23, 12 April 2019 (UTC)
Maybe a wording like "does a higher support percentage in an RfA mean bureaucrats have a stronger mandate to promote an admin?" would have worked better? That's essentially equivalent to the intended question - if the answer is "no", that creates as a natural by-product a probability curve that roughly approximates Fig. 1 or 3; and if the answer is "yes", that gives a result that roughly approximates Fig. 2 or 4. - but a wording like that would (hopefully) make it clearer why there's no false dichotomy with the right answer being neither yes or no. Sideways713 (talk) 12:58, 12 April 2019 (UTC)
You didn't read what I said above, Sideways713. There was no mention of the quality of this RfC, but there were some key words like 'drama'... Kudpung กุดผึ้ง (talk) 13:15, 12 April 2019 (UTC)
I fully acknowledged in my initial comment that many editors would like a longer cooldown period after the RexxS case before any RfC goes up; I take it that you are one of them. Equally, though, many editors do think this is an appropriate time for an RfC; and maybe a bit of extra drama now will save us from even more drama later. Sideways713 (talk) 13:22, 12 April 2019 (UTC)
What Kudpung said (and note that when it comes to RFA, I very rarely agree with him). The mechanism of RFA is that people explain why they feel the candidate should or shouldn't be trusted, and the crats assess which comments should be taken seriously and what the consensus is. Talk of "discretionary ranges" and "cutoff points" is a misunderstanding of the data, which was based on an analysis of likely outcomes at varying levels of support and never intended to be enforceable cutoff points. The issue affecting RFA ultimately stems from a mentality among some people that the job of participants is to look for reasons to oppose, rather than only opposing if they feel something genuinely precludes the candidate from being given the sysop bit; trying to impose hard limits or pass/fail cutoff points is never going to be the way to address this. It's not that we want a longer cooldown period after the RexxS case before any RfC goes up, it's that your RFC is based on an utterly false premise which you're repeatedly trying to present as fact despite the rest of us repeatedly explaining to you that it's not the case. ‑ Iridescent 13:25, 12 April 2019 (UTC)
This RfC is not intended to impose a hard limit or a pass/fail cutoff point. Sideways713 (talk) 13:28, 12 April 2019 (UTC)
I was surprised by the Crats decision, not disappointed as I believe RexxS will make a good admin. But I was surprised, and if I were a crat and a comparable cratchat came up I would probably be in the no-consensus position. I'm concerned that if we were to hold an RFC now it would be too heavily influenced by the most recent RFA, there are other scenarios where a Crat might judge consensus as being outside the normal 56-75 zone. For example here is a fictional scenario: 24 hours before the normal end of the RFA the candidate makes a disclosure that sees support fall by 24% in 24 hours with the only activity supporters striking and moving to oppose. I would hope that a Crat looking at such a situation with support dropping from 100% to 76% in 24 hours would not feel that they had to close as successful, or even start a Crat chat or extend for further consideration. I would hope that a crat would read the consensus and close as lacking consensus to promote. Going back to the recent RFA, there were at least four reasons given for giving less weight to some opposes in that RFA. I think it would be good to give the crats some steer as to which if any of those reasons are acceptable to the community. I've mentioned extrapolating the trend, which doesn't particularly apply to the recent one and I doubt would be contentious. But other reasons for going beyond the discretionary zone include:
  1. Being more lenient with longstanding members of the community. I don't agree with this, not least because the RFA !voting community is quite capable of doing this where appropriate, so no need for crats to give extra weight where the community chooses not to. I think an RFC could usefully clarify things for the crats on this.
  2. Giving less weight to opposes over an issue where the candidate has given an assurance that the issue won't recur. In the past this has been more likely to sway the community on things like a signature that doesn't follow the rules or a default to minor edits that has resulted in many non minor edits being flagged as minor. It is fairly easy to interpret a situation where after the candidate responds such as by fixing the signature there are few or no further opposes over that issue and some existing opposes strike. Less easy where some opposers strike but some new ones oppose despite the reassurance.
  3. Use of humour by the candidate or nominator. My own first RFA used too much humour for some of the RFA crowd, and I'm happy to see humour at RFA, but I wouldn't recommend it to anyone who wants to pass, and I'm not convinced that Crats should discount Opposes from people who don't get or appreciate the joke.
  4. Opposes without rationale. Personally I'm OK with such Opposes where there is already a substantial and unified Oppose section, though I'd recommend adding the words per above. Where there are two or more significant reasons being cited by Opposes then I'd prefer if Opposers said which of those reasons they agreed with. So I wouldn't have struck or discounted such Opposes in the recent RFA. But I would support discarding such votes if there are no Oppose rationales already there for them to be agreeing with. We give the typical Oppose close to twice the weight of the typical support, so while I'm happy to treat a support without a rationale as being per nom, I would be happy if an RFC reaffirmed that Opposers should do the courtesy of explaining why they oppose (there is also the practical matter that we need admins, and if people are turned down at RFA they need to know why if they are to be tempted to run again in the future).
  5. Weak strong etc weightings to votes. I'd hope we are all OK with the idea that if someone marks their own vote as weak they are inviting the crats to give it less weight. I'm tempted to argue that less weight should also be given to people who try to up the impact of their own vote with prefixes such as strong, but I'm going to assume that an RFC would simply affirm that such verbiage be ignored. I doubt if the community would be at all bothered if the crats were to close a 63% RFA as successful because there were lots of votes marked as weak and they were almost all opposes. Or conversely if a 77% support RFA failed because many of the supporters had marked their !votes weak.
  6. Something we have seen at RFA is when people !vote per another editor and that editor subsequently strikes their !vote and moves to another section. I think that in such circumstances we should have RFA clerks invite such editors to return to the RFA and possibly replace their rationale. But I can see this as an interesting one for crats, and there is an argument that if RFA is a discussion and at the end there are a couple of per user:example !votes when user:example has switched is difficult to see how one can give full weight to them.

Not all of the above have been seen in the last RFA, and that might be reason to wait a while before an RFC. ϢereSpielChequers 13:36, 12 April 2019 (UTC)

(edit conflict)Iridescent is right o the money and I'm afraid, Sideways713, that I don't really understand what you do want, and I'm not the stupidest of people when it comes to RfA reform. What I do see is that you didn't get the result you hoped for despite the drama you added to the RexxS RfA, and now you appear to be asking for solutions that are looking for a problem. Kudpung กุดผึ้ง (talk) 13:52, 12 April 2019 (UTC)

  • We get these 'emergency' or 'flood-of-bogus' scenarios that are really not in the least 'scary'. Not only crats but basically every other person would go, 'yep that's an emergency, handle it.' Alanscottwalker (talk) 14:12, 12 April 2019 (UTC)
  • I want a polite and constructive discussion on policy (as opposed to the RexxS case), giving us a clear answer to a question to which we do not currently seem to have a clear answer ("does or does not a higher support percentage give bureaucrats a stronger mandate to promote an administrator?"), so that future RfAs will create less drama than RexxS's did. If we get a consensus that the answer is "yes, it does", I will consider that a win for Wikipedia; and if we get a consensus that the answer is "no, it doesn't" I will consider that a win for Wikipedia also. Either way, future RfAs will cause at least a bit less drama, and future RfCs will have more to build on. If there's no discussion and no answer, that will only cause more drama further along the way. Sideways713 (talk) 14:14, 12 April 2019 (UTC)
Sideways713 it's all to do with the RexxS case - because that's why you began this discussion. AFAICS, you were directly involved in some of the drama there so I fail to understand how you can be campaigning for less drama at RfA. Untill the voter community gets potty-trained, there will always be drama at RfA, And due to that, RfAs will take place in ever decreasing numbers., I urge you to read some of the links you've been given and familiarise yourself with the history of RfA ad its reform over the last 8 or 9 years. Kudpung กุดผึ้ง (talk) 14:55, 12 April 2019 (UTC)
  • In principle I'm in favour of no hard cutoff at the high and low ends. But if a discussion is right at one of the extreme ends of the discretionary range, the bureaucrats should have extraordinary reasons for closing it the other way. Canvassing and sockpuppetry would be suitable grounds for chucking out a pile of oppose votes. "I don't agree that civility is a concern" is not. Reyk YO! 14:19, 12 April 2019 (UTC)
    Reyk: Could you point out which of the bureaucrats in the discussion indicated they did not agree civility was a concern? –xenotalk 14:35, 12 April 2019 (UTC)
    That's the impression I got from the fact that all the civility opposes were so severely devalued. I'f I'd meant for that to be taken as a direct quote, I'd have given a diff. Reyk YO! 14:46, 12 April 2019 (UTC)
    So you're putting forth a paraphrase (in quotation marks) that has no basis in the subject discussion, and has no adherents. I'm not sure where the impression that opposes on civility grounds were "thrown out" or "severely devalued" comes from. If a bureaucrat found consensus to promote, that doesn't mean they are ignoring the oppose section, simply that on the balance they found the discussion leaned towards promotion. –xenotalk 15:16, 12 April 2019 (UTC)
    Zeno, an example of where I personally got a similar impression that opposes were being "severely devalued" from was when you said "I’m also concerned there may have been somewhat of a pile-on effect with participants [expressing an oppose view] merely looking at the words used by the candidate without fully exploring the context and circumstances behind the comments." Well, yes... some participants expressing an "oppose" view may have merely looked at the words used by the candidate - neither you nor I have any no way of knowing. Equally, some participants expressing "support" may have been piling on as well, and may not have looked at the words used by the candidate at all - again, neither you nor I have any way of knowing! I know in my case, I read quite carefully around the various cases that came up, and I'm prepared to assume that other editors did so as well, including those who disagreed with me in the RFA. Even if you didn't mean it to be (!), I found it a bit insulting to be unilaterally told I may have been part of a pile-on effect, and therefore my view shouldn't carry as much weight. NB: I think an RfC at this point would be too soon, and feelings are running too high. Time for everyone to draw breath, think a bit, and come back to it in a few weeks. Hchc2009 (talk) 16:24, 12 April 2019 (UTC)
    Hm - I'm sorry that it came across that way - and my comment was not at all meant to disenfranchise those who opposed the candidate on civility grounds. I wouldn't say those comments need to be "severely devalued" to come to my opinion, either. In my more active past (and certainly wherever I see it), I've advocated for bitten newcomers and feel strongly that they are one of our most valuable assets and did not take the civility concerns lightly. It may help if I explain how I come to an opinion when participating in a bureaucrat discussion: approach the RfA with an open mind and project's core principles in mind, after closely reviewing the submissions from the community, determine if, on the balance, the consensus is that the community would be better served by having another administrator, sometimes in spite of legitimate concerns noted in opposition. It's a balance scale, certainly more art that science, in trying to divine the community's will, and in the present case I found the balance in favour of consensus to promote. In my mind, the opposition advanced due to civility did not need to be "severely devalued", for the scale to remain on the side of promotion. I do my best to articulate my thoughts on how I come to my opinion in each bureaucrat discussion. I must be honest in my opinion, and others may disagree with my opinion, as is their prerogative. –xenotalk 17:06, 12 April 2019 (UTC)
    Thanks for replying Xeno, and, while I do think they were unfortunate, I equally don't imagine that you intended for your remarks to come across that way. Hchc2009 (talk) 17:23, 12 April 2019 (UTC)
    I think DeltaQuad's claims that people were hypocrites and WP:POINTy was a bit of a stinky thing to say too. Reyk YO! 16:31, 12 April 2019 (UTC)
  • Too soon - I think per WereSpielChequers above, there are some legitimate discussions to be had here, and the fact that I don't agree with all his observations is testimony to the fact that some thought is needed (e.g. (a) I would have seen consensus for RexxS promotion if I were a crat, and I say that as someone who sat out the RfA because I wasn't sure how to !vote. And (b) I think oppose !votes based on the jokey nature of the RfA should be given less weight, as to me they show a disregard for WP:CIVIL and WP:AGF. Although there are very often excellent reasons for opposing a candidate it is always going to be a hurtful action if you do. By default anyone with the basic level of experience and nous should pass, and opposes should be made for genuine reasons of thinking they'd make a bad admin, rather than just that you don't like their sense of humour.) But, with all due respect to the OP of this RfC, I think feelings are still running too high and, per Iridescent, I don't think the RfC is asking the questions that most need to be asked. I therefore suggest we wait a few months and then brainstorm some proper ideas with WereSpeilChequer's questions above as a starting point. Thanks  — Amakuru (talk) 14:53, 12 April 2019 (UTC)
I did not participate in the RFA, and I would find entirely differently as I wrote on the the cratchat page, and I think that's either disrespectful or a misunderstanding of what the concern was around the 'humour issue', for lack of a better term. Alanscottwalker (talk) 15:01, 12 April 2019 (UTC)
@Alanscottwalker: Well you have your view and I have mine, that's fine I fully respect that. That's why it's a legitimate question for debate. My intention was not to insult those that cited the humour issue or even to say their opinion is wrong per se, just that for me, if you take into account the goals and pillars of Wikipedia, I personally think that is a poor reason to Oppose an adminship. Others feel differently and we can come to a consensus decision in the RFC. Cheers  — Amakuru (talk) 16:21, 12 April 2019 (UTC)
I do think this is a question that needs to be asked eventually; it may not be the question people most want to see asked, but it's one that answers to those questions will depend on. Unfortunately it does look like feelings are still running too high for a policy discussion to be possible without it getting personalized; and this RfC has clearly strayed very far from its intended purpose. Sideways713 (talk) 15:19, 12 April 2019 (UTC)
When an Arbcom member says this: "I feel strongly that this was a poor close by the crats, essentially supervoting away valid opposes because they were not convinced by them, while, as usual, being far laxer with the supports. It does look like excessive leniency for an "establishment" RfA candidate, and it is particularly grating that civility concerns were dismissed given mounting evidence that a hostile culture is stifling the project." there is every reason to examine what led to that and to resolve it. The 'crats are incapable of self-resolving so let the community do it. Speedily. Leaky caldron (talk) 15:21, 12 April 2019 (UTC)
And what about the six or so arbitrators who feel and spoke differently? Looking at the nine arbitrators, there is a stronger consensus towards this being a good call by the 'crats than otherwise. Do you disagree? -- Avi (talk) 15:24, 12 April 2019 (UTC)
As far as I can see, Joe Roe was the only arbitrator to directly opine on the goodness or the badness of the close. Several arbitrators called it a good call by Maxim to open the chat in the first place, but without giving any views on whether the chat's actual outcome was good. Sideways713 (talk) 15:39, 12 April 2019 (UTC)
(edit conflict) Avi, my reading of the declines is not so much an endorsement of a good/bad result and more as a statement that the specific actions were not within the purview of arbitration. — xaosflux Talk 15:42, 12 April 2019 (UTC)
  • They made a decision within their remit and according to their procedures;
  • In fact, I cannot see that even a first resort process is needing used: excepting the filing party, the community appears to be content that processes worked as designed in this case.
  • As someone who participated in the 2015 RFC and provided the second support vote for lowing the discretionary range, I believe the bureaucrats acted in accordance with the instructions provided to them by the community.
  • Closing RfAs is up to the bureaucrats, and I think it was a good call on their part to get wide input on the closure amongst themselves.…It would be an overstep for us to revisit the 'crats decision, which I think shows no sign of improper behavior that would justify our involvement.
  • Not going to set some precedent of running here if a discretionary call ends the way someone doesn't like.
The point being that while the close was not, and would never be universally approved, it was neither outside the burecrat remit nor some other violation. -- Avi (talk) 15:49, 12 April 2019 (UTC)
I'd appreciate if anyone who is going to advance the claim that bureaucrats are "supervoting" to head over to Wikipedia:Supervote include definitions on what constitutes a bureaucrat supervote and provide examples. –xenotalk 15:30, 12 April 2019 (UTC)
Does anyone have any historical insight on why I keep seeing RexxS called an "establishment candidate"? 1, 2, 3, to name a few. I'd never even heard of the guy until the 'crat chat, but I've been semi-active (to really stretch the definition of that word) for a number of years. Has he served in various Wikipedia capacities in the past that I'm unaware of? This is not supposed to be a knock on RexxS, I apologize if it came across that way. I also never heard of User:Joe Roe who is the ArbCom member that said the quote above, so it's nothing personal against RexxS. It's my own lack of recent activity, that's all. Useight's Public Sock (talk) 15:41, 12 April 2019 (UTC)
It's an unsubstantiated claim that RexxS is an "establishment candidate" (whatever that means in this context) whom the bureaucrats wanted to appoint as an administrator. Those of we bureaucrats who did want him as an admin, such as me, supported him in the RfA and recused from the close accordingly. Anything to the contrary starts to descend into the tin-foil hat-wearing zone. Acalamari 16:15, 12 April 2019 (UTC)
I can think of four reasons why people may consider RexxS an "establishment" candidate. He has been active on Wikipedia for a very long time. He has been involved in the UK chapter for many years, he has been involved in outreach for many years and he is involved in WikiMed. The UK link is particularly strong, looking at the supports and opposes I spotted a dozen people who I know are in the UK, and only one was an oppose. Wikipedia is big, I can't remember many on Wiki interactions that I have had with him, I know him largely from meetups and editathons in the UK. ϢereSpielChequers 16:28, 12 April 2019 (UTC)
  • Neither – I feel the question is too narrowly-focused and we would benefit from a broader conversation about RfA reform and community de-sysop before any specific proposals are made. I've started a thread with one idea at WP:Village pump (idea lab)#RfA reform: straight vote?. Levivich 16:14, 12 April 2019 (UTC)
  • I think what is meant by a 'establishment candidate' is that RexxS has been around for a long time, has been highly active in areas that are not reflected in his edit count on-Wiki, and is certainly of an age and maturity where he doesn't 'need' the mop on his CV or to brag about in the schoolyard.
On another note, WereSpielChequers and I have been heavily discussing RfA on and off, on- and off-Wiki, for a decade and we mostly agree on most things. I'm not so keen on his #4, but #5 and #6 are important - especially #6. Overall though, on the hundreds of RfA I've participated on, the community has usually reached (rightly or wrongly - and more often right than wrong) a clear outcome; where the decision has been left to the 'crats, there has only been, IMHO, one outcome which I disagreed with. It was an extremely knife-edge but very polite RfA on which a single 'crat extracted a consensus where a 'crat chat would certainly have been a much safer option. What I'm saying is however, that where I was very disenchanted with the outcome, I didn't make a song and dance about it and didn't call for RfCs to get the votes more closely examined or cut-off values renegotiated.
To suggest that the 'crats are , or have been guilty of supervoting would be highly infelicitous. RfA nevertheless remains to this day the one venue where editors can rejoice at being drama mongers and as nasty as they like with almost total impunity. That's where reform is needed. Kudpung กุดผึ้ง (talk) 16:23, 12 April 2019 (UTC)
  • I don't want to knock the OP for opening a discussion in good faith, but I think the discomfort myself and presumably several others felt about the outcome here is because of a more fundamental problem that this does not address. Essentially, we as a community have decided that consensus is always based on the strength of arguments, not on the numbers favoring each argument (the one exception that proves this rule are the ARBCOM elections). However, even with such analysis, outcomes of discussions are not clear; there is a continuum of outcomes from "unanimity in favor of (X)" to "unanimity against (X)", and encompassing everything in between. RFAs usually fall somewhere along the continuum, and so of necessity we need some sort of threshold to turn a continuum of outcomes into a pass/fail. While acknowledging that this threshold cannot be numerical, it is also obvious that this threshold should broadly be consistent, otherwise the process is meaningless. In this case, I felt it not to be; or to put it another way, if the outcome here was okay, then Jbhunley got a really raw deal, and that's extremely unfortunate for many reasons. It's unfair to him personally, it's unfair to the community as a whole because we've been deprived of the services of an extra admin, and it's doubly unfair to the community because it makes RFA (already seen as an extremely unpleasant process) also appear capricious. Both these RFAs are now closed, but we need to be discussing how to find this balance between basing consensus on the strength of arguments while also keeping it reasonably consistent. Vanamonde (Talk) 16:24, 12 April 2019 (UTC)
I supported both JBHunley and RexxS, but having just revisited the earlier RFA I don't see them as similar other than in raw numbers, and raw numbers are not the whole issue here. You could conclude that the crats were taking the reasons cited for opposing JBHunley slightly more seriously than they took the reasons for opposing RexxS. Or that they gave greater latitude to a WikiMed person who had been here 11 years over an editor who had been here 6 years. Or that one got cut some slack for commitments made during the RFA. The truth is that both were close, both could have gone either way, and the crats have to make a binary choice when there is clearly a close call. ϢereSpielChequers 17:00, 12 April 2019 (UTC)
@WereSpielChequers: While I agree the two were not identical, my comment wasn't just about the numbers (broadly similar, slightly more support for JBH) but about the nature of the opposition, and more specifically about the applicability of some of the crat comments in the Rexxs crat-chat to Jbh's RFA, and vice-versa. While I supported both, I recognize that they were both judgement calls. My complaint, though, is that the reasons to give less weight to the opposition to RexxS also applied broadly to Jbh. I'm particularly concerned by the comments of WJBscribe and Nihonjoe, who supported promotion in the latter case but not the former; and I'm concerned because AFAICS, their analyses in each case could be applied in its entirety to the other discussion. At the moment it's unclear if this result was because the crats are hearing community feedback about RFA having too many overly-demanding naysayers, or whether it's stochastic variability. If it's the former, that's great; if it's the latter, we need to talk about making it less variable. Vanamonde (Talk) 17:22, 12 April 2019 (UTC)
Quoting WJBscribe "The Opposition is largely concentrated on one issue in relation to which limited evidence is presented despite the candidate being a longstanding contributor." I see those as three arguments two of which are radically different between these two RFAs. The RFA that failed had two major arguments in the Oppose section and an editor who had been here six years rather than 11. I'm not sure that I agree with WJBscribe on either of those, I don't see a case for cutting extra slack for our longest serving editors. As for whether the argument in the Oppose section has focused on one argument or not, I could put the other case, there were two main arguments for declining the other RFA, but if they had been voted on separately it is likely that over 75% thought that the candidates contributions were sufficient for RFA and a not quite overlapping 75% thought that their behaviour was acceptable. I'm not sure if I agree that a united Oppose section should be given less weight than a divided one, but I would agree that on those two criteria those two close calls go in opposite directions. ϢereSpielChequers 17:45, 12 April 2019 (UTC)
My thoughts pretty much mirror those of WereSpielChequers, who phrased them much better than I would have. ···日本穣 · 投稿 · Talk to Nihonjoe · Join WP Japan! 18:02, 12 April 2019 (UTC)
Indeed it was. In fact your change of course, as it happens from one favorite to another, was especially poorly explained. Looks like you just wanted to go with the mood shift. I mean, repeating something is "a hard one" doesn't really explain your rationale, does it? Leaky caldron (talk) 18:22, 12 April 2019 (UTC)
@Leaky caldron: Thank you for the backhanded compliment, Leaky. It was a "hard one" because I was pretty much on the fence with it, and could have gone either way. Initially, I decided one way. After reviewing the comments made by a number of the other 'crats, I decided I agreed with the reasoning they gave, and so I changed my opinion. There's nothing nefarious here, despite what you may think. ···日本穣 · 投稿 · Talk to Nihonjoe · Join WP Japan! 19:31, 12 April 2019 (UTC)
  • I didn't participate in RexxS' RFC (because I didn't know it was happening) and I've not read the Crat Char or almost all the subsequent drama, however I am firmly against numerical cutoffs, strong or weak. Every RFA should be assessed on the strength of the arguments presented - two dozen !votes based on a subsequently-clarified misunderstanding are worth much less than a single !vote that is based on things everyone agrees are accurate (even if they disagree on whether they are good or bad). Thryduulf (talk) 16:59, 12 April 2019 (UTC)
  • I appreciate the good faith of opening this discussion and I have nothing negative to say about it having been opened, so thanks. In one sense, yes, I do think that the higher the support percentage, the stronger the rationale for closing as successful, and that trend applies pretty much across the entire range of results and not just in the discretionary zone. But in another sense, I feel strongly that, per WP:VOTE, the numbers just do not matter that much and no numerical cut-off should be regarded as binding. (The example given earlier, of an RfA in which new information comes forth late in the process, is a good example.) It's intentionally hard to pass RfB, so crats are entrusted by the community to use discretion, and they should be able to do so. I also think that there is a logical flaw in arguments that there could be a problem because there was one crat chat where the decision was to promote, and another with a very similar percentage where the result was no consensus. The logical flaw is the assumption that a given percentage always indicates the same thing about community consensus, and it's a flaw for the very reason that we have !votes rather than votes. Two different RfAs that both ended at 64% are not automatically equal in terms of how the discussions played out. And that is exactly what I do want the crats to evaluate. --Tryptofish (talk) 17:56, 12 April 2019 (UTC)
  • I am not a statistician. I don't understand the question, and the diagrams don't help. But RfAs are now advertised with a notice that has had the desired effect of increasing participation. And the hard-argued RfCs on lowering the discretionary range established that the community wanted the discretionary range lowered to start at 65% and rejected lowering it to 60%. The job the bureaucrats swore they would do in their RfBs is to interpret consensus in the RfAs. A discretionary range is a discretionary range: outside it is not discretion, it's wilfully changing the result (and persuading someone to break a repeated promise doesn't make it better). RfA is enough of a popularity/influence contest already. Deprecate the bureaucrats' changing results that do not fall within the discretionary range. Whether they do it because they like someone or because they think having more admins is so imperative it doesn't matter if we get an abusive one is immaterial. Fit that into the RfC however you statistically literate people see fit. Yngvadottir (talk) 18:00, 12 April 2019 (UTC)
  • At last!! the DISCRETIONARY Range BEGINS at 65%. Below is outside Discretionary range - obvious, simple logic. Leaky caldron (talk) 18:13, 12 April 2019 (UTC)
  • Then we're back to bean counting. In this case, after weighing the strength of the arguments, in my opinion, there were more than 65% weighted responses in favor. I'd say that Enwiki either says we have absolute hard and sharp boundaries at the bottom, or they trust the discretion and resoning of the 'crats. We've always had the latter, knowing that 30%+ of participants in an RfA may be annoyed. Switching to the former requires a site-wide RfC. -- Avi (talk) 18:25, 12 April 2019 (UTC)
  • Forgive me Avi, this is not intended to be as critical as it sounds but are you being deliberately obtuse? You guys - 'crats - have a discretionary range of 11 percent in which to do your sums, play your games, pontificate, muse and opine. Why do you need more? The community expressly rejected less than 65% in an RFC. Leaky caldron (talk) 18:29, 12 April 2019 (UTC)
  • Will you stop? They're fulfilling their roles as bureaucrats. If we wanted someone to just total up a numeric percentage, bots already do that. Natureium (talk) 18:36, 12 April 2019 (UTC)
  • That's going too far. Reyk YO! 18:48, 12 April 2019 (UTC)
  • (edit conflict) I hope he/she doesn't stop. "Bean counting" = respecting what the voters say. We are only too aware of your opinion, Avraham. It's that the RfA was an unnecessary formality because you and your friend crats know better. I'm coming to be ashamed that one of you once promoted me. This was not discerning consensus, and you can't be trusted to do so any more. Let us vote and stay out of it. Yngvadottir (talk) 18:54, 12 April 2019 (UTC)
  • Me too. As I said on BN, it seems like the crats used their discretion to ignore votes related to the process to bring the percentage up to 65%, then ignored the other oppose votes by closing the request as successful. By the current consensus, 65% should be the minimum, after which crats can engage in their arbitrary reasoning at the crat chats. -- Ajraddatz (talk) 19:25, 12 April 2019 (UTC)
  • Question are the outcomes of RFCs exempt from WP:IAR? Asking for a friend... The Rambling Man (talk) 18:58, 12 April 2019 (UTC)
    • Bureaucrats don't need to invoke IAR. Yngvadottir (talk) 19:17, 12 April 2019 (UTC)
      • Is that indoctrinated in policy, or just your opinion? Asking for a friend... The Rambling Man (talk) 19:18, 12 April 2019 (UTC)
        • I might be wrong, it has been known :), that the convention when resorting to IAR is to quote WP:IAR in the rationale for whatever the IAR is being used. Leaky caldron (talk) 19:21, 12 April 2019 (UTC)
          • Don't get me wrong, I'm not sure if IAR is a reasonable explanation here, or not, but the point is, if IAR is a policy, and 'crats aren't forbidden from using such policy, then all this bluster over 65% and discretionary ranges etc, is a complete and utter waste of time. Either get consensus that 'crats can't IAR or move on. The Rambling Man (talk) 19:24, 12 April 2019 (UTC)
            • To belabour the point, the crats have demonstrated that they consider their opinion to be the rule, and hence have no need for that nicety. Yngvadottir (talk) 19:39, 12 April 2019 (UTC)
              • What you mean is, the 'crats, like Arbcom, have exercised their own judgement, even if the community don't agree with it? Is that what you're belabouring? If IAR is something that can't be applied to 'crats and Arbcom, we should be told, right? The Rambling Man (talk) 19:41, 12 April 2019 (UTC)
                • Based on the principle behind IAR, I don't understand why it shouldn't apply to crats.Natureium (talk) 19:46, 12 April 2019 (UTC)
                  • Because they promise at RfB to merely discern consensus. Yngvadottir (talk) 19:50, 12 April 2019 (UTC)
                    • Ah, so "discern consensus" means they are not allowed to IAR? Could you be clear here? The Rambling Man (talk) 19:52, 12 April 2019 (UTC)
  • This RfC question is so badly worded that I can't even respond to it. It actually seems like you don't even understand what the issue is. The discretionary range is a general rule that is almost always followed, with strength of arguments factored in according to policy considerations and common sense, just like any other discussion. I don't think anyone disputes that. The actual question is whether 'crats can judge a 64% scenario to be in the discretionary range by disqualifying illegitimate opposes. That's what happened. That's the controversy. After discounted opposes, RexxS' RfA would be in the discretionary range, so they passed an RfC that did not even hit the minimum percentage. Some users feel that regardless of the merits of the opposition, the vote count is the vote count, and being outside of the discretionary range is be a quick-fail. This RfC does nothing to resolve this actual issue. ~Swarm~ {sting} 19:36, 12 April 2019 (UTC)
    • (I think you mean an RfA). That's not what I saw happen. One of them opened a crat chat because participants disagreed in the RfA (shock, horror). Several crats then decided to discount opposes by "weighting" and based on disagreement on the major issue raised. They persuaded the candidate to rescind his withdrawal. They persuaded at least one of their fellow crats to fall in line. They were determined to change the result, and did so. Yngvadottir (talk) 19:46, 12 April 2019 (UTC)
      • I think that's gross assumption of bad faith. I suspect, given that I trust most of our 'crats a damn sight more than the Arbcom, that they saw a corner case and needed to act on it responsibly and in keeping with their ability to do more than just divide one number by the sum of two numbers. All this "persuasion" and "determination" that's being perceived in my opinion is hysteria. The Rambling Man (talk) 20:03, 12 April 2019 (UTC)
        • I hope I'm right in seeing bad faith here, because otherwise they're idiots. Don't let's bring Arbcom into it, one set of powerful people telling common editors their concerns don't matter is enough for this week. Yngvadottir (talk) 20:14, 12 April 2019 (UTC)
          • Once again, you're not answering the question. And yes, it's important to include Arbcom here. Are 'crats (like admins, and it would seem Arbcom) allowed to IAR, or are 'crats the only subset of the Wikipedia editing community which are precluded from exercising such a policy? The Rambling Man (talk) 20:17, 12 April 2019 (UTC)
            • Michael Howard [1] and yes they are. Leaky caldron (talk) 20:34, 12 April 2019 (UTC)
              • Did you order the code red? Yes, 'crats are allowed to invoke IAR because it's policy and everyone here is bound by policy. So this is a pointless exercise. If, however, someone wants to start an RFC which says "below 65% support, no-one may even invoke IAR", then you have your RFC. Right now, this is yet another timesink. The Rambling Man (talk) 20:38, 12 April 2019 (UTC)

Arbitrary breakEdit

  • Another historical question. When Biblioworm made the edit summarizing the 2015 RFC, did anybody voice any concern with the wording? I scoured the related talk pages and archives looking for anyone lodging a complaint about the wording, but couldn't find anything. The closest I found was a short conversation that took place in August 2018 here, which wasn't about the 2015 change at all, but about a small 2018 change that ended up getting reverted. Does anyone know of any conversations that took place shortly after the 2015 RFC regarding how the results were codified and summarized by Biblioworm? Or is this the first instance? Useight's Public Sock (talk) 19:47, 12 April 2019 (UTC)
  • B? This is a very weird RfC question. In general, my guess is that the likelihood the bureacrats see consensus to promote scales with the actual vote tally. This isn't so much a policy or unspoken practice as it is common sense. More votes in support = higher probability the community came to consensus to promote. That said, figuring out current practice is hard since very few RfAs close at or near the discretionary range. By my count at Wikipedia:Requests_for_adminship_by_year, since December 2015 there have just 9 that ended up within a couple points of the range. Pbsouthwood (77%) and Oshwah (75%) closed as consensus to promote without a 'crat chat. Jbhunley (70%), Godsy (68%), and Hawkeye7 (67%) were closed as no consensus following 'crat chat. GoldenRing (67%) was closed as consensus to promote following 'crat chat. Between 60 and 65% the only closes we've had are Philafrenzy (64%) closed as no consensus with no 'crat chat, and RexxS (64%) closed as consensus to promote after 'crat chat. If you chart those, you'll find a mess. In general, I'd say the 'crats rarely find consensus to promote when the vote tally is under 75%. In some rare cases, they judge that there is consensus to promote. Put another way, the expansion of the discretionary zone in 2015 has changed the result of at most 2 RfAs. I don't think the answer to this RfC would further change RfA outcomes, so I'm not totally clear on its purpose. Some folks obviously feel that there should be a hard numerical cutoff under which 'crats should not consider promotion. If you feel that way, perhaps start an RfC on that instead. Ajpolino (talk) 20:13, 12 April 2019 (UTC)
  • This is what was actually passed [2] according to the closer, Neonjoe, by a supermajority. At the same time reducing the lower edge to 60% was firmly rejected [3]. 64.1% lands below 65% - there is no fuzzy margin here. The added verbiage was added outside of this supermajority decision. Wrongly. Leaky caldron (talk) 20:25, 12 April 2019 (UTC)
    I feel like you're replying to my question with your response, but it isn't actually answering the question. I do not and will not get into philosophical discussions. My question remains (and is not directed at you, but at anyone who might know): Does anyone know if Biblioworm's edit in December 2015 was made without incident until now or was there prior discussion about the wording? Useight's Public Sock (talk) 20:45, 12 April 2019 (UTC)
    Did that RFC state that 'crats could not IAR nor could they dismiss objections (e.g. "I don't like the guy") which clearly fall outside the normal expectations? Just asking, because all this bluster about 65% etc is fine, but once again, we're just dancing around the edges of whether or not 'crats are entitled to act with IAR in boundary conditions. The Rambling Man (talk) 20:31, 12 April 2019 (UTC)
  • I commend Sideways713 for attempting to clarify this issue. RfA has always benefited from and been plagued by the vague nature of voting/!voting. The RfA from which this RfC came forward is simply the latest iteration of many. This RfC can't and won't clarify this. For all its hideous ugliness, the system is working as intended. For those that are gravely concerned about the outcome of Rexx's RfA; be aware that RfA is an absolutely horrible predictor of whether an admin will fail and lose their bit. The last 10 admins who lost their bit for cause passed RfA with a combined tally of 928/52/29. The community expressed an enormous amount of confidence in those 10 former admins, only for them to subsequently lose their bit for cause. Rexx's RfA is not a predictor one way or another whether he will succeed as an admin. This, too, shall pass and we will continue as we have been. --Hammersoft (talk) 20:40, 12 April 2019 (UTC)
    Wow, recentism works, or is it just lies, damned lies, and statistics? I was elected second time round to admin 128/0/0. Then look what happened. And even to 'crat a couple of years later. I would take the preceding comment with a pinch of salt. I have little doubt that RexxS will do good things with the bit, especially in light of all this unnecessary heat (which, incidentally, should not be any reflection at all on him). The Rambling Man (talk) 20:44, 12 April 2019 (UTC)
    And then I would take my own comment with a pinch of salt. Hammersoft nailed it (is that allowed?). The final sentence was spot on, to whit: This, too, shall pass and we will continue as we have been., my apologies to Hammersoft for spending too much time watching The Victim on the iPlayer while editing... The Rambling Man (talk) 21:28, 12 April 2019 (UTC)
  • Hahahah! TRM, you are the first person to ever (even if unintentionally) make that play on words. Bravo sir, bravo :) --Hammersoft (talk) 23:09, 12 April 2019 (UTC)
  • The last RFC expanded the discretionary range, I don't believe that it made it more rigid as it would have done if the proposal had been to change the wording such as from "below the range will normally fail" to "below the range will always fail". What I do wonder is whether we have now moved de facto into an era where adminship is such a big deal that all close calls have to go to crat chats. ϢereSpielChequers 20:42, 12 April 2019 (UTC)
    I believe what is being sought here is hard and fast. (a) above 75% - promote (b) between 65% and 75% - chat (c) below 65% - fail. And this means completely ignoring the substance of the votes being cast. The Rambling Man (talk) 20:46, 12 April 2019 (UTC)
    Exactly. Bureaucrats are appointed to assess substance of comments made. If the community wants to go off percentages only, then the only way I can think of it working is making RFA like ArbCom elections, with established users only and hidden voting. Aiken D 20:55, 12 April 2019 (UTC)
    Indeed, at which point we might as well remove the position of 'crat from Wikipedia altogether as it's only really nowadays serving a middleware purpose of promotion to adminship. Let's just hand that off to Arbcom and actually get them to do something useful and timely. The Rambling Man (talk) 20:58, 12 April 2019 (UTC)
    Just FYI:The last RfC also rejected discretionary ranges below 65[4] [5]. Alanscottwalker (talk) 20:59, 12 April 2019 (UTC)
    But the 'crats HAVE a 65%-75% range in which to use discretion. The arithmetic has to start somewhere. 65%. Leaky caldron (talk) 21:01, 12 April 2019 (UTC)
    @the Rambling man I'm aware that there are several people on this page who want to make such a change, and even who think that this is the current policy and that the Crats have breached it. There are also people such as myself who are uncomfortable changing RFA to a rigid discretionary zone. Just to repeat the hypothetical that I gave earlier, on the seventh day of an RFA Support slides from 100% to 76% as a series of supporters shift to oppose. Currently a sensible crat would close such an RFA as a failure. Is there anyone who wants such a rigid discretionary range that a crat should close such an RFa as successful? ϢereSpielChequers 21:03, 12 April 2019 (UTC)
    Your description of a 100% -> 76% slide (or more importantly, a 74% or even 64% slide) is why we vote for and enable our 'crats to make hard decisions. I trust them. I don't trust Arbcom at all, but 'crats are worthy. They perform their duties openly and honestly, are elected openly and honestly, and sometimes have to call it, and this time they did. It's not a problem, and IAR exists for this very reason. Once again, if someone wishes to revoke 'crats' ability to act subjectively based on their assessment of various opposes (e.g. "I don't like the guy" for god's sake), then make that the RFC. Or if someone thinks IAR doesn't apply to this dozen or so of the millions of us, then make that the RFC. Meantime, get over it. The Rambling Man (talk) 21:08, 12 April 2019 (UTC)

Can I clarify with those opposed to the 65% minima - what does 65% mean? Leaky caldron (talk) 21:09, 12 April 2019 (UTC)

65% is merely a number derived from two other numbers which happen to be based on a subset of editors who chose to preface their comments with a #. To set the system such that candidates who fall even slightly below 65% automatically fail, without any regard to the actual comments behind those #’d remarks would be to allow form to triumph over substance, and I do not think this is what the community wants from bureaucrats. Since we disagree on this very fundamental aspect, until an RFC answers that question in your favour or mine, I don’t think we can come to any kind of mutual understanding, unfortunately. –xenotalk 22:28, 12 April 2019 (UTC)
In other words, what I said below about half an hour ago. The Rambling Man (talk) 22:31, 12 April 2019 (UTC)
  • I think it means that you add up all the supports (A) and then add up all total votes (B), divide (A) by (B) and if it's 0.6500000 or above, you are allowed to be "discretionary" and below 0.6500000 you must fail and you (i.e. the assessing 'crat) must not pay any attention whatsoever to the substance of any of the votes, all must be treated equally, like a secret vote, like we do for "Arbcom". The Rambling Man (talk) 21:16, 12 April 2019 (UTC)
    I answered your question regarding IAR fairly. I would appreciate your interpretation please. Leaky caldron (talk) 21:18, 12 April 2019 (UTC)
    I just gave you my interpretation. What are you looking for? The community seem to need a minimum below which they cannot trust 'crats to make a decision. 65% apparently seems to be the current minimum. The Rambling Man (talk) 21:20, 12 April 2019 (UTC)
    In which case I am sorry that I have misinterpreted your previous remarks as being opposed to that. I am hoping to gain an understanding from those opposed to 65% minimum what they believe 65% means and you responded under that heading. Leaky caldron (talk) 21:23, 12 April 2019 (UTC)
    Ah yeah, well I'm neither in favour of, or opposed to a 65% minimum. I think it's nonsense. Any numerical answer which simply relies on 'crats dividing (A) by (B) renders their role lower than a lowly Arbcom clerk. Anyone can tweak numbers, be told to refactor or redact something, divide two numbers and check it's above 0.65. Is that really what we have 'crats for these days? I'm glad there's a growing consensus against there being any issue here whatsoever, but if this ever got more traction to suggesting that 'crats overstepped their bailiwick, then I'd immediately opt to eliminate 'crat as a position now they're not needed for renames, and if judging borderline RFAs is gone too, what's the point? The Rambling Man (talk) 21:31, 12 April 2019 (UTC)
    Is a 65% - 75% borderline not sufficient though? Does it have to be a Donald Trump sized border to police? Leaky caldron (talk) 21:34, 12 April 2019 (UTC)
    I don't see the point in a borderline at all. The 65%-75% you describe is purely numerical, so it includes votes like "I don't like the guy" (wow, I mean, WOW). Should we be doing that? I don't think so. I think 'crats should be given the latitude to completely ignore such ridiculous votes, and if you can find a consensus that says we shouldn't allow 'crats to do that, I'd be surprised. So immediately, the % argument starts to lose its way. The Rambling Man (talk) 21:37, 12 April 2019 (UTC)
    [6] [7] Leaky caldron (talk) 21:40, 12 April 2019 (UTC)
    I don't see anywhere in either of those links where it says that 'crats can't discount blatantly disruptive votes like "I don't like the guy" and hence I don't see anywhere in either of those links where 'crats can't make subjective decisions over what constitutes a reasonable vote, and hence I don't see anywhere in either of those links where 'crats shouldn't be enabled to do what most of us voted for them to do, make important decisions without resorting to bean counting. And even then, if necessary, IAR. Numerical accuracy is perverse here when you have clearly disruptive voting tactics. The Rambling Man (talk) 21:44, 12 April 2019 (UTC)
    Maybe they should have said which votes they were discounting in order to arrive at the 65% to begin their discretion consideration? Leaky caldron (talk) 21:52, 12 April 2019 (UTC)
    Maybe, and then of course every user whose vote was discounted would be up in arms, and it would become a never-ending story, hence perhaps we should vote for admins like we do our "Arbs". And leave it to someone else to do the vote counting and announcements. All in secret. And as I noted, that way we could get rid of the 'crat position entirely. The Rambling Man (talk) 21:55, 12 April 2019 (UTC)
    every user whose vote was discounted would be up in arms- you say that like it's a bad thing. If someone participates in a process in good faith, and then their opinion gets chucked on the trash heap, I think they're justified in wondering why. Reyk YO! 22:01, 12 April 2019 (UTC)
    Well, the one example I gave thus far was "Oppose - I don't like the guy", so that's a prime candidate for being "chucked on the trash heap". Time to start getting real here, and avoid pandering to those who are clearly not engaged in the process of electing someone to admin. If someone is prepared to give a suitable reason for oppose or support, than we should trust our 'crats to take those votes into consideration. Votes based on whether you like someone should be thrown out and the voter should be banned from future voting as disruptive. The Rambling Man (talk) 22:04, 12 April 2019 (UTC)
    I'm no fan of trollish opposes in RfA's either. I've often objected when candidates are opposed for such extreme felonies as not being a native English speaker (though fluent), or being a fan of a particular author, or of editing articles that some people don't find interesting. Yet people tell me I need to respect those opposes. Which is why I'm annoyed when actual good faith opinions then get tossed in the garbage. BTW, your example of "I just don't like the guy" is a horrible example because Softlavender followed it up with a measured reply in response. Reyk YO! 22:27, 12 April 2019 (UTC)
    No, it's far from horrible because it's an exemplar of the kinds of votes that 'crats have to work with all the time, and in close-run cases, they need to exercise judgement which is why they were elected in the first place. Simply put, if we want RFA to work purely numerically, then say so. If we want 'crats to work purely within the 65% to 75% range after the pure numerical closure had happened, then say so. If we want 'crats to reject anything below precisely 65% from the pure numerical closure, then say so. If we want 'crats to be disallowed any latitude on interpreting the validity of votes, then say so. If 'crats are not allowed to exercise IAR, then say so. Right now, I'm not hearing any real solutions, just a lot of bitching about one instance where a few people appear to be "butt-hurt about the situation" (to use a US vernacular). The Rambling Man (talk) 22:37, 12 April 2019 (UTC)
    RfA does not follow the patterns we normally use when judging consensus. The supports are invariably variations on "I think this editor would make a suitable admin, and the opposes are insufficient to convince me otherwise". The opposes are generally more colourful, often detailing grudges and grievances, with the occasional "we have too many admins already" and "I couldn't vote for someone who urinates in the sink". Hawkeye7 (discuss) 22:33, 12 April 2019 (UTC)
  • I agree there's a hint of false dichotomy, but logically it has to lean toward the "unit" side. It can't really be that 'Crats are bound to almost always pass someone near the top of the discretionary range, and almost always fail those lower in it, or it is not in fact a discretionary range, but a pretense of one, and we have no use for fake process that creates a bureaucratic layer "just to have one". It has to actually serve a function, and the specified function is for 'Crats to carefully examine the arguments presented. We already know the numbers are iffy, because it ended up in the discretionary range at all. So, making the 'Cratchat outcome dependent on any kind of numeric approach defeats the entire purpose of remanding the candidacy to a 'Cratchat in the first place. But of course the 'Crats are not robots and will of course take numeric support into account as they're going over the RfA, so it is a factor and will remain one (that's the false dichotomy part).  — SMcCandlish ¢ 😼  01:10, 14 April 2019 (UTC)
  • My thoughts are to decline to engage directly with Sideways713's question as proceeding from a false premise. I can see why Sideways713 is taking the approach chosen here and recognise that it is being done in good faith, but trying to define and then mathematically describe the nature of the discretionary zone proceeds from a premise that the decision being taken is one capable of precise modelling – and that is false. The same fallacy is seen in a focus on specific values of the S / (S + O) ratio.
    The purpose of RfA is to determine whether there is a consensus amongst community members who choose to participate to grant access to the sysop tools. It is not about winning or losing a prize, nor about filing a fixed number of positions with editors whose views you favour, it is not about governance or politics or popularity. Now, it is true that !vote counting can provide information on which to decide whether consensus exists, but that is all it is – a tool used to provide a means that assists in making a decision. In many cases, the result of such a calculation makes consensus clear and a single 'crat makes the call on consensus and close the RfA accordingly. In others, the presence or absence of consensus is unclear and a 'crat chat is held. We as a community held an RfC some years ago to modify the "discretionary zone", which I see as another tool to assist 'crats in deciding the result. Our discussion showed that there was a majority view that standard practice at RfA at that time was adopting a threshhold for consensus that was inaccurately reflecting community views, and we asked / directed the 'crats to lower the threshhold and gave guidance as to how this should be applied. To me, the idea that that RfC dictated to 'crats that they must find consensus at and above 75% support, must find no consensus between 75% and 65%, and must find no consensus or consensus against when below 65%, is an interpretation inconsistent with the nature of the decision and instead reflects the faulty premise that I mentioned above. It implies that the RfC changed the decision from being about consensus to having a narrow range for considering the strength of arguments sitting inside a broad range where only !vote counting mattered. I do not accept that as an accurate view of what happened.
    I have no problem with the 'crats having a 'crat chat on a 77% support case, and even finding no consensus so long as the reasoning shows an honest and reasonable evaluation of consensus. I have disagreed with the outcomes of 'crat chats before and there have certainly been cases where my opinion differed from that of some 'crats or of the 'crat consensus. However, I can't recall one where I viewed the conclusion as objectively unreasonable / unsupportable and where I thought the actions had gone beyond the limits of the discretion they are granted to carry out their duties. It would only be if that happened that we'd need to act on a policy level, or through RfAr to remove individual 'crats. I am glad the 'crats concluded as they did, but much more important, I am grateful that they took their duty seriously, carefully examined the discussion, and offered their views and reasoning for all to see. We may disagree with some comments, but I think we should be respecting the collective decision reached after a serious and mutually respectful discussion that itself demonstrated the emergence of consensus.
    I understand that some editors feel RexxS is not suited to being an administrator and are disappointed that others have disagreed. I understand that the outcome reached is not the one that follows from strict !vote counting... but that has never been the basis for judging consensus at RfA. But, please, consider whether attempting to change the nature of RfA is something you really want, reflect on whether 'crats as judges of consensus have done anything worse than come to conclusion about which you disagree, and ask what there is to gain from these discussions. If you truly believe that one or more 'crats, or even the 'crat group as a whole, have so violated their obligations and granted boundaries of discretion that they are unfit for the role, you can take it up with them directly or with ArbCom and seek sanctions... but before doing that, ask if that is truly what you think. Disagreeing is not a wikicrime, it's a part of a healthy community and productive discussions and debate. Deliberately abusing discretion is a wikicrime – and in this case it's a charge that logically would apply to at least seven 'crats, or more if you see even having / participating in the 'crat chat as a betrayal – but I struggle to see how that view can be sustained. Finally, give some thought to RexxS, who didn't ask for this controversy and must be feeling stressed by all that's going on. At least, let's try to allow him the space that all new admins need to grow into their new role and its responsibilities. EdChem (talk) 03:45, 14 April 2019 (UTC)
It is a bit of a wonder that any of this is all that confusing for an experienced pedian (although sure, I think it would have been great to wait for awhile on these serial RfC's, etc.). Something happened in the community that almost had never happened before, so yes, it's going lead to things. The 'real world' does 'after-action' review all the time, so it can't be all that unexpected (we just are messier). Alanscottwalker (talk) 21:58, 14 April 2019 (UTC)

Neutrally-stated SummaryEdit

The discussions have been all over the place (literally on various pages and topically splintering into tangents), making it somewhat convoluted. I went through it and made a list of all the questions and concerns I came across. There could certainly be more (and others could arise in the future), so feel free to add to the list. I am not commenting on any of these, nor am I stating that I condone or condemn any of these questions or their (potential) answers. I'm merely listing them for the sake of clarity. I tried to phrase each concern as neutrally as possible, phrased as a yes/no question. They are also in no particular order.

  1. Should April Fool's festivities be banned from RFA areas?
  2. Should April Fool's festivities be banned from the site as a whole?
  3. Is the discretionary zone bounds of the 2015 RFC a hard 65%-75% or fuzzy?
  4. Do the bureaucrats have the ability to IAR when it comes to the discretionary zone boundaries?
  5. Was the discretionary zone fuzzy before the 2015 RFC? If so, was it still fuzzy afterwards?
  6. Was it within the bureaucrat's remit to open a 'crat chat for RexxS' RFA?
  7. Did any or all of the bureaucrats participating in the aforementioned 'crat chat act in bad faith, COI, etc?
  8. Did any or all of the bureaucrats participating in the aforementioned 'crat chat use discretion in a substandard way or explain their discretion in a substandard way?
  9. Should bureaucrats be more explicit in which !votes are given full weight and which ones are not and why?
  10. Was the RexxS 'crat chat, or any portion of it, a supervote?
  11. Is more leniency provided for longer-tenured candidates?
  12. Can RFAs or 'crat chats be compared? Is precedence a factor?

I don't know if that helps provide any clarity or not. Feel free to completely ignore. Useight (talk) 01:46, 13 April 2019 (UTC)

Order of operationsEdit

So, I don't really want too get deep in to this dramafest right now - but here is something to consider: when "weighing" votes - I don't necessarily consider all votes to be equally weighted (as I clearly stated in my RfB). So when closing an RfA: in general I weigh out the votes and if it results in <65% I'm likely closing as unsuccessful, and if its >75% I'm likely closing as successful - in between requires more consideration, and possibly, but not necessarily, extra help (the 'crat chat). So to note here, I look at the weighing first, then the calculation. I certainly don't expect every 'crat to weigh every every discussion point the same as I do, so there certainly could be variance between crats here. The point of this is: weighing may be performed prior to percentage support determination by at least some of us crats. If the community would like a strict numerical cut off somewhere (computed prior to any weighing) then continue in the section above - if not, this order of operations may be what is confusing the causal observer. — xaosflux Talk 23:53, 12 April 2019 (UTC)

  • Not sure about these tailing level three headings, but this follows the preceding post. Consensus is not about the numbers. Understandably, there was a curiosity about studying the RfA numbers and their correlation with whether the bureaucrats promoted. Unfortunately, this led to an obsession with numbers, and with time, the perception of the correlation with the numbers hardened, and now, for some, their is a belief in the numbers at the expense of consensus. Consensus is not about the numbers. It is about strengths of arguments, and whether some arguements eclipse others, and whether some arguments overcome others, and whether some arguments are more persuasive to others in the discussion. The recent RfC on shifting the discretionary zone should have been written differently, and should be interpreted as the community telling the bureaucrats that they should be passing more RfAs. —SmokeyJoe (talk) 13:55, 13 April 2019 (UTC)
    • But there is still a percentage element to consensus. Unanimity is the ideal, which is 100%. 2 people have different views and they come to agreement and you have a consensus of 100%, if the two don't come to agreement you don't have a consensus, and they are at 50-50. (See also meeting of minds.) As you scale larger, it becomes harder to have agreement of all, so it becomes acceptable to expect less, but not down to 50% - 50% again, and 75 is between 100 and 50. (And the issue, here, of course, is an admin is an admin for 100% of us. Also, a discussion among 300 is, well, a bit imaginary). As I have said elsewhere, it would be good, if the cratchat modeled making consensus in a group, like explicit statement that an !vote is given 0 weight (say, by majority of crats), and a group statement of consensus that, say 75% of crats can agree too. All over the pedia, small groups of editors come to consensus on what to say, how to say, and why to say something, all the time. Overtime you would get a better informed !voting population, too. Alanscottwalker (talk) 14:40, 13 April 2019 (UTC)
      Drawing the line at 75% would require 3-out-of-4 but also 4-out-of-5, 5/6, 6/7, 6/8, 7/9, 8/10, 9/11, 9/12, and so on. As practically applied to small groups of voters (like crats), it would require near-unanimity. It would be strange if five-out-of-seven or seven-out-of-ten people agreed on something, and yet that wasn't considered to be "consensus". Levivich 21:24, 13 April 2019 (UTC)
      So, 70%. But are you assuming, for say 10 people of similar interest and background to discuss something together and come to agreement is either impossible, or a miracle? -- Alanscottwalker (talk) 21:35, 13 April 2019 (UTC)
    • Alan, the percentage element of consensus is overblown. Unanimity is not consensus, unanimity is what happens when you have unanimity. If you have unanimity there is not need and no point talking about consensus. Consensus is about making a decision, taking into account disagreement. If there is no disagreement, the feature function of consensus is irrelevant. Counting numbers is about discounting a minority view based on numbers. Consensus is about considering more seriously the minority disagreeing view. Many Wikipedians have trouble with the concept of consensus, and try to substitute numbers. —SmokeyJoe (talk) 21:44, 13 April 2019 (UTC)
      Partly right. Consensus is still agreement, if you have 100% agreement you have consensus. But, yes if you don't have 100% at the beginning, you work on the differences, until you come close enough. Alanscottwalker (talk) 21:52, 13 April 2019 (UTC)
      100% right. —SmokeyJoe (talk) 23:11, 13 April 2019 (UTC)
      All-right! Right-on! Alanscottwalker (talk) 23:18, 13 April 2019 (UTC)

Time to look at the bare facts - and really look...Edit

 
Successful RfA drop by almost half each year, while the pass rate has increased from 41 to 55% since 2015 – graph by Oshwah

The RexxS RfA has created a lot of comment and spawned many mini RfC recently, but the significant facts and stats have not been taken into account - more precisely clouded by the facts that the drama is due to the RexxS opposers not having got their own way, a lot is based on conjecture, and nobody appears to be bothering to check out the history of research into RfA as a process or to read some essential recent articles about it before commenting:

Something dramatic must have happened in 2007/2008 after which both the successful RfA and total number of RfA began to drop roughly by half each year shown on the (Wikipedia:RFA by month overview table by WereSpielChequers) until we were left with only 10 'promotions' last year. Research has failed to reveal the root cause of this phenomenon, but it could lie in a combination of forces such as the gradual unbundling of tools, a perceived sharpening of voters' criteria, the irresistible attraction for voters to be disingenuous and offensive with impunity, the doubling of the number of participants since the 2015 reform, and perhaps simply a general apathy for maintenance tasks - although this latter is not reflected in the regular clamour for minor rights. What ever it is, editors of the right calibre who have been approached have very often clearly stated that they are not prepared to go through hell on wheels for seven days however good their chances may be. Some of the key people such as Bishonen, Spinningspark, Jo-Jo Eumerus, Amakuru, Seraphimblade, Guy Macon, Cullen328, MER-C, Risker, TonyBallioni, Anna Frodesiak, MelanieN, and Spinningspark who don't regularly weigh in here my wish to come up with some ideas.

 
Only 10 new admins in 2018

The stats page at WP:RFAY created 2016 by Hammersoft provides yet again a similar overview with the additional undeniable evidence that the vast majority of successful RfA generally pass by a large consensus irrespective of pre or post-reform, and that very few were anywhere near the discretionary zone whatever it happened to be at the time. In fact while over 2,000 adminships have been created, only 28 RfA were ever subject to a 'crat chat, something like only 1.4%. Extrapolate that, and the next borderline RfA might not be for years at this rate.

This all seems to demonstrate that the current spate of discussions since the RexxS RfA are based on a set of artificial premises, none of which really need to be addressed. As I mentioned earlier, the community is looking at the problems associated with RfA down the wrong end of the telescope. Fix the voters and RfA will fix itself: that is the more pressing issue, and has been since Wales' famous statement about RfA being "a horrible and broken process". Horrible, certainly, but perhaps not beyond repair.Kudpung กุดผึ้ง (talk) 23:20, 18 April 2019 (UTC)

Ah yes. RfA reform. I have some observations:
[1] If you ask "is RfA broken" you get 60% to 80% consensus, depending on exactly what you ask and how you ask it.
[2] Any specific proposal for fixing the problem never gets even 10% consensus.
[3] You rarely if ever see any specific proposal that has not been proposed and rejected several times already.
[4] I have been asking for over ten years for somebody to give me a good reason why I or anybody else would want to go through hell to become a Wikipedia administrator. Nobody has ever presented a good argument. Mostly nobody even tries.
[5] I know exactly how to pass an RfA if I ever go batshit insane enough to want to become a Wikipedia administrator (see above). Spend one or maybe two years never disagreeing with anyone on anything. Stay away from all noticeboards. Spend a couple of hours every day writing new articles on noncontraversial subjects, and abandon the article the moment anyone disagrees with you on anything. Participate in things like AfD, new page curation, vandalism reverting, etc., but never ever on anything where there is the slightest chance that there will anything other than overwhelming agreement with you. Finally, having demonstrated that you have zero interest in doing any of the tasks that admins actually do, run for RfA. Once you get the bit, do the bare minimum of noncontroversial admin work required to avoid an inactivity desysop.
I would love to see someone argue that I am wrong on any of the above points. --Guy Macon (talk) 02:55, 19 April 2019 (UTC)
Maybe I'd disagree a little on #5; I think some people will want to see how you handle a conflict.
Me, I am wondering if people just are less interested in becoming admin. Or perhaps more exactly, that the kind of editor who wants admin buttons thinks they'll struggle in RfA while the type of people who will pass uncontroversially just doesn't need access to the admin tools. Jo-Jo Eumerus (talk, contributions) 08:07, 19 April 2019 (UTC)
EC@Kudpung. We know why RFA dropped off a cliff in early 2008. Rollback was unbundled, in that era you had to have Rollback in order to use Huggle, and before the unbundling of Rollback it was possible to pass RFA simply as a "good vandalfighter", after that particular RFA reform it ceased to be possible to pass RFA without some content contributions (there have since been various abortive attempts to raise that further to "some featured contributions" but the change in early 2008 is stark and known). There are multiple theories as to subsequent decline in RFA numbers, and yes an obvious anomaly in that effective vandalfighting does really require the ability to block vandals and protect pages.
@Guy Macon There is an old trope at WT:RFA to bemoan the inability of various proposed reforms to get consensus, and yes there are a set of perennial suggestions that keep recurring. But one reason why some keep recurring and getting in some cases over 50% support, is that there are several obvious and sensible reforms, and eventually we work through the opposition and fine tune the proposal or the idea comes up when the community consensus has shifted. Even a reform as obvious as making the page the equivalent of extended confirmed protected took about a decade. The last big reform package included two reforms, advertising RFAs on watchlists and lowering the discretionary zone. The greater advertising has greatly increased the voters, though sadly not yet the candidates. The change to the discretionary zone led directly to the recent RFA success. As for your method of passing RFA, yes that would work, but as recent RFAs have demonstrated, it isn't the only way of passing. for starters, how you respond to controversy is important. In terms of motivation, some people want the extra bits, others are content not to have them. As a nominator I start with the former group. As for it being "hell". RFA is timeconsuming, a contentious RFA is stressful, but most successful ones are more of an inauguration, and some of the last few have had support levels approaching 100%. I'm one of those who passed at the second attempt, and the contrast between my two in stress terms was marked. ϢereSpielChequers 08:12, 19 April 2019 (UTC)
  • The chances of any kind of revolutionary change are close to nil in my opinion, but since I have been (double) pinged, I'll give you my views (entirely without evidence) as to why there is a problem. There are two broad reasons, neither of which can be fixed by tinkering, and neither of which the community is capable supporting a radical solution for.
  1. As others have mentioned, the RFA process is hell to go through. That is no way to appoint people to a task, most especially volunteers. Having the whole community interrogate is just crazy. The candidate will never know what the requirements are because any random issue can be raised by any random editor. This is not how things are done in the real world. The solution is to give the task of appointing administrators to a finite sized appointment board. The appointment of the board itself can be open to community participation, but once they are in place they are solely responsible for appointing admins. The board would be small enough that we don't need to worry about finding enough candidates to go through the process. After all, the Arbitration Committee does not seem to have this problem. Hell, this could even be a function of the already existing ArbCom. SpinningSpark 11:13, 19 April 2019 (UTC)
  2. Administrators have very limited powers. I know this is an unpolular thing to say; the general perception is that administrators are too powerful. They particularly do not have any power to resolve content disputes. Think about it, we have been demanding content experience from candidates for years, but these content creators are given no powers in the very area that is of interest to them. In fact, there is a problem here that goes beyond the lack of administrators. That problem is that Wikipedia has no formal process for definitively resolving content disputes. This results in even minor disagreements becoming major wars with no hope of resolution. I have suggested elsewhere that Wikipedia should have an Editorial Board (you know, like the ones all grown-up encyclopaedias have) and if such a board were implemented it could set the boundaries for administrator intervention.
When Wikipedia was young and developing, it was an exciting thing to become an administrator. Now the community has matured, I think there is a better understanding of just how limited the administrator role is. That combined with the difficulty of going through an RfA pretty much explains, to me at least, why there are so few candidates. SpinningSpark 11:13, 19 April 2019 (UTC)
  • As long we have the occasional Arbitration Case over administrator conduct (for example, the one that just started), there is going to be that perception among random voters that candidates have to be extraordinarily vetted and qualified to ensure they don’t abuse the powers. I would argue that the arbitration case linked above, necessary as it is, will simply add fuel to the opposition fire. If Enigmaman was able to get away with the giant mound of examples in the evidence section, how can we be sure a new admin is not going to get away with the same? I know there are a lot of problems with that argument, but tell that to the pile-on opposers who can make a ridiculous claim look just credible enough to be usable at RFA. As an aside, I do not think most of the recently appointed admins in the last couple of years have been overly bad at abusing power. The bigger problem I see with newer admins as a normal content editor is their being over-zealous in using the tools and throwing their weight around. While an overzealous admin with good intentions is easier to handle (and repair damage from) than one who skits the fine line of abusing power, I don’t think either one is helping RFA candidates at the moment. ZettaComposer (talk) 12:03, 19 April 2019 (UTC)
  • Radical or semi radical?
Quarterly selection dates (and Crats may call selection, at any time, generally related to the number of candidates expressing interest). Crats will also decide on incidental timings related to quarterly selection, and generally act as election committee in addition to final consensus maker).
Selection run similar to Arbcom elections provided that for winners, there be at least 60% of secure poll, and 75% Bureaucrats in the finalizing chat agree. (Crats who participate on a candidates page or in secure poll are recused) Candidates are still well advised to have nominators. Others may write voter guides if they wish. Crats may run the chat anyway they see fit provided that it is done openly (and it is expected that discussion will center on the issues raised on candidate pages, the level of secure poll support and anything else they see relevant), they are free to use straw polls among themselves any other consensus building mechanism. They may delegate scrutineer of the secure poll itself to remove votes to an internal group or single crat. (The secure poll question: Should [candidate] be forwarded to the crats for consideration?)
Further, it is the general expectation given to Crats to that about 1/3 of the candidates standing at the quarterly will be selected, barring good reason. -- Alanscottwalker (talk) 14:10, 19 April 2019 (UTC)
(I can't see any way to make admin more attractive to those of us who don't find the job itself attractive, and I have never seen anything that convinces we actually need more admins, because we do not know how many we need, but if the desire is for new blood and the thinking is the current selection process is broken, let's try something different with a melding of the past). Alanscottwalker (talk) 14:10, 19 April 2019 (UTC)
  • Spinningspark has an exceptional breakdown above: the problem is that RfA is too big of a deal, and adminship is too small of a deal proportional to the amount of work it takes to become one. As I've said before, nobody wants to spend years working on Wikipedia like it's their full-time job only to be able to mindlessly delete pages and block IPs at the end of it. Past RfA reforms were passed with good intentions but have been complete policy failures in terms of attracting new candidates.
Because this is a website and we have literally invented all of our policies and processes, there is endless possibility for reform. The obvious solution here is to make RfA proportional to adminship -- make the selection process closer to PERM. This would involve setting criteria for adminship, and a community comment period that is focused on those criteria. But I also think that deep down, the vast majority of the community really likes the status quo. Admins like it because it makes them feel good about passing such a big hurdle. The Content Creators™ like it because it means their "enemies" can't get access to more buttons. The dispute resolutioners at AN/ANI like it because it means they can continue to act as the final authority without pesky new people running around with the block button. It prevents bad people from becoming admins at the expense of preventing an uncounted number of good people, or good enough people, from accessing the toolset as well. We can put as many bandaids on RfA as we like, but in five years we'll still be wondering why nobody wants to step forward. -- Ajraddatz (talk) 16:53, 19 April 2019 (UTC)
  • In many ways I agree with Spinning Spark, but note an asterisks for one (semi-large, now, area). We have been unable to grapple with having a content board, mainly because there is no verifiable expertise, so what did we do instead, Arbcom stepped into the breach to create wide swaths of the pedia subject to plenary discretion of admins. (Further note, it looks like the job is unattractive, in part because large swaths of admins don't do it, we don't even expect them too, and basically say, 'here you go, after all that, be functionally inactive.') Alanscottwalker (talk) 17:40, 19 April 2019 (UTC)
  • Our RfA standards are currently the lowest they have been in at least a decade, and while to one degree or another I agree with both Ajraddatz and Kudpung, I will again point out that I think the biggest problem is we need people to run. To address some of the points above:
    • The level we expect now is about right for what you'd want someone with administrative access on the 5th largest website in the world. Yes, we are just a website, but we're not the same just website we were in 2004. +sysop can't really do anything too terrible from a technical perspective, but it can create a ton of PR problems, which is just as important, and why I can't really fully get behind Ajraddatz' view here.
    • The January 2017 crop showed that we are willing to promote acceptable candidates en masse and if anything, the standards have remained the same or gotten lower since then. The question just becomes why aren't people volunteering? And there are several possible answers to this, but three come to mind:
      • We're losing people who are active editors to recruit
      • People who are qualified aren't interested or are scared off
      • We already have most of the people who would meet a reasonable RfA standard as sysops
I think the second one isn't really the answer: RfA has had a terrible reputation for years, and we haven't had this low of a number for a while. The real answer, I feel, is somewhere in between the first and third options.
There are plenty of editors who could make decent admins: somewhere between 10-20 a year who are interested, apparently, but while we are maintaining stable editor numbers, a significant number of new editors, even those who stay around, wouldn't meet a reasonable standard for administrator status on this project. I'm not talking full-time work, I'm talking basic understanding and competence of culture and policies and who don't have other issues that are incompatible with adminship on this project.
I suppose the point I am making is that as a maturing project, this is to be expected, and thanks to the advent of bots, etc. we don't really need 1200 active administrators anymore. We do need more admins, yes, but we aren't in a crisis by any means, and reflecting on where we are as a project today compared to where we were as a project even in 2007 will put this in perspective. That being said, I'm happy for anyone to reach out to me if they are considering RfA and I'll give me advice and see if we can work out a nom. TonyBallioni (talk) 17:43, 19 April 2019 (UTC)
That's probably a better description of why the community likes the status quo than the straw men I constructed; you're saying that the results of the current system are good for Wikipedia's current status as a top-five website. While I don't necessarily agree that a system designed by some random dudes in 2004(?) is ideal for that, I think that is a good explanation for why we aren't going to get any meaningful change. And maybe that's a good thing. -- Ajraddatz (talk) 17:52, 19 April 2019 (UTC)
Yeah, I'm not saying it is an ideal system, but there is no such thing as an ideal system. I'm saying that the status quo works reasonably well at preventing 14 year olds with a clean block log and 3 months of Huggle experience from hard blocking the U.S. House of Representatives without realizing the implications, etc. I'm sure every one of the groups you listed above could find their own version of this action would not be ideal on a top-5 website. The system has flaws, but it does generally well at keeping disastrous choices out, which at this stage of en.wiki's development is more important than getting someone who might be good in. TonyBallioni (talk) 18:23, 19 April 2019 (UTC)
I've typed up a number of responses, but none of them really hit the mark. I obviously fundamentally disagree with your assessment. There is no reason why adminship should be more important than other toolsets from a risk management perspective. There is no reason why people who were promoted in 2007 should be more responsible than users who joined more recently. Preventing good people from becoming admins stagnates our leadership and enhances some of the worst parts of the Wikipedia culture that exists today. But the reality is that you speak for the majority of the community − better to let 50 good candidates be dissuaded from the process than promote one bad admin. And that said, I'll go find more meaningful discussions to have :P -- Ajraddatz (talk) 18:56, 19 April 2019 (UTC)
Re "no reason why people who were promoted in 2007 should be more responsible than users who joined more recently": If anything, we have evidence to the contrary, because early admins who went through less of a vetting have more often been desysopped. There's a possible statistical skew in there, in that the more years you are active the more opportunity you have to pooh the screwch in a way that causes people to seek your head on a pike (though that idea would seem to deny that admins could get better through practice and experience, so I don't buy it). Regardless, it's unquestionable that the criteria (both formal and "my personal RfA criteria are ...") were generally much more lax back then.  — SMcCandlish ¢ 😼  20:30, 6 May 2019 (UTC)
  • (edit conflict) Since the English Wikipedia's edit rate has been fairly stable, I think it would be better to ask whether the situation is sustainable, rather than whether it's possible to increase the number of RfAs. Is the number of high-activity administrators stable? We know that the total number of administrators isn't necessarily relevant here, because the bulk of the work is handled by about a quarter of the current administrators.
    I don't know how to use Quarry, but presumably you could generate a graph of the rate of admin actions over the past five years. If it's realistically not going to cause the project to descend into chaos, then it's not really necessary to spend all this time debating it when it's not really going to spur anyone into magically wanting to dedicate hours of their time becoming one of Wikipedia's janitors. Jc86035 (talk) 17:49, 19 April 2019 (UTC)
presumably you could generate a graph of the rate of admin actions over the past five years As it turns out, I did such a thing! Check out User:Amorymeltzer/s-index; you can see some graphs (last updated in January) there of total actions, both including adminbots and not. A ton of caveats — not all sysop actions are equal and they aren't a great way to measure usefulness or difficulty — but I think both the totals and the s-index itself are interesting to have. ~ Amory (utc) 18:18, 19 April 2019 (UTC)
I’m not sure I quite understand those graphs. Do they tell us whether it is the same subset of admins who are consistently active over time, or whether there are a significant group of admins who are sometimes active and sometimes not? ϢereSpielChequers 19:17, 19 April 2019 (UTC)
Nope! It's the project/community as a whole, so it doesn't care who is doing what when. We can graph an individual's number of monthly actions over time, but I don't think that really means much? Or did I misunderstand the question? ~ Amory (utc) 23:58, 19 April 2019 (UTC)
@Amorymeltzer: - you get bonus points for not merely having created what I think this discussion needed, but adding it in so that the next paragraph I read actually contained it. Nosebagbear (talk) 20:21, 25 April 2019 (UTC) Now I feel this discussion needs some cheese paninis...
  • It's always been an issue. As Kudpung correctly states, we have an interesting conundrum, where there is a general consensus that something should change, but very little on what changes should be made or how to implement them. One of the primary problems I see is that there isn't any clarity on what makes a vote (and let's not kid ourselves by putting the bang before it, there are even codified numerical thresholds) valid versus invalid. At AfD, there's at least some general guidance, in terms of for example arguments to avoid, and those are more or less permitted to be "enforced" by the closing admin. If a bunch of brand-new users flood an AfD discussion, all making very poor arguments, that can be considered and weighted (or de-weighted) accordingly. On the other hand, there's little if any guidance on what makes for a persuasive RfA argument, and many are even just a bare "support" or "oppose".

    I think part of the problem is the idea of a "big deal" after a few administrators either had their account compromised or went off the deep end. If anything, I think that those incidents actually bolstered the "not a big deal" argument, since the disruption they caused was undone easily and quickly. Something like the Wifione and Runcorn incidents, on the other hand, certainly do show that an administrator determined to do harm can do actual and insidious damage over a period of time—but, then, the RfA process didn't exactly stop them, did it?

    So, I think the questions about whether RfA is fit for purpose are well in order. I don't know an easy answer; if I did, I certainly wouldn't have kept it a secret. Perhaps the best way forward, but by no means easily done, is to figure out what we, as a community, actually want to see in an administrator, but also realizing that just checking whether someone "ticks all the boxes" isn't a great way forward. I think, also, that there's some focus on things only tangentially related to suitability as an admin, such as content work. While I wouldn't want an admin with absolutely no content experience, there are plenty of people who don't write featured articles but could make great admins, and more than one who has written a ton of featured articles but who I wouldn't want within a million miles of the tools.

    Maybe the place to start is to ask that people put forth at least some rationale for why they support or oppose the person becoming an administrator, and giving bureaucrats greater leeway to discount arguments which are either completely unsupported or nonsensical. That would probably require some substantial notification and instruction, at the very least, as right now people are especially used to putting a "support" with no rationale. If you think someone should be an administrator, one would hope that you have at least a few words to say as to why you believe they should. I think that might also help the pass rate. As things stand currently, if you want to find actual reasons, you're much more likely to find them in the oppose section than in the supports. More indication as to why the supporters believe the candidate should be an admin could allow someone undecided to see both sides of the issue, not just the reasoning behind those against.

    That also brings up the other issue. Raising the pass rate doesn't help if good candidates won't run to begin with. When asked, potential candidates quite often note the ugliness of the process as a reason for their decision not to undertake it, and even many who ultimately passed noted that as a reason that they were reluctant to go forward with it. If RfA is keeping away good candidates who would be able to pass, that certainly indicates a problem. I think, there, that the solution already lies in our existing policies. We are expected to treat one another civilly. That certainly does not mean it's not okay to disagree with someone. It means that if you do, you focus on what you disagree with and why, rather than attacking them. Those standards of behavior should be enforced at RfA, and participants there (both candidates and voters) should expect that they will be. It is okay to say you're concerned about a candidate's history and why, and for that reason oppose their request. It should not be acceptable to say "Oppose, this moron completely sucks in every way." The same should apply to any discussion regarding a particular vote, or discussion of the candidate in general. I think an expectation of baseline civility at RfA would help in convincing good potential candidates to be willing to undergo the process. It's stressful enough without namecalling and nastiness; there's no need to actively make it even worse. Seraphimblade Talk to me 19:40, 19 April 2019 (UTC)

The problem is not so much "Oppose - candidate is an asshole", which happens rarely. More often, the problem is more like "Oppose - candidate made a typo [1]", "Oppose - only 1 GA", "Oppose - I have my reasons and I do not appreciate badgering" or "Oppose" (no reason) which is then jumped on with excessive badgering like "What the **** are you doing, haven't you got a ****ing clue how to behave here, people like you should be topic banned from RfA". I've done this, as has Kudpung, so I speak from some experience when I say this is neither big nor clever. Ritchie333 (talk) (cont) 10:47, 23 April 2019 (UTC)
  • There's an obvious missing cause (maybe several missing causes, but one's clear and I understand it well enough to outline it) missing from Kudpung's otherwise spot-on list of them at the top of this subthread under "Something dramatic must have happened in 2007/2008 ..."). Around this time is when meta:Eventualism stopped being viable. By this point, most of the "sexy" Wikipedia articles had already been written, most WP:P&G was firmly set, even most influential WP essays already written in much like their present form; ArbCom was solidly in its bureaucratic pattern, and Jimbo's "benevolent dictator" activity was being sloughed off (while along with it went the "anyone sane and clueful can be an admin, which is no big deal" idea – definitely already a thing of the past in practice despite mantra-like lipservice to that folklore); vandalism was already well-in-hand, and most mass-scale attempts to do stupid stuff on Wikipedia were over. In short, WP had aged out of the first phase of the organizational life-cycle, and the heady buzz of the wild and wooly era of "visionary founders" and "we don't really need rules" had worn off. This is when the general editorial decline began in earnest (starting as a trickle in 2006[8]). It's also, pretty much necessarily, when the RfA decline started in earnest, candidates and RfA participants being a statistical subset of the editorship. WPs editorial ranks had been massively swelled in the early 2000s bY a large influx of Slash/dot users, fascinated and very enthused by the idea of whether the crazy experiment to collectively create a free, volunteer, self-organizing encyclopedia could actual work. That enthusiasm and fascination, in a large block of the editorial pool, faded quite quickly. By the period at which WP actually had become one of the general public's most frequently used sources of information, very rule-bound, and with much less to do at it that was new rather than polishing and maintenance, it had become an institution not a "what if", so for many of those early adopters it became boring old news, or (for some having more of a "my job is done" than "yawn" reaction) like watching your kid finally ride that bike away from you, training wheels still attached but without you holding the offspring and bike upright. PS: The checklist in Guy Macon's tongue-in-cheek post after Kudpung's, for how to pass RfA, is missing a denouement point: after you get the bit, avoid making any potentially controversial decisions that actually require judgement (you know, the reason the community needs admins at all).  — SMcCandlish ¢ 😼  20:30, 6 May 2019 (UTC)

Arbitrary and capricious breakEdit

  • Re "...what we, as a community, actually want to see in an administrator", Maybe "we" want the wrong things. We have a number of highly vocal editors who create articles and think that this somehow makes them better than Wikignomes who make smaller improvements on a larger number of articles. Some people are simply not very good at composing paragraphs but are great at interpreting policies and dealing with people, but the content-creation-bigots would have you believe that such a person is somehow not qualified to be an administrator. :( --Guy Macon (talk) 20:00, 19 April 2019 (UTC)
It always seems to me that a pretty high proportion of "opposers" have created very little content themselves, though of course there are exceptions. Complaints about edit-count, an area where Wikignomes do very well, are common, but these, and excessively high requirements for actual content, are much fewer that they were a few years ago, imo. Johnbod (talk) 20:09, 19 April 2019 (UTC)
Re "Oppose, this moron completely sucks in every way." We have admins, who are suppose to redact cmts/block people who write that. Are people not writing that or are admins not doing their job? Alanscottwalker (talk) 20:15, 19 April 2019 (UTC)
"Oppose, this moron completely sucks in every way" is indeed blockable; "Oppose, the user has not demonstrated good judgement in any of the areas" is not.--Ymblanter (talk) 20:23, 19 April 2019 (UTC)
I sometimes think it would be a good idea to come up with a kind of a rubric - that an oppose or a support vote should be supplemented by a set of "partial" votes. I saw that users have RfA criteria, but, as far as I am concerned, they are often too rigid and too detailed. For example, often users say that they require GA and FA experience. This is not really necessary - for example, on my RfA I was very open that I am never going to have GAs and FAs, and I did not get a singe oppose on this ground. The real reason why people ask for this is that they want to see some content creation experience - and it is needed, in its turn, to demonstrate, that the candidate (i) knows what copyvio, vandalism etc means; (ii) has certain respect to people who actually create content and to the content creation process. This can be reflected in a rubric as "content experience" and be one of say ten criteria. If we would be able to come up with such a rubric and run it say during a year at all RFAs (without any effect on the outcome, the rubric must merely explain the vote), we would understand much better what voters actually expect.--Ymblanter (talk) 20:20, 19 April 2019 (UTC)
No, sorry, users do not often say "they require GA and FA experience"! Some used to, but there was a bit of a campaign slapping this down a while back, & it's much less common (part of my point above) - especially FA. These days 2-3 DYKs will satisfy the great majority of voters. Johnbod (talk) 20:32, 19 April 2019 (UTC)
This is ok, I just used it as an example to illustrate my concept. One can city DYK instead.--Ymblanter (talk) 20:35, 19 April 2019 (UTC)
?? To complete the point, take for example, from a nom this January "Enterprisey has written a handful of articles, often related to tech, of which four were promoted at DYK..." - 252 Support, 2 Oppose, neither mentioning content creation. Johnbod (talk) 20:38, 19 April 2019 (UTC)
Opposes for a lack of (interest in) content creation are nothing new and in fact something I myself experienced in my RfA back in 2008, i.e. the "glory days of RFA". So I don't think these kinds of opposes were ever responsible for the decline in people running for adminship. As for Ymblanter's point, I wrote the essay Wikipedia:Content awareness, not content creation back in 2011 and I think it's advice still applies.</shameless selfpromotion> Regards SoWhy 18:42, 21 April 2019 (UTC)
All I ask is that there is a little bit of Wikipedia that they look at and are proud for their efforts.--Wehwalt (talk) 19:10, 21 April 2019 (UTC)
I remain unsure of why I had such an easy time of RfA, given that I had very little major content creation to my name (though I went to the opposite extreme after my RfA, I have yet to get any DYK/GA/FA credit at all) and was quite open about the fact that I Absolutely Just Don't Care about such things. As far as I know I haven't laid waste to anything, so why would anyone else standing for RfA in my situation do so? I was basically the last admin chosen based on NPP work (TonyBallioni arguably excepted), and the people deleting things and blocking username violations are still overwhelmingly all the same people who were deleting things when I tagged them (plus me!). Plus, I've come to see that getting enough experience to become an administrator drained a lot of parts of my life, which I frankly regret but can't take back now; only relying on people willing to take part in Wikipedia as a de facto second (or even first!) job, with enormous expense to their personal lives, is not a way to keep people interested in what tasks admins are needed for (since I am one now I happen to enjoy those tasks, but I'd have never sacrificed as much as I did to become one). The Blade of the Northern Lights (話して下さい) 03:11, 23 April 2019 (UTC)
I somewhat wrote the rule book on "why admins should create content" WP:WRITE, but I have never mandated GAs (eg: Wikipedia:Requests for adminship/331dot), and would indeed support a user who poo-pooed the process with well thought out and justifiable reasons (even if I didn't personally agree with them). Ritchie333 (talk) (cont) 10:53, 23 April 2019 (UTC)
Based on my experience, Blade, I never would encourage anyone to put themselves through an RfA. It's not worth the pain from the personal attacks. I guess I'm glad I passed but the RfA changed me, changed how and what I edit and how I view everyone else who edits here. Everything is different now and not in a great way. Working on Wikipedia is more duty than pleasure now. And RfAs aren't any less brutal now. Kudpung is right that it's the voters, not the candidates that are the problem here. There is no call to dutifully go through every area where a candidate is lacking and enumerate their faults. This is not Festivus, there doesn't need to be an airing of grievances but everyone that can find an edit where the candidate made a mistake will find a way to bring it up years later. Support, Oppose, I don't care, but one can oppose someone's attempt to become an admin without trashing the person. Liz Read! Talk! 00:59, 25 April 2019 (UTC)
I remember that all too well. Since then I haven't made a nomination because it was hard to even watch, I felt absolutely horrible that it got so vicious; I can only imagine what it was like for you. And it's not even for anything significant, despite the pompous claims of adminship being Very Serious. It's a few buttons, everything (save botched histmerges!) is easily reversed, and the number of ZOMG VANDALS CONSPIRING TO DESTROY US!!!!! ascribes an utterly unwarranted self-importance (as if they're lined up everywhere trying to get admin tools so they can blow up Wikipedia). It's a website. Seriously. The Blade of the Northern Lights (話して下さい) 01:54, 25 April 2019 (UTC)

The issue raised by Alanscottwalker about whether admins are doing their job, is probably due to admins preferring to vote rather than clerk the process. They can't do both. Kudpung กุดผึ้ง (talk) 23:19, 23 April 2019 (UTC)

@Kudpung: - this is a similar issue to the fact that 'Crat chats have been recusing major numbers, because the more controversial (and thus likely to warrant a 'chat) an RfA, the more Crats end up participating in it. Nosebagbear (talk) 20:21, 25 April 2019 (UTC)


  • Do we have an ability to see what the median (not mean) number of edits is by an admin in general, and the median & mean number of edits by an admin created in, say, the last 2 years? Nosebagbear (talk) 20:21, 25 April 2019 (UTC)
    • Any stats you can draw up on it are a bit problematic as the data set is rather small (see WP:RBM). For what it's worth, the mean edit count of a successful candidate at the time of their RfA for 2017 was 43k edits. The median was 33k edits. You didn't ask for this, but on the fail side...isolating for those that failed without SNOWing or NOTNOWing; 40k edits mean, and the median was 13k. I don't have full data for 2018. --Hammersoft (talk) 20:36, 25 April 2019 (UTC)
      • @Hammersoft: - thanks for working that out. That's an impressive mean/median difference for the unsuccessfuls, must have been a seriously high candidate. You're right, of course, as to small sample size. Nosebagbear (talk) 21:42, 29 April 2019 (UTC)
Meditations at the graves of dead nuclear plant managers
Actualy, "sample size" has nothing to do with it. This isn't a sample of RFAs. It just is the RFAs. Period. It's the whole population. Sampling don't come into it. EEng 23:40, 29 April 2019 (UTC)
  • It has everything to do with it. Small data sets, even if they are complete, do not lend themselves to determining trends. You need larger sets to do that. consider an extreme case; let's say we had only one passing RfA in the time period. That one RfA could show a result in the six sigma range and we would never know it. Simplifying; flip a coin once and get tails, and the "sample set" would conclude that heads is impossible. Nosebagbear can amplify; but I suspect this is why he was asking for the median, and not the mean. Note how the median and mean in both sets are off by a fair bit. That's to be expected with small data sets. --Hammersoft (talk) 02:54, 30 April 2019 (UTC)
I have a feeling that EEng (for those unfamiliar with his, um, sense of humor) was capriciously yanking everyone's chain. He just means that it's the entire data set as opposed to a sample of data taken from within the whole. (See Sampling (statistics).) I don't think anyone disputes that more RfAs would yield sturdier conclusions. --Tryptofish (talk) 23:42, 30 April 2019 (UTC)
(edit conflict) Capricious isn't quite the right word, but here's the response I composed a few days ago and just now found, unsaved, in a forgotten browser tab...
Well, since I have a degree in statistics I don't really need any explanations, simplified or not. The concept of sampling can make sense here only if the RfAs we actually see are a subset of some larger set, this larger set being a "population of all RfAs" existing somewhere beyond our view. But that's not so: the RfAs we have are all the RfAs there are, period. In particular, when you say that a lone RfA might "show a result in the six sigma range" ("six sigmas" being a management buzzword with no practical application in statistics, but we'll let that pass), the sigma could only mean the SD of this hidden full population, which again doesn't exist. There is no sample.
You speak of "determining trends", but trends in what? You may be trying to infer a shift in some parameter or distribution which somehow represents the community's willingness to give adminship to editors with various edit counts, and that might make sense if we could see RfAs as some kind of random process like a coin flip. But though there are random mechanisms at work in RfAs (e.g. the random events of which particular editors happen to see the RfA notice, decide to participate, etc.) I have no idea how to model the RfA process as a random process in any statistically meaningful sense.
Or you might use # of edits as an input to a regression exercise, after which you might try to make statements such as "The edit count of editors earning adminship has been declining at about X thousand edits per year" but again, without a coherent model of RfAs as a random process that's just feeding a lot of data into some impressive mathematical machinery, letting that machinery clink and clank a while, and then exhibiting betas with no real idea of what they mean.
Loose misuse of statistical concepts such as sampling is what leads people to do silly things such as apply hypothesis tests to full populations, and as I said somewhere else once, that kind of thing has caused a lot of problems, as the ex-managers of the Chernobyl and Fukushima nuclear plants would be able to tell you first-hand (if they weren't both dead, of course). So please do your part to stop the senseless slaughter of nuclear-plant managers, and don't talk about sampling when there's no sampling involved. (I realize that's a bit over the top, but I'm in a puckish mood.) EEng 14:30, 3 May 2019 (UTC)
You're lucky I didn't say "arbitrary" instead of "capricious". (wink) --Tryptofish (talk) 18:30, 3 May 2019 (UTC)
  • oooooommmmm <meditating> oooooommmmm :) --Hammersoft (talk) 15:42, 3 May 2019 (UTC)

RfA voting motivationsEdit

I think Seraphimblade is absolutely a spot on, and a discussion why and how we vote on RfA should happen. Hopefully it could lead to some understanding of the community what is actually going on.--Ymblanter (talk) 19:59, 19 April 2019 (UTC)

  • I personally rarely oppose (I believe I opposed twice, and in both times the candidates finished well below the line), and I often support if after research (admittedly, not always extremely thorough one) I do not see any crucial problems. The question is of course what are crucial problems. I do not expect the candidate to know all the details of the policies (which can be read at any moment), but I do expect them to have a clue - meaning they should know how to behave in various, sometimes unexpected situations. None of us is perfect, but at the very least one expects that if an admin really screws up and gets pointed out to that, they admit they screwed up and do not repeat.There are some things which would be absolutely unacceptable for me - such as past documented experience of vandalism, socking, or, I do not know, battleground mentality without proper subsequent reflection - but, to be honest, this is a non-issue, we do not get such candidates anyway. I can be lenient on experience, I can be lenient on past conflicts, again, if conclusions have been made, but if a candidate is clueless and not capable of reflection, I am not going to support them. See also my rubric suggestion above.--Ymblanter (talk) 20:21, 19 April 2019 (UTC)
I also think Seraphimblade's comments are quite accurate, but everyone else who has commented up to now has made interesting observations and suggestions. I particularly find Seraphim's analogy with AfD a pertinent idea - it might be useful to revise WP:AAAD ad talk about it. My WP:RFAV is aimed generally at voters who are very new and/or unkind, and I often drop a link to it to voters who IMO have acted in an inappropriate manner; if anyone reads it, I belive it does its job quite well, but I'm not sure if either of the essays get read very much. One of the debatable effects of Biblioworm's reforms was the large increase in exposure of RfA - does doubling the umber of participants double the effectiveness of the process, or does it simply attract more drive-by and potential rubbish votes and user questions? Kudpung กุดผึ้ง (talk) 01:34, 20 April 2019 (UTC)
As long as it isn't required to give any rationale for a support, we'll never know the answer to that. If all someone needs to type is "#Support ~~~~" (or even just "#~~~~") how are we supposed to know if it's drive-by and/or potential rubbish? ansh666 01:58, 26 April 2019 (UTC)
I mean I'd be perfectly game for support voters having to justify, however as the pass-rate is calculated off the current status quo, we'd have to rejig it before implementing Perhaps a push for support !voters to justify would be beneficial in multiple ways, without the drawbacks? Nosebagbear (talk) 21:44, 29 April 2019 (UTC)
I don't see how. Justification for support can be pretty much finding nothing to make you distrust the person. per Nom does that. Do we really want the RfA full of 'I didn't find anything bad' comments? That is basically 'per nom' so not achieving anything we don't already have. If someone comments that they support because of 'criteria a' are they then going to draw lots of oppose responses saying yeah, but what about b, did you find some of that, what about c candidate never worked in c....do supporters then need to list everything?. Oppose is different; you only really need to agree one significant thing makes them unsuitable....what I'm rally trying to say here is it is hard to prove a negative. I for one don't plan on writing a big diatribe for support along the lines of 'good AfD work, good template work, good edit count, made dyk, made fa, worked with bots, good hit rate at npp, polite talk page responses, helpful to new editors, etcetera ' ClubOranjeT 08:20, 7 May 2019 (UTC)

How this interplays with personality disordersEdit

The current method favors the narcissist-leaning user, while a more mechanized voting method favors the psychopathic-leaning user. The current discussion method gives an advantage to the higher degree of networking which narcissist-leaning users are capable of. Psychopathic-leaning users, being more lone wolfs, are at a disadvantage. The existing rules concerning the limitations and conduct of admins are currently adapted to check the vices that narcissist-leaning users are more prone to succumbing. If you change to a voting system, some tweaking may be needed. As for my own preference, I suggest we continue to let the problem get worse. If it gets bad enough, people may start treating others better on the RfAs on their own. If the admin shortage gets very bad, in coming years there may be an algorithm based system which does a better job choosing admins. Epiphyllumlover (talk) 21:48, 7 May 2019 (UTC)

I understand your argument, but which system would favour down-to-earth, reasonable, perceptive, and caring users? ---Sluzzelin talk 21:55, 7 May 2019 (UTC)
Jimbo Wales' country school (quick answer, I need to think more).--Epiphyllumlover (talk) 23:05, 7 May 2019 (UTC)
Some sort of real-world program where prospective admins were vetted from a pool which had taught Wikipedia classes based at local schools, libraries, places of worship, neighborhood centers, half-way houses, etc. Classes could be for adults or children, but priority would come for working with youth. People would need to be evaluated by the teachers/librarians/clergy/other administrators for their people skills. A written form completed by these people would be signed, scanned, and uploaded to Wikipedia. A real world example which approaches aspects of this method would be the GLOBE program.--Epiphyllumlover (talk) 23:14, 7 May 2019 (UTC)
A drawback to having down-to-earth admins is that we need near-personality disordered admins with the grit to fight hard against near-personality disordered editors. Possibly the feature with enhanced sanctions for certain topics could be strengthened--it could become sort of a mini-Wikipedia inside of a Wikipedia, with more rules and easier adminship. Admins would need to choose whether they want to be a good cop on the larger Wikipedia or a bad cop on the controversial pages. The tough ones could decide "Better to reign in Hell than serve in Heaven." Meanwhile, the rest of Wikipedia would become more pleasant due to safety valve theory.--Epiphyllumlover (talk) 23:37, 7 May 2019 (UTC)
We want admins to be balanced not imbalanced — I'm not sure you're depicting reality or anything realistically-implementable with the rest. El_C 00:09, 8 May 2019 (UTC)
Yes, that was for fun! The serious comment comes from the first one: "If you change to a voting system, some tweaking may be needed" to account for the difference in personalities you get from each system. In particular, Wikipedia culture tends towards cerebral narcissism as it is (See Sam Vaknin#Views on narcissism). The relation of this, broadly speaking, to the RfA process has already been discussed by Stvilia et al. (Wikipedia administrators#Requests for adminship, and see the scientific studies section at the bottom of the page if you are curious). If you change the system to disfavor near-cerebral narcissists in favor of near-psychopaths (see Psychopathy#Society and culture), expect to adjust some rules to account for this.--Epiphyllumlover (talk) 00:22, 8 May 2019 (UTC)
Whatever all that is, down-to-earth it is not. Which is to say: hello, my name is Mr. Snrub! El_C 00:58, 8 May 2019 (UTC)

Are we asking the correct question?Edit

That there are very few RFAs, and very few new admins, is a fact. That the reason for this is problems with the RFA process isn't - before a user reaches the point of doing an RFA, (s)he genereally goes through the following steps:

  1. Doing a first edit on Wikipedia, and feeling that this edit was appreciated.
  2. Becoming a member of the Wikipedia community
  3. Handling areas of enforcement of policies, including requesting help from admins for this purpose.

I believe that we have serious problems with step 1, which result in fewer potential admins reaching step 3. עוד מישהו Od Mishehu 13:23, 15 May 2019 (UTC)

What you say is absolutely true. Sadly the fact that no one has replied to your post shows that, with over 250 archives of this page, people are much more interesting in discussing the minutae than getting to the nub of the issue, which is that we don't have enough editors. If you want lots of top-class sportsmen or pianists or whatever, you need a large base at the grass-roots. That's what we're missing. Nigej (talk) 22:05, 18 May 2019 (UTC)
1,000% agree with this. Levivich 19:53, 20 May 2019 (UTC)
Od Mishehu's point is a good one, but its not only about their first edit. Everyone needs to feel like their work is appreciated in order to keep working. This thread reminded me of a TED talk by the genius Dan Ariely called What makes us feel good about our work? and how spot-on it relates to every editor on Wikipedia. To sum it up, he finds that its not rewards that motivates people to work - its the sense of progress and the satisfaction that others appreciate the fruits of their labor. I think that applies to both a novice editor, making their first simple tweaks to a page, to an experienced editor, who perhaps puts some extended effort into something. We have the "thank" function conspicuously right next to "undo" on the page history, but which do we use more for new editors and old. Are we thanking editors even for changes we disagree about? -- Netoholic @ 21:08, 20 May 2019 (UTC)
Before one can be appreciated for doing an edit, (s)he must actually do it. Sincd I joined Wikipedia, I believe this became much harder. Today, for a user to do this first edit:
  1. The page must not have any protection. The page, however, has a high chance of being a BLP, a controversial issue or a current event - all of which have a high risk of being protected.
  2. The IP address must not have any block. With soft blocks being available, we are much quicker to block IP addresses and ranges than we were when I first joined. Additionally, the CheckUser tool gives certain community members access to reasons to block IP ranges which no one had back then.
  3. The page must not have a BITEy edit notice. Every BLP has one now. Any edit notice which appears to say "before adding any information, please read a difficult policy page" (such as WP:BLP) is inherently BITEy.
  4. The edit must not trigger any edit filter warning/disallowing or capcha. Yes, even a warning or a captcha, which only make the edit harder and not impossible, make it less likely the user will actually save the edit.
I believe that this process is much harder than it was when I joined (semi protections, soft blocks, CheckUser, edit notices, edit filters, captchas for certain types of edits), which means we have fewer new users and fewer potential candidates for RFA. עוד מישהו Od Mishehu 04:45, 21 May 2019 (UTC)
I suspect the biggest reason for the apparent decline in raw edit numbers in the 2007-2014 era and its subsequent stability at around 5 million edits a month were that edit filters lose us a lot of vandalism, the move of intrawiki links to wikidata lost us a lot of edits, especially bot ones, and the smartphone is not a viable editing device for most smartphone users. Yet the community is not much smaller than it was at peak. If our problem was recruitment of new editors not RFA then we would have a community where most people were admins and we would be bemoaning the lack of new non-admins within our community. Instead we have a broadly stable community with a declining number of admins and a growing wikigeneration divide between those who have been here over a decade and are mostly admins and those who have been here less than a decade and are rarely admins. ϢereSpielChequers 09:31, 10 June 2019 (UTC)

Just change it alreadyEdit

Nice idea @MelanieN: but most of those you would approach aren't masochists. The main thing that makes RfA hard for the candidate is that it is a lifetime appointment and too much of the community gets to nash their teeth at them for a whole lot of (often frivolous or just downright nasty) reasons on the back of that premise. Mother Theresa would fail RfA if a couple of users dug up evidence she once kicked a rabid dog. @RexxS: was one of the best experienced candidates to run in recent times and look how that went. However, there are many other editors out there that have run out of ideas or enthusiasm for articles they want to write, or get bored with the monotonous vandalism reversion or NPP, or have reviewed all the current AfDs and are looking for things to do, but won't go through hell week and who could blame them. A way better idea would be to bypass the RfA process in its current form. Take the heat out of it by returning to "Admin is no big deal", and taking away the idea that the community is stuck with bad ones forever

  • Existing Admins should identify, approach and get agreement from potential candidates for said areas, and commit to providing some guidance.
  • Candidate should then be vetted by 'crats and given Adminship for 12 months.
  • Then they could have a 'ratification RfA', where they are generally likely to pass as they have shown they "ain't breaking it". I would expect all admins to pass this ratification as long as there are not generally popping up as the perpetrator in ANI cases or losing Arbcom cases against them.
  • I'd also like to see ratification RfAs for all admins every , say 3 years...same thing, unless ANI or ARBCOM cases are showing an unsettling trend, I'd expect opposers to be shut down by the community.

Never happen though, too many people like the drama.ClubOranjeT 11:18, 20 May 2019 (UTC)

  • The need to undergo a periodical ratification RFA would probably cause thosr admins who deal with our most controversial areas to fail, even if they had handled the situations correctly; these are the admins we need to retain more than anyone else. עוד מישהו Od Mishehu 05:18, 21 May 2019 (UTC)
    That's my same issue with it. If you handle a tough issue, you will handle it wrong. If you had handled it differently, you would have also handled it wrong. That would lead to admins only handling uncontroversial situations, since wading into tough ones guarantees some opposes next time around. No matter what you do in those cases, someone's going to be mad about it. Seraphimblade Talk to me 05:47, 21 May 2019 (UTC)
    Well I disagree with both of you. Define the criteria for opposing. If an admin is taken multiple times to ARBCOM and the community decides there they acted appropriately, then no issue. If the community finds they have gone outside the bounds of acceptable behaviour then perhaps they shouldn't have the tools anyway. Same at ANI. Time for people to be a bit mature. If a user complains about being blocked it should take two minutes to see there was some justification and that user can be told to pull their head. Too much pussyfooting around. And any admin who thinks they might lose the tools at a ratification RfA is exactly they type of admin that probably should because they are likely abusing the privilege. Admins ask the community to trust them, but they don't trust the community. Bit of a one-way street. I know a number of Admins who ran under a flag of admin recall. Don't remember voting them out yet. It's not about a couple of people having a grudge with an admin. Most of the community will see right through that and back the admin. The current system is not not working and hasn't been for some time. But like I said, know one wants to know. They want to protect their own little world and complain endlessly that not enough people want to put themselves on the pedestal to get stoned for public amusement, instead of getting on with the job of building a quality encyclopaedia. ClubOranjeT 06:49, 21 May 2019 (UTC)
    An admin reviews a discussion requiring admin closure. Significant claims were made in both directions, and the admin chose one. Months later, the admin is up for reconfirmation. Quite likelt, the "losers" in that discussion would be more likely to show up than the "winners".
    This admin has actually closed many such discussions. A user "won" in one discussion, but "lost" in an other one. This user would be more likely to remember the case where (s)he "lost" and oppose the admin because of that. עוד מישהו Od Mishehu 12:55, 21 May 2019 (UTC)
    Od Mishehu, I personally wouldn't know whether you've closed any discussion I've been involved in as I don't normally even look. I do know of a couple of admins who have closed contrary to what I might agree with, and I know a couple I find slightly rude or dicks, but that doesn't mean I'd vote against them; they'd need to do something legitimately wrong. You know, abuse the tools. But I'm not a long tenured admin, so I still think AGF is a valid guideline. ClubOranjeT 07:15, 22 May 2019 (UTC)
I agree. I think what's generally forgotten is that it will almost certainly be the same body of editors turning out to !vote for each one; I am wholly unconvinced that someone who can gain adminship without annoying people is then going to start annoying 50%+1 of them; or, likewise, that an admin who works in difficult areas will be unapprecaietd by the vast number of commenattors. I would expect the number of editors making WP:IDLI (read=dumbass) !votes at the reconfirmation to be the same proportion of editors who do so now at RfAs, viz, a tiny minority except in the cases where the candidate is problematic. The only adins who would have anything tofear, rightly, would be the legacies; not because they would have to defend maing bad calls, but because they would have to defend making no calls at all (cf this episode), and by it's nature that's not going to har the project. Frankly, the community—at least those who make up the backbone at RfA—are shit hot and spotting trolling !votes, and I see no reason why RRfA would be different; except the Discussion moved to talk page template would continue to be kept busy :D ——SerialNumber54129 13:15, 21 May 2019 (UTC)
@Serial Number 54129: I don't often disagree with you, but I think your description of what such a recall process would look like is missing something about the nature of Wikipedia editors as a body; that people are willing to go to far greater lengths to spite their "enemies" than to defend colleagues they have a good relationship with. There's a fair number of people who have expressed appreciation for the actions I have taken as an admin, but they are not as likely to !vote at a reconfirmation or recall as those editors who have long-standing grudges. I can't even blame them; those folks who have supported an admin in the past are far more likely to be spending their time constructively rather than monitoring their contributions for opportunities where they can !vote against said admin, whereas those who bear a grudge all crawl out of the woodwork whenever anything resembling an election occurs. You commented at WP:CUOS2018; you've seen this happen. Vanamonde (Talk) 17:42, 21 May 2019 (UTC)
@Serial Number 54129: You're overlooking the fact that it's nowhere near good enough to have a majority of editors in support; the threshold is not 51%, it's 76%. The minority of opposers get super-votes at RfA. Hawkeye7 (discuss) 20:18, 21 May 2019 (UTC)
Hawkeye7, you're overlooking the fact the threshold is not 76%. If you are changing the system to get with the times, then get with the times. You get to set the threshold for a new system. Personally I'd be happy with 75%....Remember you are voting to take tools away, not give them. If 75% think they should be stepping down then that shows a problem. If you. Think that's too low, suggest a different number. ClubOranjeT 07:00, 22 May 2019 (UTC)
A wise individual once said: "Friends may come and go, but enemies accumulate." The stress of dealing with difficult areas is great enough without adding the concern that it may reduce the odds of reconfirmation. UninvitedCompany 20:52, 21 May 2019 (UTC)
  • And another wise individual once said "Be careful what you ask for. You just might get it." I'm going from fuzzy memory, but as I recall when the German Wikipedia instituted reconfirmation RfAs, they lost 40% of their admin corps from admins who didn't even want to stand for a reconfirmation. With respect, I don't see how we solve what ails RfA by making admins run through the gauntlet repeatedly. Further, if we did lose 40% of our active admin corps, the result could be catastrophic. --Hammersoft (talk) 22:31, 22 May 2019 (UTC)
  • Maybe I don't explain things very well, because some of you just don't seem to get it. It doesn't have to be "RUNNING THE GAUNTLET". It's only seen like that because we let it be what it is. Most of the issues come from frivolous opposes. There will no longer be frivolous opposes; *oppose, he blocked me for 3RR, R= "so, not abuse of tools, you deserved it, vote zero rated", *oppose, I didn't like how he closed that AfD, R="So, arguments were for and against, there's DrV if you don't agree, but he didn't abuse admin tools so vote zero rated", *oppose, he blocked me in a content dispute he was involved in because I held different view, R="you and 30% of other voters have raised valid issues, let's evaluate the arguments." Won't take too many of these before the culture changes.
Its also not about existing admins; write in a protectionism clause grandfathering right to perpetual power for pre-existing admins if you wish. It's about getting other good editors who want to help become admins do they can help, without them having to be torn up and spat out. I just don't see how that is seen as a good thing by so many in an environment where civility and such like is a core principal.
Open your minds a little. New methods can have new rules it doesn't have to be locked into the old ways. It doesn't have to be 70% pass rate. It doesn't have to be any oppose counts. It doesn't have to be lots of questions for a candidate to answer - in fact I wouldn't see any need for questions, because they'll have already answered by their admin actions during their trial periods. It can be as simple as #crats giving the tools on recommendation of (2?) admins, #(2?) admins coach the trialist for 2? 4? 6? months, #crats pull the plug if coaches report issues, or present to community if they don't. Most by then would fly through as there will be no grounds for oppose if they are using the tools right. ClubOranjeT 13:07, 23 May 2019 (UTC)
No, it literally is not. An RfA-like process is a requirement for deletion and viewdeleted rights (the tools are tied together) per the Wikimedia Foundation, and I would not trust a "trainee" with blocking rights. —A little blue Bori v^_^v Bori! 18:48, 9 June 2019 (UTC)
P.s. And I'm pretty sure those mentioned above that admins have been asking to run would be less reluctant if you adopted a system that actually aligned with the core values of Wikipedia. ClubOranjeT 13:13, 23 May 2019 (UTC)
If anybody is interested in my experience, I was the only admin to ever run a reconfirmation on the Russian Wikipedia, a couple of years I have passed RfA (a couple of other people tried but their reconfirmation was speedily closed by crats). Between RfA and reconfirmation, I was pretty active on an analog of ANI, on AfD, and I also sat for one term in the ArbCom. I passed RfA with smth like 98% votes, and I passed reconfirmation with 97% votes.--Ymblanter (talk) 13:17, 23 May 2019 (UTC)
  • Vanamonde93, with endorsement from Hawkeye7, UninvitedCompany, and Hammersoft, has hit the nail on the head. Kudpung กุดผึ้ง (talk) 02:02, 23 May 2019 (UTC)
  • Well, it would certainly be interesting if RfA was refactored such that Oppose !votes would be discounted unless they were accompanied by explicit incidents demonstrating potential unfitness, and then for reconfirmations those had to be incidents of abuse of tools. I'd find it even more interesting if no two Oppose votes could cite the same incident. I.e. rather than number of !votes having any bearing on an RfA, it would be the number of incidents that was considered relevant. Everyone makes mistakes, and perhaps there should be more focus on the number of mistakes rather than the number of people who view the mistake as such. DonIago (talk) 13:59, 23 May 2019 (UTC)
    • Interesting indeed. Currently a single undisclosed or recent example of sockpuppetry or personal attack can derail an RFA, whilst Deletion tagging errors need to be both multiple and recent. I'm not sure the community is ready to have two RFAs, one passing because we could only find a dozen incidents of sockpuppetry and personal attacks, and the other failing because though all the CSD tagging errors were stale and current CSd taging was exemplary, in the candidates first few months they did multiple A7 tags on articles that asserted significance such that they merited going too AFD, but which by our policy should be deleted via AFD not A7. I'm all for requiring Opposers to give examples or oppose per other people's examples, but requiring each oppose to find a fresh example is OTT. ϢereSpielChequers 09:21, 10 June 2019 (UTC)
  • I find it interesting you use Mother Theresa as an example because... Galobtter (pingó mió) 08:08, 10 June 2019 (UTC)

What is the record for most admins lost in a month?Edit

Is this tracked anywhere? (beyond going through each month manually) Enigmamsg 15:58, 12 June 2019 (UTC)

See Wikipedia:Desysoppings by month. -- Pawnkingthree (talk) 16:01, 12 June 2019 (UTC)
July 2011 was a particularly...firm-bodied vintage :) ——SerialNumber54129 16:03, 12 June 2019 (UTC)
Ah, it slipped my mind that whenever the 'inactivity' policy was implemented would be without a doubt the high mark, as it would be a mass exodus of admins who hadn't edited in a long time. Enigmamsg 16:11, 12 June 2019 (UTC)
Also interesting to note that since the inactivity policy was implemented, our net loss of admins has been at least 20 every year. Enigmamsg 16:12, 12 June 2019 (UTC)
  • Comment Seems we're up to 11 now for June which probably means it's going to be the high mark for 2013-2019 (guessing the carnage isn't over quite yet). Enigmamsg 14:19, 13 June 2019 (UTC)
  • Comment We are seeing a gradual decline in the number of admins .It is 1161 Admins now as against 1045 former admins at this rate we will have more former admins than admins.Pharaoh of the Wizards (talk) 16:26, 13 June 2019 (UTC)
    • I have no doubt that former admins will outnumber admins, because the trends are unlikely to reverse themselves. This year alone a reasonable forecast for the final number in net admins would be -52 or so, which would set the record (obviously ignoring the first year of the inactivity desysops, which had inflated and misleading numbers). Enigmamsg 16:42, 13 June 2019 (UTC)
      • The answer to the header title question is (not counting July 2011 of course, since that's when the policy was implemented) March 2012, when 23 admins were desysopped. Though this doesn't take into account how many admins we (may have) gained that month or any other month, just the number that were lost. IntoThinAir (talk) 17:04, 13 June 2019 (UTC)
        • That would be "one". ‑ Iridescent 17:11, 13 June 2019 (UTC)
        • (Reply to IntoThinAir) Right, I used the information from the chart for my previous comment here. I thought it would be worth noting how this bloody WMF-inducted month stands historically. One could argue, though, that inactivity desysoppings are not on the same level as what we're seeing now, which is a bunch of people voluntarily stepping down (as opposed to inactivity). Enigmamsg 17:19, 13 June 2019 (UTC)
Return to the project page "Requests for adminship".