Wikipedia:Rfc requests for adminship

RfC: English Wikipedia needs to produce more admins

Just recently, I wrote an op-ed for the Signpost about how Wikipedia needs more admins. I now want to see if the community generally agrees with me. There is rational backup for this position, but the details are a bit too long for this page. I strongly recommend that you read this (and the "Stats" subsection) before voting here. I have intentionally removed any personal opinions about causes and solutions from there; it is pure statistics. A large portion of my op-ed is simply my opinion as to why the problem exists and how it can be fixed, but if you haven't already you may wish to read it anyway. Therefore, you may want to read the "Why?" and "Stats" sections of my op-ed. However, realize that the purpose of this RfC is not voice opinions about how to fix the problem, but rather to determine the level of support for the position. Therefore, please do not interpret this RfC as aiming to determine support for my opinions concerning the reasons for, and solutions to, the problem, but rather to simply determine if the problem exists. It is essential to determine if the community agrees with the basic idea before making further proposals which assume that the problem exists. --Biblioworm 16:50, 4 October 2015 (UTC) Edited @ 01:14, 7 October 2015 (UTC) --Biblioworm 01:14, 7 October 2015 (UTC)

Support

1. Support, per this. --Biblioworm 16:50, 4 October 2015 (UTC)
2. Obviously. Sam Walton (talk) 16:53, 4 October 2015 (UTC)
3. Yes because my RFA failed but I am capable of doing the admin work. There are are many users who are capable of doing the job but failed in RFA. This is very sad. I am not saying this because my rfa was failed. I saying this in general. Supdiop (T🔹C) 16:57, 4 October 2015 (UTC)
4. Less admins are doing the same job as before. This manifests itself as less diversity in the judgement used by admins. More admins means more oversight, and a greater likelihood of an uninvolved admin being available. I can think of at least 3 issues on Wikipedia where we are rapidly running out of uninvolved admins. HighInBC 17:05, 4 October 2015 (UTC)

   @HighInBC: Could you give those three examples? They might be useful for the future. --Biblioworm 18:10, 4 October 2015 (UTC)

   I said "issues" but I was thinking of people. I would rather not name them, these people respond negatively to criticism. Suffice it to say that they are people who have been here long enough to encounter pretty much every admin, yet the community has failed to deal with them. HighInBC 18:13, 4 October 2015 (UTC)

5. Agreed, in general. DES ^(talk) 17:07, 4 October 2015 (UTC)
6. Yes--Fauzan^{✆ talk}_{✉ mail} 17:40, 4 October 2015 (UTC)
7. Yes. This thing is usually too big and if you pick an admin area at random (e.g. RFPP), most of the requests/issues/whatever will be dealt with by just a few specific admins. Fortunately, these admins usually happen to be very competent and skilled, but as many different admins as possible should be dealing with an admin area to avoid any chance of bias, so issues can be dealt with efficiently at all times of day etc. — Bilorv_(talk)^(c)(e) 18:35, 4 October 2015 (UTC)
8. Support the wider notion that there aren't enough editors, admins/functionaries/bureaucrats included, to maintain content and to organize Wikipedia. Although many still oppose unbundling the admin toolkit, the community has set a pretty high standard for adminship so why not unbundle the toolkit for editors specializing in other areas? That way, the lack of candidacies at RfA would become less of a problem and editors specializing in specific areas actually will be able to maintain their areas without as much admin intervention. Esquivalience ^t 19:07, 4 October 2015 (UTC)
9. If we want admins to be members of the community who spend a minority of their wiki time doing admin chores then we can't afford to see their numbers keep falling. I think it would be unhealthy for the community if admins became a separate caste who had to spend all their wiki time doing admin stuff. ϢereSpielChequers 21:47, 4 October 2015 (UTC)
10. Support. Everyking (talk) 22:19, 4 October 2015 (UTC)
11. Support Something's wrong, and this (again) acknowledges that fact. I hope this time we can press on to a solution. Miniapolis 22:46, 4 October 2015 (UTC)
12. Support: Absent enough admins, the ones that remain are overburdened, leading to poor decision-making and burnout, even among the best-intentioned. Montanabw^(talk) 01:18, 5 October 2015 (UTC)
13. Support. It's obvious that 250 active admins are not enough to handle the workload. It's more than that, though, because even if we get the number of admins up to the number needed to handle the workload, we'll still need more admins. Admins should not have to spend a huge amount of time using the tools. They should have enough time to work on other personal projects of their choosing without feeling as if they might be letting the rest of us down. So when this process here, now, today leads to a discussion to try to determine how many admins are actually needed, the admins' abilities to spend time writing articles or whatever else they like to do that doesn't require the mop must also be an important factor. Paine  ^{(talk – contribs)}  01:40, 5 October 2015 (UTC)
14. Support. The number of active admins is not enough considering the backlogs. utcursch | talk 02:13, 5 October 2015 (UTC)
15. Support. In the past I was always in denial that RfA was "broken". I've always thought, the process works very well.. it's the people that need to change. But it seems people are changing for the worse! The bar keeps getting raised higher and higher, and people are forgetting that trust should trump all else. I must admit defeat. Something really needs to be done. Or else the backlogs will grow to such a size that the hopelessness of ever getting ahead of the curve will deter even more people from wanting to do admin/maintenance work. That is, if we're not already past that tipping point.. we very well may be. -- œ^™ 04:29, 5 October 2015 (UTC)
16. I'm Mailer Diablo and I approve this message! - 05:49, 5 October 2015 (UTC)
17. Support It needs to be the right kind of admin, obviously. Ideally people who'll be prepared to do the grunt work of clearing the backlogs at WP:CFD and WP:ANRFC to name just two. These don't need to be "content creators", just good people who've got the time to get stuck in. Lugnuts ^{Dick Laurent is dead} 07:12, 5 October 2015 (UTC)
18. Support We need more admins as stated above, particularly per Paine & Utcursch. --Tom (LT) (talk) 08:37, 5 October 2015 (UTC)
19. More admins please. Less backlogs = less burnout = longer activity of those admins we have. Also, more admins can mean more expertise and diversity. —Kusma (t·c) 13:33, 5 October 2015 (UTC)

    That would only be true if we could find perfect recruits. Samsara 19:36, 5 October 2015 (UTC)

20. Duh. -- zzuuzz ^(talk) 18:23, 5 October 2015 (UTC)
21. Strong Support --AmaryllisGardener ^talk 19:15, 5 October 2015 (UTC)
22. Support per User:Lugnuts; we're particularly short of people who feel comfortable assessing consensus, and people who are interested in working with files. -- Diannaa (talk) 19:35, 5 October 2015 (UTC)
23. Support I'm dumbfounded at the Panglossian reactionaries below, who say we have just the right number of administrators. I'm sure they'd say the same if we had half as many or twice as many. I am also sick of the cabal of long-time admins-for-life who've been here a decade or more, and are encountered with fear by everybody else, since not only do they know the arcane rules and the levers to pull and friends to round up in a pissing contest, but they've written a lot of what is now WP policy, and point to it as though handed down by god. Anybody who thinks RfA is a great system should be willing to go through it freshly every two years, so we don't have the same kind of low turnover and lockup on WP that we do in the US congress. Which, indeed, we do. SBHarris 21:20, 5 October 2015 (UTC)
24. Support There are people very experienced and knowledgeable at Wikipedia whose RfAs get close as WP:NOTNOW or who are opposed because of some (relatively) minor incident that happened a while back, and that they have learned from. I, for example, am a very experienced new-page patroller despite my only having 500+ mainspace edits. In general, those Wikipedians who do mostly patrolling work, and not really article-editing work, are underappreciated despite their vital role in keeping the site clean and running. Gparyani (talk) 23:06, 5 October 2015 (UTC)
25. Support. Per everything stated by Biblioworm, above. Good luck, — Cirt (talk) 06:14, 6 October 2015 (UTC)
26. Yes. Cla68 (talk) 06:40, 6 October 2015 (UTC)
27. Yes - the correct statistics here are the response time for real-time issues (vandalism, AN3, UAA, RFPP) and the admin backlogs; the latter clearly shows ever-increasing level of backlogging. עוד מישהו Od Mishehu 06:51, 6 October 2015 (UTC)
28. Yes - as evidenced by backlogs. Cas Liber (talk · contribs) 12:47, 6 October 2015 (UTC)
29. Of course we do. This has been evident for a few years now. The requirements are far too strict as it is now. Rcsprinter₁₂₃ (shout) 10:01, 7 October 2015 (UTC)
30. Support - I've been watching WP editing trends for a long time. I've long argued that the evidence is that total editing numbers have more or less plateaued and do not represent a frightfully plummeting population — as naysayers would have it. At the same time, the crop of administrators has been steadily and inexorably falling. If 60 are lost and 30 replaced (and many of those nominally remaining burning out and stopping with the administrative tasks) year after year after year, pretty soon the surplus of admins are gone and a real shortage presents itself. That's where we are now. The difficulty will be figuring out what to do about it. Carrite (talk) 10:30, 7 October 2015 (UTC)
31. Support largely per Montanabw and Kusma. In the early days of 2005, getting adminship was a lot easier, and yet very few incompetent admins were let in. There's no reason for the extreme paranoia that has set in over the last five years. And I actually disagree with people who want to reform adminship by instituting term limits or making it easier to depose sitting admins: this will just encourage admins to be lazy and afraid to take on challenging situations, which would be as bad as them not being admins in the first place. I think it will also perversely decrease RFA applications even further as people see that adminship is even more stressful with term limits than without it. Also, I think comparing today to 10 years ago is more fruitful than comparing today to 5 years ago, because 5 years ago we also had very few successful RFA's; its just that more admins from the old days were still around then. —Soap— 12:39, 7 October 2015 (UTC)
32. Obviously. This is something we've known for many years. If you think 250 active admins are enough to police five million articles, you are living in a fantasy world. It shortchanges the encyclopedia, it shortchanges the living individuals we write about, and it shortchanges our readers. Gamaliel (talk) 14:16, 7 October 2015 (UTC)
33. Yes we've tried before, but if at first you don't succeed .... Johnbod (talk) 16:50, 7 October 2015 (UTC)
34. The promotion rate isn't the only statistic we need to consider (the number of articles deleted also started to decline after 2007, for instance). However I don't think a promotion rate of about 20 admins a year is going to be sustainable, even assuming we can stabilise at that rate. We are still making heavy use of the admins promoted in the period when RFA was more productive, and those people aren't going to stay around forever. A well populated admin corps takes some time to establish and the last thing we want to do is suddenly adopt drastic measures to quickly create a load of new admins. Hut 8.5 22:00, 7 October 2015 (UTC)
35. Support. The need for more admins is independent of the process used to promote them, and independent of a need to make it easier to remove adminship in a few cases - those are possible reasons why the problem exists, not indicators that there is no problem. Thryduulf (talk) 12:10, 8 October 2015 (UTC)
36. Support Assuming we could find enough qualified admins, we could certainly use more. I find it also particularly interesting that of the opposes, only one is an admin (with 50 admin actions in 3 months), and would have direct insight to how much work admins do and where there are shortfalls. Mkdw^talk 13:14, 8 October 2015 (UTC)
37. Yes, definitely. As well as making it easier to keep the backlogs down, having more admins will mean that individual admins can take more time over each action, thereby improving communication and making mistakes less likely. — Mr. Stradivarius ^{♪ talk ♪} 05:59, 9 October 2015 (UTC)
38. Support. Although I don't often need to call upon admins in the course of my editing here, it's clear to me that the same old faces are doing the same old admin work year after year - and with every year that goes by, less and less admin work seems to get done in a timely fashion, as our sysops are faced with ever-growing backlogs. As Lugnuts observes above, second-tier, non-urgent admin backlogs are in a terrible state (just open WP:ANRFC to see what I mean). Many processes are "staffed" by just one or two admins, and they grind to a halt when those admins take a well-deserved break. My hope is that those users who feel inclined to oppose RFAs at the drop of a hat will take a moment to stop and think about each candidate and what good they could bring to this great project. — This, that and the other (talk) 12:57, 9 October 2015 (UTC)
39. Support. More admins mean that administrative actions can be peer-reviewed to determine if they are appropriate. sst✈ 14:16, 9 October 2015 (UTC)
40. Support. Those stats look pretty convincing. This has worried me a bit for a while. I don't see how this is sustainable in the long run. Jimbo said that in the beginning he envisioned that everyone who'd been here long enough and wanted it would get the admin bit providing they weren't a total asshole or moron (in which I case I suppose they aren't a net asset anyway... so "up or out", maybe). In any case I don't get how having more admins could be a negative thing... right? How to forward om this I don[t know... I doubt the community will want to make RfA any easier unless there's a recall provision, which the community has shown time and again it doesn't want and won't accept... so dunno. Herostratus (talk) 15:30, 9 October 2015 (UTC)
41. it's really surprising to come back after more or less years of absence and see the same few people handling the same massive backlogs (e.g. the xfd's). Despite the surprising lack of attrition, there's going to be some, and if we're not replacing those people, the backlogs are just going to grow. delldot ∇. 17:55, 9 October 2015 (UTC)
42. Yes. Make RFA much easier, (I'd personally drop the bar to 50%, which works fine for Arbcom), and either the sky won't fall in, or undesirables will get through and it will force Wikipedia to finally create the recall process it's been procrastinating over for the past decade. ‑ iridescent 22:29, 9 October 2015 (UTC)
43. Support a lower threshold, maybe around 60% - but still a "majority". As to whether 50% "works fine" for Arbom, I'm not sure I completely agree - but the point is valid. And YES - a better recall process is sorely needed. — Ched :  ?  22:43, 9 October 2015 (UTC)
44. Not just a greater level of workload shared among a smaller number of people but a lower admin-editor ratio makes adminship more of a status symbol and big deal than it should be. A bigger divide between the two groups is not conducive to growing the encyclopedia. Gizza ^(t)(c) 22:50, 10 October 2015 (UTC)
45. Support, though I don't have anything to say that hasn't already been said. —⁠烏⁠Γ ^(kaw), 19:23, 12 October 2015 (UTC)
46. Support. We reached our peak years ago. Now Wikipedia is on a decline and backlogs are piling up like litter in New York City. At this rate, an average active admin has to watchlist 20,000 separate articles in article space alone (ensuring no overlaps at all) to ensure quality. More admins would solve that problem. But you'd need to drop the bar to 2/3 of !voters. Epic Genius (talk) 21:06, 13 October 2015 (UTC)
47. Cautious Support I do think we could use more admins. And I agree that we have become a bit too demanding in terms of what we look for in candidates and their track record. (I say this as someone who has in the past been guilty of this.) But I also think we need to be very careful about promoting dramatic changes to the way we get our Admins. Part of the problem may be excessively exacting standards, and an often unfriendly atmosphere in RfA. Another sizeable chunk of the problem is simply a reflection of falling levels of participation on the project. More editors are abandoning Wikipedia than joining, and active participation has been declining for years. That's a much bigger problem than RfA. Also we would do well to address some of the more common problems that Admins are often spending excessive amounts of time dealing with, including most prominently, IP vandalism. I have long believed that we would severely cut down on petty vandalism by requiring registration as a precondition to editing. All of this said, I think the best answer is to just look for decent editors who have a reasonable amount of experience and a track record that demonstrates both good will and that they have a clue with regards to guidelines and policies and then encourage them to step up. -Ad Orientem (talk) 21:48, 14 October 2015 (UTC)

Oppose

1. RfA produces the right number of admins and I maintain the process is not broken. See my comment below. Chris Troutman (talk) 19:21, 4 October 2015 (UTC)
2. What is needed is an overhaul of wikipedia's basic editing practices (pending changes implemented as standard for all BLP's, required registering a user to edit etc) which would negate the need for many more admins. There are/would be enough admins to handle the workload if the workload was sensibly reduced. Only in death does duty end (talk) 10:06, 5 October 2015 (UTC)

   Implementing PC for 1,091,763 pages would require more user hours, not fewer. And requiring people to register accounts before editing would make sockpuppetry a bigger problem, also requiring more user hours to deal with problems. עוד מישהו Od Mishehu 06:41, 6 October 2015 (UTC)

   Na, a BOT could do it in a jiffy. Unless you are talking about reviewing the pending changes, in which case that does not have to require an admin. Remember the goal is to reduce admin workload, see unbundling. As for registering accounts, almost all online communities/websites that require registering (with greater or lesser amounts of verification) report drops in vandals/trolls/bad hat individuals. Its already a solved problem. Only in death does duty end (talk) 16:04, 6 October 2015 (UTC)

   I strongly suspect that many of the PC reviews are done by admins; of the 15 most recent non-automatic reviews (as many as there were on the first page, since I can't filter out the automatic ones), 4 were done by admins. Even if admins only do about 20% of the reviewing (my statistics here are 27%, although clearly not based on enough data to be meaningful), that's still a lot of admin time. עוד מישהו Od Mishehu 16:23, 6 October 2015 (UTC)

   4 out of 15, but they didnt *have* to be done by admins, so its not an admin task. You could argue applying PC to all BLP's might increase the workload by editors, but the point is to reduce admin-related tasks like applying protection, blocking vandals etc which would disappear once the vandal knows they cant vandalise a BLP. Only in death does duty end (talk) 16:36, 6 October 2015 (UTC)

3. Oppose because the wrong question is being asked. See below for further explanation. --Hammersoft (talk) 15:09, 5 October 2015 (UTC)
4. Oppose Admin bots and unbundling have reduced the amount of work that needs admins. Backlog size seems to be stable. No need to lower our standards. Samsara 17:50, 5 October 2015 (UTC)

   As I mentioned in my reply to Hammersoft a couple of section below, humans are still required and will always be required for many tasks. To address your mention of unbundling, I cannot think of any significant unbundling of actual core admin tools. Sure, we have unbundled rollback, template editor, etc., but we haven't unbundled, and are very unlikely to ever unbundle, the core components of the administrator package. (Deletion, protection, blocking, etc.) And we are unlikely to ever do this (especially deletion), because of the sharp rift that exists amongst the community and intervention by the WMF. Concerning the backlogs, we want to reduce or even eliminate the backlogs by getting more admins and therefore prevent backlogs while at the same time keeping the work load down for any one administrator. As for your last sentence, we are not voting on lowering our standards. That will come later. Right now, we are voting on admin production in general. --Biblioworm 18:43, 5 October 2015 (UTC)

   we are not voting on lowering our standards. That will come later. So there is the agenda everyone was asking about. Which would be why it's important to stop this nonsense in its tracks. Samsara 19:08, 5 October 2015 (UTC)

   Well, reading further above would have made it clear that we are having this RfC for this very reason. (Namely, to establish that the issue exists before starting new RfCs which assume that it does.) I said it from the very beginning in my opening statement, and I also told Liz that very clearly. This is not some "aha!" moment that has revealed some secret. And I suppose you are entitled to think of having open, healthy discussion about an issue to be "nonsense". --Biblioworm 20:17, 5 October 2015 (UTC)

5. Oppose I've not noticed any significant problems arising from a lack of admins. Compare this with AFD, say, where there's a lack of editors commenting on the topics and so discussions are rolled forward repeatedly. It's ordinary editors that we seem most short of and it might be that all the attention given to admins is counter-productive, making the ordinary content creators and maintainers feel unloved and unwanted. Consider RAN, for example. The poor guy just wants to crank out bios of obscure individuals but there seems to be no shortage of admin types determined to stop him. Andrew D. (talk) 18:14, 5 October 2015 (UTC)

   Of course, a non-admin might not notice any problems for a very obvious reason: they're not admins and don't have to deal with all the things that admins must take care of. The statistics are very clear in showing that our admin production is insufficient. Our backlogs never go away, even though we have dedicated admins working around the clock to keep them down. See this category if you don't believe that we have backlogs. Researching the topic will reveal things that casual observation will not. --Biblioworm 18:43, 5 October 2015 (UTC)

   I am aware that there are backlogs but they mostly seem to be busywork rather than anything that matters much. For example, the biggest backlog seems to be for usernames but that seems mostly a waste of time. For example, see the complaint at User talk:Jeff.jeff.jeff.jeff.jeff.jeff.21. This says "Usernames must not exceed 40 characters" which seems to be an invention not found in the policy. And, in any case, the name is shorter than 40 characters. We shouldn't be wasting time on such stuff when the real work which needs doing is in article space. Andrew D. (talk) 19:00, 5 October 2015 (UTC)

   Backlogs are not meant to go away. They're just supposed to remain stable, which they have been for a long time. There are many reasons for backlogs that have nothing to do with availability of hands. Samsara 19:22, 5 October 2015 (UTC)

   The statistics are very clear in showing that our admin production is insufficient. Except Hammersoft has shown that that is not the case, and you seem to feel you have a license to just ignore his relevant evidence and pretend it doesn't exist. Samsara 19:26, 5 October 2015 (UTC)

   They are not meant to go away? We are to be satisfied with simply keeping them stable, no matter how much work there is? The definition of backlog is "an accumulation of something, especially uncompleted work or matters that need to be dealt with."[1] The template for marking backlogged categories says that backlogs need attention from admins. If what you say is the case, then why do we have backlog drives in all aspects of Wikipedia? Why be satisfied with an inferior solution ("just keep them stable"), when getting more admins would result in more even distribution of work and perhaps even elimination of the backlogs? And finally, I addressed Hammersoft's evidence in exhausting detail and explained the discrepancies between his data points, and the inconsistencies between the short-term data and the more important long-term data, so to the contrary no one can in any way pretend that I'm ignoring what he said and pretending that the data is not there. --Biblioworm 20:17, 5 October 2015 (UTC)

   In reply to your comment, Andrew, it is true that many of the backlogs are busywork, but there was also a previous discussion about this in this very RfC. It was pointed out the even simple tasks can become extremely tedious and boring when there is a lot of it to do. For instance, when I edited another wiki, I was one of the very few people who dealt with the reference error backlog. Usually, it was just a matter of adding the same chunk of text on articles. But there were dozens of these articles and the work felt very difficult after a while. As for the username example you gave, that probably was a silly invention, but otherwise there are some seriously disruptive usernames to deal with. And the maintenance work is extremely important as well, so it's hardly a waste of time; otherwise, the site would descend into chaos. --Biblioworm 16:48, 6 October 2015 (UTC)

6. Oppose. The community is understandably reluctant to elevate editors to lifetime positions unless it is positive that they will behave properly. Accordingly, RfAs make it as hard to become an admin as it is to remove one. Improve the process for dealing with abusive and inept admins and this problem will self-correct. Coretheapple (talk) 13:12, 6 October 2015 (UTC)
7. Oppose Hammersoft, Samsara and Andrew D. have the better arguments here. Backlogs are part of life, any problem can be solved only after it has appeared, was acknowledged, was debated/considered, had solutions proposed and the latter eventually implemented. The backlog is the amount of problems that are within this time window (from appearance to solution), so, by definition, there can not be any time without a backlog, except in Utopia. Kraxler (talk) 15:30, 6 October 2015 (UTC)

   Excuse me, Kraxler, but I put much time and effort into refuting Hammersoft's and Samsara's arguments. They are invalid and not supported by the facts and data. Please read this diff for a short summary of why this is the case. --Biblioworm 15:35, 6 October 2015 (UTC)

   You have certainly put much time and effort into trying to refute them, but you failed. And thanks for pointing to useful links, but I'll use my own brain here. Kraxler (talk) 15:39, 6 October 2015 (UTC)

   The problem is, you look at statistics that show increases in base numbers and conclude that this means more admins are needed. For a start that is one possible solution to lowering the individual burden, however it makes no reference as to what the burden actually is. Is it onerous? Were the actions taken in (for example) 2007 more time consuming compared to today's suite of tools? Were the admins previously working at 30% of their available capacity and are now working at 50%? Have the tasks changed significantly? Do we now have more invested admins at the top end? Essentially looking at your posts and rebuttals it comes across more as a 'this is the conclusion, here are the statistics to prove it', rather than 'these are the statistics, what is causing this?'. Only in death does duty end (talk) 15:48, 6 October 2015 (UTC)

   @Kraxler: How did I fail? I showed rather clearly that they reached completely illogical and unsupported conclusions with their "data" and "evidence". I hope you at least read the link. @Only in death: Yes, I am trying to prove a point. A person trying to prove a point should show statistics to prove it. In regard to your assertion that I should have a "these are the statistics, what is causing this" outlook, I did that already. I looked at the statistics, and I reached the conclusion that the evidence is quite unambiguous. In fact, even Hammersoft's data, intended to prove the opposite conclusion, could in fact be construed as evidence for my position. For instance, since we seem to be very interested in correlations, does anyone find it interesting that the average burden per admin in the top 30 has increased by 29% since 2005, and by 10% for every admin since 2010? It's an interesting correlation that the individual burden has been increasing as promotions are decreasing. In regard to "burden", I'm afraid that's something that cannot be accurately quantified in numerals, but the stats I just showed to demonstrate that it has been increasing. We know, however, that there is a burden if the backlogs still exists even though admins are performing hundreds and thousands of actions while not having any noticeable difference on the amount of work. --Biblioworm 16:48, 6 October 2015 (UTC)

   Taking the 29% figure in isolation is misleading. The fact, in isolation, that average admin burden over the last ten years has remained static is misleading. The fact, in isolation, that average admin burden over the last five years has increased 10% is misleading. None of these figures mean anything in isolation. Using any of them to prop up any conclusion is a false conclusion. Extraordinary claims require extraordinary evidence. We don't have it. In fact, rather the opposite. --Hammersoft (talk) 16:59, 6 October 2015 (UTC)

   ...which is why I said "since we seem to be very interested in correlations". I was saying it in response to the contrary attempts to prove a point using this data. So, I gave an "according to that logic" example. You're right: alone, they mean nothing. But overall, it is very interesting that was have these interesting correlations (burden increases as promotions decrease) and the presence of data to show that the average admin is doing more work than they were doing a few years ago. Curiously, they seem to support the idea that we should get more admins to reduce the average workload. --Biblioworm 19:31, 6 October 2015 (UTC)

   That is your interpretation, yes. --Hammersoft (talk) 20:21, 6 October 2015 (UTC)

   @OID; I removed myself from the conversation below because it was apparent that we were talking past each other with statistics. There's the old adage about "There's lies, damned lies, and statistics" and it's true. The statistics can be used to support lots of views, even ones that directly contradict one another. The idea that statistics from a very small data set absolutely, irrefutably conclude anything is one I find objectionable. It's kinda like saying the flow rate in the Colorado River below Lake Havasu is too low, and therefore concluding there isn't enough rain in the state of Colorado. There's a blizzard of factors that play into that rate. Data taken in isolation is insufficient to conclude anything, much less in a system as complex as Wikipedia. --Hammersoft (talk) 16:59, 6 October 2015 (UTC)

   FWIW, there's more on the distribution of admin actions in 2014-5 vs 2006-7 in these graphs based on data collected in May. My interpretation of that data was that admin actions are being concentrated in fewer hands: compared to the old data, we now have many more admins who take only a few actions, and many fewer at the top of the activity distribution. There is wiggle room there for redistribution of some actions into flagged adminbots, but the basic conclusion is pretty clear. Of course that makes no normative claim about how many admins there "should" be.

   From what I can tell the Foundation's data science people produce self-contained research projects rather than data in a form that is useful feedback for the community, but they really ought to pay a couple of Berkeley kids to be the on-call data gofers (er, ahem, "Applied Data Science Interns") for this kind of relatively small-scale question so people don't waste their time gathering redundant datasets too small or arbitrary to draw conclusions from. Opabinia regalis (talk) 20:49, 6 October 2015 (UTC)

   @Opabinia regalis: Your beliefs seem to be based on unstated and untested assumptions. Not long ago, I was briefly interested in page protections, and ended up concluding that it was futile trying to get a significant share of that cake because the admins who've been doing that work for longer are just that much better at it. Which is consistent with what Hammersoft found. Every backlog I've ever looked into has been dealt with and, on long-term observation, been stable. In fact, there is a problem in that some "backlogs" do not really exist, such as Category:AfD debates relisted 3 or more times, which is mostly populated with items that have been dealt with (I looked at ten items, and nine had been closed). Nonetheless, it is officially designated a backlog. Which goes back to what I've said before: we love tagging things, and we forget to untag them. Samsara 22:22, 6 October 2015 (UTC)

   @Samsara: I don't follow. I described the difference in distributions. That's not a belief. I also didn't say anything about backlogs. It could be true that having fewer admins taking a greater share of the admin actions is a good thing: maybe higher consistency and specialized skill development is worth the tradeoff against greater diversity and less vulnerability to changes in individual admins' activity level.

   As for that AfD category, it's always cited as the Dumbest Backlog. Why does it still exist? Opabinia regalis (talk) 23:33, 6 October 2015 (UTC)

   @Opabinia regalis: I alluded to the admins who've gotten extremely good at certain tasks. This is why they claim an increasing share of admin actions. I don't see any demonstration that it's anything other than a simple training effect. Samsara 23:43, 6 October 2015 (UTC)

8. Oppose in a perfect world, no admins would be needed, and actually most things adminy should be able, and many are, done by non-admins (we should devolve more), yet, recognizing this is not a perfect world - we do have some admins - the present admin corp appears reluctant to do the tool needed work, so it is doubtful "new admins" will solve anything - address the type of work (automate?) or the admin corps' reluctance to do it (incentives and accountability -- eg. one idea, out of many -- in some areas make identified admins responsible for this or that board, backlog, or say, the main page - so everyone knows who to talk to about it - and we know there are enough interested in doing it - or some other assignment mechanism - with perhaps rotation; thank the non-doers for their service and give them a t-shirt, and tell them it's time to move on to bigger and better things on this project than holding a mop). Then too, this idea that we need "more admins", without answering the 'how many' and 'for what' is backwards. Alanscottwalker (talk) 15:53, 6 October 2015 (UTC)

   @Alanscottwalker: The hard facts and data show that most new admins really do contribute to administrative work (see my reply to Axl above), so to the contrary there is no proof for the idea that the present admin corps are reluctant to the needed work. The AdminStats tool shows that hundreds of admins are performing hundreds or thousands of action within a couple of months. Finally, I can answer the two questions you presented at the end. Although it is difficult to say how many admins we need (that's really something unrealistic to expect), this was already discussed. The answer is that we know that we are somewhere around the desirable number when we can eliminate most of the backlogs and the burden on individual admins is reduced (and before I hear the tired answer that backlogs are not meant to be eliminated, I would advise any would-be repliers to read my second reply to Samsara here). To answer the "for what" question, we are doing this so that we can significantly reduce or even eliminate some backlogs. (I repeat: wouldn't you feel like your hard, dedicated work was going to waste if you were performing hundreds to thousands of actions to handle the backlogs and yet they never went away? To anyone thinking of objecting, think about yourself in the position of the admins.) Just as importantly, we are also doing this to reduce the burden on individual admins. As I mentioned above, the statistics show a general increase in the burden of each individual admin since 2005 and 2010; the statistics show this for both the top 30 group and the admins as a whole. Admins are WP:VOLUNTEERs like the rest of us. They shouldn't feel like they have to do so much work just to keep this site running. Perhaps some want to work on other things on-wiki, or maybe they want to do more things in real life. Obtaining more admins will contribute to distributing the workload more evenly; as I mentioned above, the statistics also show that most admins elected via RfA go on to become active. --Biblioworm 16:48, 6 October 2015 (UTC)

   No. They don't want to do it, that's why it is not getting done. More admins, as experience demonstrates, will be more admins that do not do it. Sure, they are volunteers -- they volunteered for the tool, and by their (in)actions, they are volunteering not to do what it is needed for. -- Alanscottwalker (talk) 18:39, 6 October 2015 (UTC)

   With all due respect, you continue to assert that admins aren't doing any work even though the data is plainly against the notion? And furthermore, my detailed response to the questions you presented is just dismissed with a "no"? Of course, many of our admins from past years have moved on naturally, and there's not much we can do about that. But we still have a very considerable number of dedicated admins who work hard. I'll give a hard statistic: over the past month, there have been almost 180 admins who have performed at least 30 actions, 115 admins who performed at least 100 actions, 35 who performed over 500, and 20 who have performed over 1,000. This is just over the past month. And yet people are saying that most admins don't work hard? And even if what you say is true, how do you suppose that we'll get more admins to jobs they don't want to do? What if they don't want to? They're not getting a paycheck for it, so they can refuse to do anything they don't want to do. The best solution, more realistic than forcing admins into jobs they dislike, is to get more enthusiastic new blood that will reduce the workload across the board so that each individual admin will have do less work individually. As I said earlier, the data presented by Hammersoft shows that there has been a general increase in burden for each individual admin over the past 5–10 years. --Biblioworm 19:31, 6 October 2015 (UTC)

   No. I do not say that individual admins are not doing work. On the whole the corp of admins is large, but it's apparently not getting done (or, perhaps these backlog tasks do not need to get done, and if that is the case - who cares). I also did propose ways to address it -- deal with attractiveness of it and their accountability for the work flow. Alanscottwalker (talk) 21:38, 6 October 2015 (UTC)

9. Oppose - I believe 250 is fine, I do wonder at times if RFA is too strict but then again if it wasn't strict it'd be hell here!, Anyway I don't think there's anything to worry about & don't really think there ever will be. –Davey2010^Talk 15:56, 6 October 2015 (UTC)

   Of course you don't worry about it, since you're not an admin and don't have to do the work. Certainly, I have written tens of thousands of words within a few days that expressly discuss why 250 active admins is not enough for a website that has endless work to do and for a website that is two places more popular than Twitter. Don't be deceived by plain numbers. Finally, I don't believe there is any hard proof for the assertion that "it'd be hell here!" if we let some more people become admins and loosened our requirements a bit. We have methods for holding the occasional consistently abusive admin accountable, if people would get over their idea that we don't have any such thing and actually try it for themselves. Opposers have three times more power than supporters. Why should we have such a negative outlook? WP:AGF is an offical guideline. Paying three times more attention to the negative is not really assuming good faith. Like I said in my op-ed, we have a more unrealistic bars for "consensus" at RfA than virtually any other real-life group outside the wiki-world of Wikipedia. Many of these groups perform much more important tasks than electing admins for a wiki. In fact, even on-wiki, RfA is the only process with such a strict requirement. The range for so-called "discretion" is in fact very narrow. In other areas of the project (even proposals with wide-reaching effects, for instance), the closers have much more discretion that a bureaucrat does when closing the discussion. --Biblioworm 16:48, 6 October 2015 (UTC)

   Admins don't "have" to do the work either. I've seen a number of ANI threads in which admin action was urgently needed, and no admin will step up to the plate. That's not because there aren't enough admins but because admins don't want to intervene. Ditto for SPIs. Some sit there for days. They can be long and convoluted. You could double the number of admins and I doubt it would change the reluctance of admins to step into messy disputes. Coretheapple (talk) 17:13, 6 October 2015 (UTC)

   Of course they don't have to do the work, so maybe that wasn't the best wording. A better way of saying it would be, "You don't worry about it since you don't do the work that some admins do to keep this website running properly." And as for matter of admins not intervening, I replied to Alan about that just now. --Biblioworm 19:31, 6 October 2015 (UTC)

   Something that Coretheapple just said strikes me as an excellent insight. Come to think of it, I see lots of cases where there appears to be a backlog, but it isn't because there are no admins around, but rather, it's because the issue needing admin action is a messy one, and many admins would rather just leave it to "someone else". The problem is not having enough "someone elses", as opposed to not having enough admins. In other words, we need more admins who are comfortable dealing with the ugliest situations, including the situations where they will come under fire for taking action. --Tryptofish (talk) 20:26, 6 October 2015 (UTC)

   And it is one that merits further praise from me as well. I remember once talking to, I think, The Blade of the Northern Lights regarding AE. He at the time said in many of the matters brought to that page, it took him several hours to read through everything before he felt comfortable making a statement in the Admins only section. The number of individuals of any sort, admins or not, who are willing to spend several hours of their time on a single act here is probably very, very small. Particularly when in the same time that same individual could probably do several page locks, page moves, XfDs, and so on. And, honestly, I am not sure that simply increasing the number of admins will in any way address the question of having enough admins to do some of the time-intensive, or in the cases Coretheapple spoke of controversial, messy, or difficult cases, which need to be addressed. How could we, basically, make it possible to find, well, more admins with guts? (acknowledging, much to my regrets, the temporarily loss of User:Floquenbeam, who struck me as one of the best of that sort.) John Carter (talk) 22:02, 6 October 2015 (UTC)

   I had in mind a specific ANI thread from a few months ago concerning a problematic user. The thread went on and on and on. In such situations, one frequently gets "tl;dr" comments from admins. Ditto very long SPIs. Can't say I blame them. It's a bit like when I use my STiKi vandal-fighting tool. Some really convoluted edits are hard to decypher, so I "pass." We're all volunteers and our time and energy are limited. As for getting admins who are willing to weigh into messy disputes - just because an editor is willing doesn't mean he is necessarily the best person to do so. Editors with the "guts" to do such things may also be objectionable on other grounds. That brings me back to the need to make adminship less of a "roach motel" lifetime position, so that admins can be removed with less drama than the process currently requires. Coretheapple (talk) 13:25, 7 October 2015 (UTC)

   I think that you both made very apt points, and for me, this part of the discussion is becoming very interesting. It's like there are two sides to the coin: On one side, we need more admins who have both the willingness and the competence to spend significant amounts of time going carefully through difficult cases, without crying TL:DR (which is essentially the same thing as kicking it upstairs to ArbCom, where they have to spend the time), and who can do so in a way that results in a good outcome, that upholds community norms. That's a tall order, in part because it also requires that the community support such admins, instead of beating up on them, especially since we have a bottomless supply of users who will falsely claim abuse. The other side of the coin is that we actually need fewer admins who will step in, but do it the wrong way. In that regard, I do not think that editors opposing here are necessarily paranoid, because there really are cases of abuse. Here's a question about data: I bet someone has compiled data about those admins who have been desysopped by ArbCom for misuse of tools: are there any trends of those ex-admins having passed RfA a long time ago, versus recently? --Tryptofish (talk) 14:39, 7 October 2015 (UTC)

   Hi Tryptofish. In regard to your inquiry about data concerning desysoppings in relation to the year they passed, there is a fundamental flaw with that system. The flaw is that admins elected a long time ago have had a much longer time to be desysopped, whereas more recent admins have not had as much time. Therefore, there will almost certainly be more admins desysopped from longer ago than more recent admins, because the probability of being desysopped would increase as time goes on. I will try to put together some data on this, though. --Biblioworm 15:17, 7 October 2015 (UTC)

   Thanks for that! I guess it's inevitably messy: on the one hand, the greater amount of time for something to go wrong, but on the other hand, the arguably rising community standards at RfA and the increasing effectiveness of ArbCom over time. Personally, I don't put much weight on the argument that long-serving admins are by definition at greater risk, because if someone is trustworthy, they will tend to continue to be trustworthy, but that's my opinion and I probably cannot prove it. --Tryptofish (talk) 15:24, 7 October 2015 (UTC)

   Would there be any sort of way to maybe, somehow, create a system which can generate, for lack of a better comparison, a system in which we can get some sort of "drafting admin" similar to the drafting arb who can maybe go through some of the complicated and lengthy discussions and summarize the data for one or more other admins or others to review and decide upon? Maybe something like a tribunal, or, God help us, "ArbCom-lite"? Not all those others involved would necessarily be admins, of course, and, I guess, maybe, making it up as I go along here, maybe individuals chosen for the MedCom or DRN or other pages, along with maybe some others, might be able to do some sort of possibly "rotating service" on the tribunals. They wouldn't necessarily have more of a !vote than anyone else who might take part in the discussion, but they might at least point out what they see as the relevant policies and guidelines, how they might be involved, and what if any form of solution seems reasonable to them. John Carter (talk) 20:05, 7 October 2015 (UTC)

   Well that might be worth exploring. And it doesn't have to be a lengthy or complicated discussion either. Some perfectly reasonable things come before ANI and are just ignored and archived after only cursory attention. Well-intentioned busybodies might be chosen at random to weigh in on such things, just as RfC commenters are summoned by the RfC bot. Coretheapple (talk) 20:22, 7 October 2015 (UTC)

   I don't think the problem is really that admins need someone, like a "drafter", to decrease the time it would take for them to deal with problems. I think the problem is the ability and willingness to deal with difficult cases. I can picture a "drafted" dispute that admins would still be reluctant to touch, and it's really better for the person making the decision to base it on an unfiltered examination of the facts. --Tryptofish (talk) 21:10, 7 October 2015 (UTC)

10. Oppose question as phrased, although I acknowledge there might be another, more useful, way of phrasing it. The number of admins is not the single best indicator of admin work output. Therefore, on that basis, I think the question being asked is wrong. If there were another question asked, perhaps along the lines of "can we find some way to reform RfA?" or "can we find a way to reduce the workload on admins?" or "can we find a way to encourage more people to seek adminship?" all those questions may conceivably be able to generate some useful results. But there seems to me to be an inherently flawed presupposition in the phrasing of this question, if it is in some way intended as a prologue to further discussion on how to maybe maximize the results and "success," however that is defined, of RfA and of the admin corp in general. John Carter (talk) 20:22, 6 October 2015 (UTC)
11. Oppose. When I first joined Wikipedia I remember a fair amount of sqwaking that the project is in imminent danger of catastrophe due to something or other. It's now over 10 years later and some folks still think that way. If the sky is falling, it certainly is taking its time about it. Andrew Lenahan - Starblind 18:26, 8 October 2015 (UTC)
12. Oppose Adminship is supposed to not be a big deal. I think endorsing that we need more admins is the wrong viewpoint. We need the current admins to step up and deal with the issues that the community has reported and come to a rough consensus about. We need our current admin corps to stop taking unilaterial decisions that they know are going to be contentious and cause them to end up at any of the drama magnets (WP:BN,WP:AN,WP:RfArb,WP:ANI, etc). We need to see what admin-lite actions are the most time consuming that don't require as high of level of consensus on (such as page protection) and give those as things a qualified and trusted user could do (while still leaving it in the Admin toolkit). "Given enough volunteers, all backlogs are only a day old". Furthermore I do also agree that the standards for candidates at RfA are ludacriously high and duplicious. You can't be an Admin if you don't do content creation (meaning you need to be in article space), but you also can't be an admin if you can't demonstrate that your judgement calls in relation to admin tasks align with consensus (meaning you need to be in WP project space). Hasteur (talk) 18:51, 9 October 2015 (UTC)
13. Oppose I'm not convinced that we need to lower the bar to let more people pass RfA, in fact I think that's a disastrous idea. The minimum qualifications should be more stringent if anything. I also think we have lots of admins that aren't actively helping with the backlogs, and those admins should be desysoped or asked to help out more. To simply add more admins will not solve anything, because there is nothing that requires them to be active in backlogged areas. I think admin is a just badge for many, and they use it primarily to help their friends. Make admin backlogs a required duty to retain the bit, then you'll see the backlogs disappear. RO^(talk) 19:07, 9 October 2015 (UTC)
14. Oppose Adminship is a vanity project. I'll only see it getting worse the more lenient it becomes. JAGUAR  20:25, 9 October 2015 (UTC)

    @Jaguar: ...the more lenient it becomes? What? Adminship has not become more lenient; that should be obvious to anyone who has studied RfA. Please show the real evidence for this position. --Biblioworm 20:31, 9 October 2015 (UTC)

    I think I could have phrased that better. I don't fully understand what is being proposed here in principal; the majority of people agree that more people have recently failed RfAs and less people have been successful. That's fair enough, but I was under the impression that we have enough admins here as it is? Less people have been successful at RfAs, but that doesn't mean we're running out of admins! There are plenty of admins here to keep Wikipedia here running and to perform administrative tasks such as protecting pages, deleting articles and closing threads per its consensus reached. I really don't see the need to "lower the bar" as people here have mentioned. I know RfAs are meant to test every aspect of an editor, and wasn't suggesting that they're in any way lenient (would that be bad for business), but I don't know why we need to let more people pass RfAs just for the sake of more people being unsuccessful. I don't understand how we would enforce the idea to bring more admins in the community? Do we need more? What other way is there to help people be successful without altering the criteria of adminship? I'm not sure if this is a good idea as such a change might only disrupt the balance of things. JAGUAR  20:43, 9 October 2015 (UTC)

    Jaguar, I have dealt with all these objections very extensively before. I have written thousands of words in the past few days which detail why we don't have enough admins. Many admins are in the support section, and relatively few in oppose section, so it seems like the ones who actually do the work generally feel that they could use more help. I've made charts showing how we're losing active admins; see this one, for instance. It shows that we have less than half the number of active admins we had at the beginning of 2008, and yet we have over two times more articles today. I've pointed out that our backlogs continue to exist, with as much work to do as ever, even though we have dedicated admins working all day every day. Hammersoft, although interestingly trying to prove the opposite point, presented data which showed that the burden per admin has increased over the years. Desysoppings have outpaced RfA promotions by over 400% this year (see this chart. What more evidence can you want? --Biblioworm 21:08, 9 October 2015 (UTC)

    But you haven't addressed the issue that many admins are active, but not helping with the backlog. Create annual duty rosters that must be completed to retain the bit, and the work will get done, but as it is now we could have 3,500 admins, but if they aren't working on backlogs the backlogs will never end. RO^(talk) 21:18, 9 October 2015 (UTC)

    Apologies for not addressing the issue sooner. I'm still trying to get up to speed with all the discussion here. But here is my answer: The plan being proposed will likely not fix any problems. Many inactive admins left for reasons beyond the control of Wikipedia. They might not even watch their talk page any more, so they might not even notice if we request that they help with the backlogs. Even if they did watch their talk page, they might decide they don't care and not respond. For instance, we currently warn admins before we desysop them (more than once, in fact), but the majority do not respond. The ones who do usually only make an edit or two just so they can keep their bit. The ones who request the bit back after desysopping usually don't become active. As I've mentioned before, admins don't get paychecks, so they can do as much as little as they please. It's a WP:VOLUNTEER service. The only on-wiki consequences we can impose are desysopping, blocking, and banning. Of course, no one but a purely ridiculous person would ever suggest that we block or ban inactive admins, so that leaves us with desysopping. Making the requirements for retaining the bit more stringent and carrying out a mass desysopping of those who currently aren't meeting an activity threshold for admin actions will in fact show us how few admins are actually active relative to our deceiving statistical number. The "annual duty roster" you suggested would probably be full of problems. How do we decide how much admins should do? How do we decide what they should do? What if we require them to do things they don't want to do? (Remember, WP:VOLUNTEER.) What if admins become frustrated and resign in protest? What if the strict requirements for maintaining the bit discourage people from trying to become an admin? All these potential problems, coupled with tightening the RfA standards as you suggested in your oppose vote (I don't see the logic behind that at all, by the way), will simply speed up Wikipedia's plunge in active admins and will in fact prove my point even more. (But some people would still be in denial even if we had only fifty active admins.) The fact is, even if we did get some currently inactive admins to become active, we are still relying immensely on the old generation of admins promoted during the old days of RfA. We would not be anywhere near where we are not if RfA had been the same back then as it is now. The tenure of these old-timers cannot last for ever. In time, they will start retiring and get desysopped because of inactivity. It's already happening now, but this will likely become very apparent over the next few years. The number of admins we're now producing via RfA will be nowhere near enough to replenish the number we're losing, as I showed in a chart I linked to above. --Biblioworm 22:14, 9 October 2015 (UTC)

    Yes. It's volunteer, I know that quite well as a content creator. You are so certain that the problem is we don't have enough admins, but maybe we have too many that use it primarily as a badge, and don't keep up their end of the bargain. I'm not talking about admins that are so inactive you can't contact them at their talk page, but lots of admins do little more than play schoolmaster at An/I, which I know is needed, but working on backlogs is also important. I don't know what the answer is, but the idea that we need to lower the bar when we had a candidate pass RfA yesterday at 94% seems misguided. The bar isn't too high, but many of the candidates are unsuitable. RO^(talk) 22:25, 9 October 2015 (UTC)

    Yep, I knew it. I was waiting for it, just as I predicted in my Signpost op-ed. The reform efforts are going to inspire more RfAs (this always happens), people are going to say "it works after all, no reform needed!", and therefore reject the reform proposals. Then, the hype dies down, everything returns to normal, and the cycle starts all over again. I will emphasize this very boldly, so that more people don't fall into the vicious cycle: Brief surges in RfA promotions are not representative of the process's long-term performance. Do not be deceived when this happens, as it always does. I also fail to see how a candidate passing at 94% percent is related to the arguments at hand here. It's an illogical connection to say that since a candidate passed at 94%, we don't need to fix RfA. Much more accurately, Ian was one of the few users (yes, I say few; look at our promotion numbers) who managed to pass RfA because he had edited for a few years, had tens of thousands of edits, and played out the whole thing rather nicely. He also got the right crowd to vote at his RfA. (Realize that RfA is an incredibly inconsistent process; it all really depends on who decides to vote.) An editor just like him might run a few months from now and fail, because a different group voted there, or because there was some particularly avid dirt-digger who found a problem deep inside the candidate's history, inflated its importance, and dragged several more users to the oppose section with him, causing a characteristic pile-on and a failed RfA. --Biblioworm 22:46, 9 October 2015 (UTC)

    "An editor just like him might run a few months from now and fail, because a different group voted there, or because there was some particularly avid dirt-digger who found a problem deep inside the candidate's history, inflated its importance, and dragged several more users to the oppose section with him, causing a characteristic pile-on and a failed RfA.". Incredibly insightful and observant. I'm impressed. — Ched :  ?  04:55, 10 October 2015 (UTC)

In between

1. I agree with the statement that administrative backlogs are longer than optimum, and that's a good reason to support. And I found the Signpost piece to be very good reading. But I see a problem with some of the evidence presented in the background reading, in that the fact that many more candidates used to pass RfA in the past than do today is not really evidence that the RfA standards have become too high. It may very well be the case that our problem is that there simply are not enough non-admin editors who could really be trusted with the tools (and who are interested in using those tools) to fill the need that is seen in the backlog. I'd hate to see a significant lowering of the passing percentage, but I welcome a discussion amongst editors, in which some editors may perhaps be persuaded to change their RfA expectations. --Tryptofish (talk) 21:46, 4 October 2015 (UTC)

   Hi Tryptofish. First of all, I plan to study RfAs from previous years and see if there were considerable differences in their level of experience was lower. Then, I will analyze whether or not there is genuinely evidence to suggest that admins elected back then (2007 and before) were considerably more troublesome than admins elected when the standards may have started rising (2008 to present). Secondly, I never voiced any opinions of my own in the user page you linked to. I intentionally removed all opinion from it. I simply presented hard, statistical evidence to show that our number of admins is insufficient. Perhaps you are confusing it with the Signpost article, which does contain my personal beliefs? --Biblioworm 22:45, 4 October 2015 (UTC)

   Not that it really matters, but I was considering, for example, the graph that appears in both the Signpost and the userspace page, showing the numbers of successful RfAs per year. I meant it when I said that I was very interested to read the Signpost piece, so thank you again for it. As for where I'm confused, it's not really about that, but I do agree with what Liz said just below. --Tryptofish (talk) 14:09, 5 October 2015 (UTC)

   Just FYI: you are going to have to do some pretty careful normalization in order to be sure your observed "differences in experience" aren't artifacts of unrelated trends. I looked into doing this awhile back but didn't have the time to do it right. Opabinia regalis (talk) 04:00, 5 October 2015 (UTC)

2. I'm not sure what we are voting on here. Could you make a statement in one or two sentences that summarizes what you are proposing? Because right now, it seems like a validation of your Signpost op-ed and that's not really an effective focus for making future plans. Liz ^{Read! Talk!} 01:58, 5 October 2015 (UTC)

   @Liz: We are voting on whether or not RfA needs to produce more admins. We must know whether or not there is general consensus on this issue, because there may be discussions and/or RfCs in the future that assume this fact. I do not consider it a validation of my op-ed, because I am asking for confirmation of the issue itself, not my opinions about the causes and solutions. --Biblioworm 15:18, 5 October 2015 (UTC)

   @Biblioworm: okay, just so there is no confusion on what people are supporting or opposing, at the end of your paragraph, can you add: RfC:English Wikipedia needs to produce more administrators. Support or Oppose? Liz ^{Read! Talk!} 16:53, 5 October 2015 (UTC)

   @Liz:  Done --Biblioworm 14:00, 6 October 2015 (UTC)

   @Biblioworm: I must be having bad vision because I don't see that you added the statement to the paragraph. Liz ^{Read! Talk!} 16:43, 6 October 2015 (UTC)

   @Liz: I decided to change the title instead. The change should be more obvious in the title than in the paragraph. Besides, I wrote the paragraph over two days ago and it would look somewhat awkward to simply add a phrase at the end asking if the editor supports or opposes. --Biblioworm 16:52, 6 October 2015 (UTC)

   You're right, I just assumed the title was unchanged and read over the paragraph. It seems clear to me now. Thanks! Liz ^{Read! Talk!} 17:03, 6 October 2015 (UTC)

3. Neutral. I've said it before, and I'll say again: Looking at this as an "Admin" vs. "No Admin" problem is missing a good part of the issue. The answer is to revamp the project so that it can get by with fewer (active) Admins. Unbundling more of the tools, so that more routine actions like page protections and page moves are done by trusted veteran editors instead of Admins, so that Admins aren't wasting their time on stuff like this, is the solution. --IJBall (contribs • talk) 03:49, 5 October 2015 (UTC)

   @IJBall: I think we should address the heart of the problem rather than simply attempting to "get by" with a few admins. Long-term, centralized fixes are always preferable over short-term patches. I would be better to focus on improving one process rather than trying to overhaul the entire project, which will consume much more time and energy, and furthermore is much more likely to ultimately fail. --Biblioworm 15:18, 5 October 2015 (UTC)

   FTR, I'm not opposed to what you're trying to do here. I just think we're coming at this from opposite viewpoints: you seem to think that there will never be tool unbundling, so we need to get more Admins in order to keep this place running; I think there is no hope in getting more Admins at this point (meaningful RfA reform will never happen – there will just never be real consensus for it; and, additionally, Adminship becomes less and less attractive to most veteran editors with each passing second!), and unbundling of the tools will ultimately be an inevitable necessity in order to just keep this place running. I wish you luck. But I'm skeptical there's any appetite to fix what's broken here. --IJBall (contribs • talk) 03:10, 6 October 2015 (UTC)

   Thank you for the best wishes. You are indeed correct that we are coming from opposite viewpoints. I personally find it very unlikely that we will ever succeed in getting the major admin tools unbundled, mainly because that would require many discussions/RfCs in different areas (blocking, protection, deletion, etc.) and therefore has a much lower chance of succeeding than a more focused and centralized fix. There also seems to be more general objection to tool unbundling than reforming the RfA process itself. --Biblioworm 14:00, 6 October 2015 (UTC)

4. Neutral. Biblioworm states that only 20% of admins are "active". I am interested to know if newly-promoted admins are active, or if 80% are just here to collect the hat. If most candidates just want a hat, the best solution may not be to distribute more hats, but rather to investigate the differences between active and semi-/in-active newly-promoted admins. Axl ¤ [Talk] 10:40, 5 October 2015 (UTC)

   @Axl: I have determined that most candidates who passed RfA this year (newly-promoted) would be considered active. The resysopped users are the ones who mostly remain inactive. I can give some numerical details, if you want them. --Biblioworm 15:18, 5 October 2015 (UTC)

   Thank you, I am indeed interested to see the numbers. Axl ¤ [Talk] 18:21, 5 October 2015 (UTC)

   @Axl: I will give the details shortly. --Biblioworm 14:00, 6 October 2015 (UTC)

   @Axl: I just finished calculating the details of the statistics. So, I found that of 15 users promoted via RfA this year, 13 of them (or about 87%) would be considered active admins. In contrast, of 22 resysopped users, only about 7 (or about 32%) would be considered active. So, it appears that users promoted via the process of RfA are more motivated to use the tools than users resysopped quickly as a matter of routine. Therefore, I think we can rather safely conclude that most people who bother to go through RfA are not there to just get the "hat". --Biblioworm 14:22, 6 October 2015 (UTC)

   Well, it means that those who passed RfA weren't there to get the hat. I suspect it may also really mean that RfA is doing a reasonable job of weeding out the hat seekers. --Stfg (talk) 15:02, 6 October 2015 (UTC)

   Thank you, Biblioworm. It is good to know that most newly-promoted admins actually use the tools. On the other hand, I am concerned that a recruitment drive or lowered standards may attract more hat collectors. Axl ¤ [Talk] 09:03, 7 October 2015 (UTC)

5. Need to offload work from admins either by unbundling, admin-bots that anyone can use (possible example: to delete a redirect with only 1 entry in its edit history to make way for a move), or admin-bot-implemented-de-facto-unbundling (if only select users can "trigger" the admin-bot, you've got de facto unbundling). This will free up the human admins to do things that require the community's review of their good judgement (which RfA is a form of) and those things that the WMF would require an RfA-like process to unbundle such as seeing or un-deleting deleted edits (I think we could get the WMF to go along with allowing selected editors to use an admin-bot that delete things and undo the deletion if they made a mistake and caught it within a short time, but that's likely going to be their limit on "non-admins seeing deleted edits"). davidwr/(talk)/(contribs) 02:54, 8 October 2015 (UTC)

Comments

But to say we are passing fewer admins per day than before is not the same thing as saying we don't have enough, so what metric indicates that we currently do not have enough active admins to take care of the workload? RO^(talk) 16:58, 4 October 2015 (UTC)

A quick look at how many inactive admins are being desysoped every month shows our ranks are dropping. I think if you look at who is making all the difficult decisions these days you will see the group is less diverse than it used to be. As the number of active admins drop we will see that the work still gets done, but by less people. Less people doing the same amount of work means less people using their judgement. The community will always benefit from more admins than less because a larger group is more capable of recognizing the mistakes of others. HighInBC 17:02, 4 October 2015 (UTC)

Read the links I recommended above. There, I discuss exactly what you're asking: I talk about why our number of active admins (not our pure statistical number) is insufficient relative to the size of this website. --Biblioworm 17:03, 4 October 2015 (UTC)

Okay. I'm trying to get up to speed on this, but how can we say we don't have enough admins without providing an estimate of how many admins we need? How many do we need? RO^(talk) 18:15, 4 October 2015 (UTC)

It's quite difficult to give an estimate of how many admins we would need, but we could always use more no matter how many we have. Right now, though, the situation is especially obvious because of the pitifully low amount of promotions via RfA. However, we could get an idea of when we have "enough" when the number of active admins reaches a point where individual users do not have to be online for hours at a time and perform hundreds or thousands of actions within a relatively short period. --Biblioworm 19:00, 4 October 2015 (UTC)

• A good standard is: aside from the occasional outlier, we have enough admins when the administrative backlog category remains consistently empty. We currently have nine backlogs at the moment. Esquivalience ^t 19:12, 4 October 2015 (UTC)

• Unsure Moved to oppose On one side, there are indeeed less admins around, art an average promotion of 1 and desysopping for inactivity of 5 to 10, per month. But, since decisions are supposed to be made evaluating the merits of a case, I doubt that there is much discretion/variance in admin actions. I suppose more than 90% of all admin actions would be the same if done by a different admin. We really do have a lack of admins who can combine good judgment with technical expertise, as seen in the terrible backlog at TfD where most of the work is now done by non-admins. RfCs are nowadays very rarely closed by admins. At AfD there are not many admins voting, they appear mostly just to push the delete button after all the votes are in. CSD doesn't seem to be backlogged too much, anything I tag is deleted within a day. Non-admins are carrying most of the workload now in many places, vandal-fighters, and page patrollers make the reports, and admins just spend time pushing the buttons after making a quick check, maintaining an encyclopedia is not really rocket-science. Still pondering... Kraxler (talk) 18:04, 4 October 2015 (UTC) I'm certain now. Kraxler (talk) 15:26, 6 October 2015 (UTC)
  • Sure, 90% of admin actions are simple tedious bullshit, 9% are difficult tedious bullshit, and 1% are genuinely challenging situations. The rarity of the last case is in my view an argument in favor of having more admins, so that a broad range of views will be available when the need arises. Opabinia regalis (talk) 18:45, 4 October 2015 (UTC)
  • It might sound simple (just press a button, right?), but even simple tasks can become tiring and difficult if there are many of them to do. For instance, when I edited another wiki, I seemed to be virtually the only person who bothered to deal with the long backlog of reference errors. Oftentimes, it was just as simple as adding "==Sources and Citations==<br/>{{reflist}}". It sounds easy. However, there was a few dozen articles on which to do this, and after a while the work became very tedious. Therefore, even for simple button-pressing tasks, we should have more admins so that the work will not be so long and tedious for any one person. --Biblioworm 19:00, 4 October 2015 (UTC)
• RfA produces the right number of admins and I maintain the process is not broken. The real problem is that too many editors can't win over enough support from !voters and too many of those promoted to admin return to content creation rather than work a backlog. Un-bundling the "block" button from the "delete" button is probably the way to solve the backlog issue although even that won't pass muster. It's a political problem. I as an editor have to be able to trust anyone I empower to potentially block me or delete content I create. That's going to be a high bar and I'm happy with the vetting that occurs at RfA. It's silly to suggest that we all need to collectively lower our expectations when we all have (in Biblioworm's words) an "irrational fear" of being railroaded by today's well-meaning candidate. Why aren't we actively harassing our current admins to pick up the mop they were handed? I will, however, agree that the 75% number for RfA is far higher than the standard for ARBCOM election and perhaps ought to be lowered to something between 60-66% to recognize we can't all be comfortable with every admin that gets promoted. Chris Troutman (talk) 19:27, 4 October 2015 (UTC)

  "too many of those promoted to admin return to content creation"? Can you blame them when most unsuccessful RfAs include a comment that the candidate is not active enough in content creation? The successful candidates want to avoid such criticism being piled on them as well after they become admins. As I noted in a recent successful RfA, the admin bit does not grant anybody more hours in the day. To do meta work - even if only a small amount - means that time has to be found for that by not doing something else. --Redrose64 (talk) 20:37, 4 October 2015 (UTC)

  As I have emphasized here, the tiny number of admins we're promoting now relative to past years (see the line graph), coupled with the continual existence of the backlogs despite the constant efforts of dedicated admins, should cast serious doubt on the idea that RfA is promoting a sufficient number of admins. And, actually, many of the admins promoted this year would be considered active, so we really don't need to "harass" them "to pick up the mop". Of course, I do agree with you that we need to trust the admins we elect. But the issue is that we have a very high bar for "trust". As I mentioned in the op-ed, almost no legislative processes or other real-life procedures (and even other bars for consensus on-wiki) have a bar so high as RfA's. I see that you agree with me on that issue, though. --Biblioworm 20:44, 4 October 2015 (UTC)

  We both agree the number of active admins is insufficient to our backlog issue. While you want to change the process in order to have more active admins, I say the process is how it is out of necessity. Why not address the issue of inactive admins? I supported three different proposals (here, here, and here) to strengthen the community's ability to hold admins accountable. Why not question why many Wikipedians that could be trusted choose not to become admins? What's the problem with swinging the mop at our dirty areas? If manning the noticeboards is such a chore then what you'd really want to change with RfA is selecting people who will work those backlogs without raising the ire of content creators. I think lowering the passing percentage and un-bundling with a sunset clause is the way to go. Chris Troutman (talk) 21:48, 4 October 2015 (UTC)

  (ec)RFA does not produce the right number of admins if it can't produce as many admins as leave. It doesn't produce the right number of admins if we only have 24 admins who started editing since January 2010. As for the comparison with Arbcom elections, its a different electoral system and the two are not comparable. ϢereSpielChequers 22:09, 4 October 2015 (UTC)

  Now that blows my mind. Opabinia regalis (talk) 04:05, 5 October 2015 (UTC)

• (edit conflict) It may be worth considering/reframing the problem in a way that pares it back a step: instead of "not enough admins", it's "not enough people performing the duties currently assigned only to the admin usergroup" or somesuch. I.e. we need admins to do admin stuff, not to be admins. Indeed, Biblioworm gets at the issue of admins who don't use the tools in the op-ed. If the point is RfA, reframing this way could be distracting, but if the point is to be as inclusive as possible at this stage (inclusive of approaches to the problem), it may be helpful. As Un-bundling the "block" button from the "delete" button would be one of the possible ways to address the problem, would it then ring truer, Chris troutman? — Rhododendrites ^talk \\ 21:56, 4 October 2015 (UTC)

• @Rhododendrites: Yes, the "we need admins to do admin stuff" statement is definitely the issue. I'd be happy to discuss how that might be accomplished. I just don't think changing RfA is the method. Chris Troutman (talk) 22:42, 4 October 2015 (UTC)
• This is what I've found to be the issue. I've looked at some of the same data Biblioworm has, and it's the RESYSOPS part of the equation that's the problem. From what I've looked at, I considered RESYSOPS to be a somewhat of a joke. Now, it's not true of all of them, but a substantial portion of the former Admins RESYSOPS'ing (and I'm talking 50% or more) seem to be doing pure "hat collecting" from what I've seen, and never use the tools when they get the bit back. Yes, some do; but many don't. That implies to me that the RESYSOPS procedure is actually what I consider to be broken. We shouldn't reflexively be giving the bit back to former Admins who then don't use the tools they are given – there needs to at least be a question included in this process: "Do you actually intend to use the tools we're giving you back much?!" If the answer is "No", then the response should be, "Well how about Rollback instead?!", not "Oh well – here you go!!" --IJBall (contribs • talk) 04:02, 5 October 2015 (UTC)

Who cares? It's no skin off your nose if someone has admin tools they're not actively using at the moment. Opabinia regalis (talk) 04:05, 5 October 2015 (UTC)

Yes, that's pretty much the perennial response to the POV I expressed above (note: I'm not the only one who shares it). To say that I find it uncompelling probably understates my reaction. Admin rights are a "big deal" in so many ways (there's just too much opportunity for malarkey...), and they shouldn't be handed out to anyone who isn't going to actively use them to improve the project. (The same could probably be said about Rollback and Pending Changes Reviewing as well, now that I think about it...) In any case, I know many Admins, especially, share your view on this. On my end, I couldn't disagree more. --IJBall (contribs • talk) 04:23, 5 October 2015 (UTC)

Unless you intend to show that resysopping is doing some damage, I don't know why you'd bring it up. Samsara 04:29, 5 October 2015 (UTC)

Because my standard for a "good idea" isn't "net neutral" – it's "net positive" for the project. (And, no – I'm not going to waste my time on a RfC for this – I already know the current Admin corps would shoot it down cold...) --IJBall (contribs • talk) 04:33, 5 October 2015 (UTC)

The case in point has already replied to you, though you apparently didn't notice. Samsara 04:38, 5 October 2015 (UTC)

@IJBall, Have you collected any statistics of admin activity by returning former admins? From what you are saying this is a glass half full/half empty situation where some would look at it and see an important part of the what has enabled our remaining admins to continue to offer a good service, and others look at it and wonder why certain people bothered when they don't seem to be using the tools much. We've had related discussions here in the past, one conclusion being that if you ignore people who aren't admins but made the admin stats due to certain anomalies, and the people whose active admin period was before December 2004, very very few admins get the tools and then don't use them. The other discussion was about the risk of long absent admins returning and making mistakes because they'd forgotten things or things had changed; The conclusion there if I remember right is that such scenarios rarely reach a level that requires a trout let alone a desysop. ϢereSpielChequers 12:02, 5 October 2015 (UTC)

Of course I can't remember where I saw it now, but I semi-recently stumbled upon a page that had the stats for recent RESYSOPS'ings, and about half had zero Admin actions since getting the bit back... It might be buried in a thread in the WT:RfA archives for all I know... --IJBall (contribs • talk) 14:04, 5 October 2015 (UTC)

Biblioworm just looked into this and found that 32% of resysopped editors became "active" admins. Axl ¤ [Talk] 13:56, 7 October 2015 (UTC)

• The possibility of unbundling the "mop" stuff from the "cop" stuff might allow more folks to do some of the tasks where there is a big backlog (page protection, moving articles, DYK queues, TFA, ITN, etc...) Montanabw^(talk) 01:23, 5 October 2015 (UTC)
• That said, I also agree with WereSpielChequers that we also don't want a situation where admins are so swamped that they only can do admin work, thus creating a separate cadre; one needs to keep their toe in the water to recognize the issues that content editors deal with on a daily basis. Montanabw^(talk) 01:23, 5 October 2015 (UTC)
• Regarding unbundling the tools, I'm going to say the same thing I said the last time this came up. I'd be wary of splitting up the core functions. In a few cases I've had to delete, protect, block, and look at deleted edits all for one incident (typically a page protect request that leads to a SPI). Each action has to be thought through. Having the tools to do only half the job means another admin will have to familiarize themselves with the situation. If we do unbundle these rights, we're going to have a "if all you have is a hammer, everything looks like a nail" environment. Every day, there are reports to AIV that are better served by page protection or page deletion than a block. Likewise, every day there are reports to RFPP that are better resolved by blocks or deletions. If the editors patrolling these boards only have access to one tool, we run the risk of implementing sub-optimal resolutions. --NeilN ^{talk to me} 04:00, 5 October 2015 (UTC)

• I've seen you make this argument before, Neil, and I have to say I consider it a weak rebuttal to the idea of unbundling (at least some of) the tools. Will unbundling, say, Page Protections solve every issue at WP:RfPP? No. Will it eliminate the need for Admins to do anything at RfPP? Obviously, no (for one thing, I'm pretty sure "unsalting" is likely to continue to be an Admin-only purview, though maybe I'm wrong...). And, yes, sure – some situation will require an Admin to come in a look at the more complex situations. But if 70%, 80%, 90% of the tasks at RfPP could handled by veteran editors after unbundling (and, in many cases, they'll likely be declining requests, as Admins have to do now), rather than Admins doing it, that would move the ball forward. (Then, if you unbundle more of the bit, then it's possible veteran editors with both Page Protector rights, and say Vandal Fighter rights, could handle some of the more complicated situations you mention...) Unbundling isn't about solving all issues at once – it's about incrementally moving some tasks to be primarily Admin-only actions to being primarily handled by trusted veteran editors, freeing up Admins to do other tasks (like deletions, or more complicated blockings) more... --IJBall (contribs • talk) 04:12, 5 October 2015 (UTC)

• I've seen repeated mentions of a "backlog at RFPP". Is this still a problem or is everyone repeating what everyone else says? I'd have less of an issue if these semi-admins were given good guidelines on what they could and could not do. They would also have to run through a mini-RFA. Too many veteran editors post inappropriate requests for protection or blocking to consider bestowing these tools upon request. --NeilN ^{talk to me} 04:22, 5 October 2015 (UTC)

• Proper guidelines (i.e. esp. qualifications for such a right) would absolutely be key if this is going to work. However, a mini-RfA seems like overkill to me for either Page Protector or Page Mover rights, though I suppose it could work if the focus of the "mini-RfA" was narrower. Personally, though, I think it could be granted like a "user right" – it's just that whoever would be granting a "Page Protector" user right would better come from an Admin like yourself that operates at WP:RfPP a lot. But, getting back to the original point – the key issue would be figuring out what would be needed for "qualifications" for the right. --IJBall (contribs • talk) 04:30, 5 October 2015 (UTC)

• Figuring out what is needed: a mini-RFA focused on examining the editor's history in the area. If they want the page protector right, examine their past protection requests. --NeilN ^{talk to me} 04:35, 5 October 2015 (UTC)

• Comment regarding unbundling: Personally I am a little wary of granting additional admin permissions to non-admins except when it comes to moves. I never really understood why we couldn't add a page mover permission. We already have a file mover permission. Why not have a page mover permission? Admins are required to move a page when there is existing revisions on the target. Essentially they are deleting the target and moving in one step. Why can't a trusted user be granted that permission? The requested move board always seems to have a backlog and having additional people with the ability to work on those requests can only be helpful. For every other admin function I feel the same way as NeilN. There is just too high a chance that multiple semi-admins would have to be involved in one situation to make it worthwhile. --Stabila711 (talk) 04:06, 5 October 2015 (UTC)

• Purely from a political standpoint, I see the chances of "rights unbunlding" going like so (in decreasing order of likely successful proposal): 1) Page Protections, 2) Page moves (but the pesky issue of Redirect deletions is a stumbling block), 3) Vandal fighter (there is still substantial resistance to granting even limited "blocking" rights to non-RfA-approved Admins), 4) Page (and Template) Deletions (basically, forget about it: from everything I've seen, the consensus is that significant Article Deletion rights require either a full RfA, or something like "RfA Lite", and most seem to be against the idea). So, the place to start with unbundling is probably on the Page Protection proposal... --IJBall (contribs • talk) 04:16, 5 October 2015 (UTC)

• I actually have more of an issue with page protection than page moving. Page moves are simple. There are only two answers, move or no move. Page protection has more outcomes that have to be taken into account. Length of time and type of protection being the two categories. Do we protect for a day, a week, a month, indef? Do we do semi, full, pending changes, or move? All of these have to be answered every time a page gets protected. Allowing non-admins to page protect things freely makes me really uneasy. There would have to be a few very important restrictions to make me comfortable with that. Disabling the ability to protect indefinitely being one of them. In fact, I would say any protect that last more than a week should be left to someone who has gone through RfA and has shown that they have the community's trust. These issues is why I put page moves first in my "probability" rankings. There is just less fuss to deal with when it comes to page moves. As to blocking and deletion I highly doubt those will ever be unbundled. There is just far too high a chance for abuse, even in a limited capacity, that granting them to even experienced users who have not gone through RfA would be a bad idea. --Stabila711 (talk) 05:33, 5 October 2015 (UTC)

• The problem with "Page Mover" rights gets back to the issue that it would require the ability to delete Redirects (and possibly Disambiguation pages – not sure about this second one...), and we're back to either, 1) any page deletion "right" requires "some kind of RfA", or 2) there needs to be a technical solution to this that would create a user right that could only delete Redirects (and Disambig. pages?) but not any other pages and apparently that could only be implemented on the WMF side of things (as I understand it). On the latter, I think people in the "know" seem to think the second point isn't as much of a technical stumbling block now as it was a few years ago. But until this issue can be resolved, I'm guessing a "Page Mover" rights proposal isn't going anywhere... --IJBall (contribs • talk) 14:00, 5 October 2015 (UTC)

• This is a problem with user attrition, not with any aspect of adminship or RfA. Wikipedia is not recruiting enough editors.

Samsara 04:19, 5 October 2015 (UTC)

• • If it was an attrition problem there would be some link between the number of active editors and the number of admins. I'd like us to be growing faster, and until the beginning of this year the number of editors saving over 100 edits in mainspace per month seemed to be in gentle decline. But by that measure we are now stable or even growing again. Of course there could be a time lag between that growth and an increase in RFAs, especially if the growth is from new editors as opposed to returning ones or long term ones becoming more active. But if RFA wasn't broken I would expect an increasing proportion of our active editors to be admins, instead we have the opposite, and that's despite some evidence that simply making someone an admin is good for editor retention as admins tend to stay longer than other editors. ϢereSpielChequers 10:06, 5 October 2015 (UTC)
• Support unbundling... seems to me a straight-forward way to deal with the problem often encountered in RfA ("candidate does not have enough experience in X"). We even ask numerous candidates what area they will use their admin tools in, which is almost doing this by proxy. I think this is a sensible way to start, decentralising some of the tools. --Tom (LT) (talk) 08:35, 5 October 2015 (UTC)

  Unbundling is difficult, as many situations require several of the tools. For example, the correct answer to vandalism on an article is sometimes block and sometimes protection. An editor who has access to only the block or only the protection button is likely to use their suboptimal tool instead of the correct one they don't have access to. Similarly, in order to decide whether an account needs to be indefinitely blocked, the viewdeleted user right is often needed. Deletion is the function that needs the others least, although having access to deletion and not blocking or protection is going to be a terrible waste of several people's time. (Often when you delete something, you either notice it has been repeatedly recreated and need the protection button to WP:SALT the page. And of course, when you decide to delete all of an editor's attack pages / ban evading contributions, you also decide that editor has to be blocked). I don't understand where the popularity of unbundling comes from -- I strongly oppose any unbundling that separates block/protect/delete. —Kusma (t·c) 13:30, 5 October 2015 (UTC)

• Looking at those figures again, I'm slightly puzzled at the interpretation. It seems that unbundling coincides with fewer admin promotions. Did anybody not foresee this, and is it really a problem? Samsara 13:31, 5 October 2015 (UTC)

• I think it's a case of unintended consequences. Even in simple systems, unintended consequences are common. Wikipedia is far from being a simple system. That's one of the problems I've had with every RfA reform process that's been put forward; no anticipation of potential unintended consequences and how to mitigate them if they are damaging. --Hammersoft (talk) 19:40, 5 October 2015 (UTC)

Wrong question being asked

With every ounce of respect to Biblioworm; This RfC asks the wrong question(s).

• Anchoring bias:The statistics we are asked to review state opinions have been removed. This is false. The very section linked to starts off with "We need more admins". That's an opinion. Further, the second paragraph concludes that "we don't have enough admins" in bold. We're then asked to voice our opinion on whether we have enough admins. This is anchoring bias. Sorry.

• Wrong analysis: Just looking at declining admin counts doesn't tell the whole story. We have to look deeper than that. For just one example; I took a look at admin actions for the time period of August 5 to October 4, a time period of two months. I compared admin actions for that period to the same period from five years ago, 2010. There's a few things that leapt out at me; (1) the total number of admin actions went down slightly by 1.7%. (2) Bot admin actions went up a whopping 169%. (3) The average number of admin actions per all admins with at least 1 admin action in the time period went up a modest 10%. (4) The average number of admin actions per top 30 most active admins went down 19%. This data points to one possible refutation; we do not need more admins, as our burden, per admin, has remained more or less the same for the last five years and we've reduced the burden on our most active admins. I say possible because considerable more research needs to be done. We can not conclude anything based off of just WereSpielChequers's excellent chart. The data from that is insufficient to support any conclusion. The data I've provided above isn't sufficient either, but it shows at least that we can not conclude, based on statistics provided so far, that we need any more admins than we did five years ago.

• Other interesting bits: Bots perform 76% of restores, and 60% of blocks. In total, these categories of admin actions comprise 40% of all admin actions (including bots). There's another big, big hitter in the room though; deletions. Deletions comprise 55% of all admin actions, yet bots conduct only 6% of all deletions. A possible conclusion from this is that rather than looking at a process which seems (one possible conclusion) to be producing enough admins to maintain the status quo for the last five years, let's look at ways in which bots can be utilized to perform deletions. Do that, and we reduce the burden on administrators. Sidebar: I took a look at the top 30 admins for the last two months. A curious data point evolved; the average date these admins began editing was just over nine years ago. It doesn't mean anything in isolation, but I thought it interesting.

Conclusion: I'm chipping at the iceberg with a toothpick here. However, just a quick analysis of some figures compared to five years ago shows the anchoring bias statement "we need more admins" to be provably false. The issues are considerably more complex than just a chart, which is but one data point to pull from. We can look for other ways in which to reduce admin burden (which has remained close to static for the last five years). Note that I have not spoken at all regarding the declining participation issues we face. A longer term plan for solving our administrative issues has to address this, and produce plans for maintaining the project after we're all gone. --Hammersoft (talk) 15:07, 5 October 2015 (UTC)

@Hammersoft: The evidence is very clear that we need more admins. I have every statistic backing me up. Since your comment is very long, I will take some time to read through it carefully and give a point-by-point response to it. --Biblioworm 15:21, 5 October 2015 (UTC)

• Some other interesting data points. I took at look at the same two month period in 2005. The number of actions per admin has remained effectively static for 10 years; it dropped 1% from 2005 to 2015. The number of actions per top 30 most active admins went up 29% from 2005 to 2015. Whatever our means of handling admin actions, the burden on administrators has remained static for 10 years. --Hammersoft (talk) 15:37, 5 October 2015 (UTC)

Here is my reply to Hammersoft's data:

1. I will first answer your accusation of "anchoring bias". First of all, when I said that I had "removed all opinion", I believe I clarified even before this that I had removed all opinion about the causes of, and solutions to, the asserted problem. I presented the statistical evidence, and made a logical conclusion, which is namely: we need more admins. Secondly, I asked voters to read the section on the user page (and, if they wanted, the op-ed) so that they could be informed about my side of the story, and to head off any automatic oppose votes by anyone who might not be aware of the evidence. I did not say, "You must read my composition, and you must support!"

2. The data presented above is full of problems. Indeed, we do need to go deeper than simple data which shows that admin promotion numbers are declining. And I did. Secondly, you contradict yourself and reach illogical conclusions.

Now, I will address your numbered points: (1) and (4) – This relatively short-term decrease (2010–2015) proves absolutely nothing. First of all, I note that for a bot admin action increase of 169% noted in (2), the decrease in the number of admin actions is not comparatively significant. And most importantly, there are contradictions in your data. The general decrease in admin actions (1.7%) mentioned in point one is contradicted by (3), which shows a 10% increase per admin. There is a difference of 9% here. Next, the shorter-term data in (4) is contradicted by the longer-term data that you presented . The long-term data is more important by the short-term data. So, the shorter-term decrease of 19% for the top 30 active admins is contradicted and completely canceled out by the longer-term data you presented here. Over the longer term (2005–2015), there has been a 29% increase in admin actions for the top 30 active admins. The shorter-term data is completely canceled out with 10% to spare by the more important longer-term data. Ultimately, this data does not at all lead to one possible refutation (namely, that we do not need more admins). Your conclusion is logically unsound and is not even supported by this data. It does not destroy the fact that despite the efforts of dedicated admins, our backlogs continue to exist. It does not destroy the fact that we are losing admins faster than we are gaining them. It does not destroy the data that WereSpielChequers presented here.

3. Bots cannot do everything. We are concerning ourselves with the actions that humans must do. We cannot rely on AI for everything. First of all, when you say that bots perform 60% of blocks, you are almost certainly talking about open proxies. I imagine that would be possible to set up via coding. But humans will always have to use intelligence when closing AfDs; humans will always have to use intelligence when deleting most CSDs; humans will always have to use intelligence when deciding to (un)protect a page; humans will always have to use intelligence when deciding to (un)block a user; humans will always have to use intelligence when closing noticeboard discussions; and the list goes on and on. You get the point. Bots cannot do everything. If we try it, our quality is certain to decline, because by their very nature bots are fundamentally less intelligent than humans. They get their limited intelligence from us; we can be pragmatic, while they rely on a set of rules.

Conclusion: To the contrary, I believe the assertion that we do not need more admins is demonstrably false. It is supported by data. If this is too long for some people to read completely, read point two of this reply; it is the most important, as it shows why the data presented is in fact meaningless, does nothing to refute my position, and should not be used as a rationale for opposing this RfC. --Biblioworm 16:57, 5 October 2015 (UTC)

• Biblioworm; Please have a read through Anchoring. If it's too long to read, just read the first sentence. Now, in presenting this RfC the very first sentence in your proposal links to your op-ed "We need more admins". You then said "I strongly recommend that you read this (and the "Stats" subsection) before voting here.". You then asked us to answer the question of whether "RfA is not producing enough admins". The very first pieces of information you give us are that we need more admins. Then, you 'strongly' recommend we read your assertion that we need more admins. You then ask us if we need more admins? This is a classic case of anchoring bias.
• As to contradictions, please read the two statements and understand there is no contradiction: I said the total number of admin actions went down, but the total burden per admin went up 10%. There is a serious difference in the meanings; they are most emphatically not the same and there is no contradiction.
• As to backlogs existing; they will always exist. To conclude that we need more admins because backlogs exist is a non-starter. To begin to verify the need, we'd need to know what backlogs existed over time. Do you have data for that?
• As to bots doing everything; I never suggested that. I suggested we look at ways that bots can be used to perform deletions. Given that only 6% of deletions are done by bots, and deletions are the majority of what admins do, there is potentially fertile ground for reducing admin burden there. I think it's worth looking into.
• In abstract; I think the question being asked is the wrong question because it looks at one reliable data set in isolation. The data I presented shows the average burden per admin that does anything (which actually favors your interpretation by the way) was effectively unchanged from five years ago and unchanged from 10 years ago. Also, the question of whether we have enough admins has been a perennial discussion; 2005, 2007, 2010, 2014. I'm sure it's been discussed several other times. --Hammersoft (talk) 17:31, 5 October 2015 (UTC)

• As for anchoring, that's the voter's problem, not mine. They are not obligated to support, although I will defend my position. The job of any person making a case for something is to present evidence and reach conclusions using the evidence. Nothing would ever get done if everyone trying to prove something was careful to avoid this thing called "anchoring bias". Secondly, my point was that your data is filled with different sorts of conflicted data over different time periods. There are some increases, some decreases. It becomes a confusing incoherent jumble of factoids. Ultimately, the minor details matter very little. The big picture is that over ten years, there has been only a 1% decline in admin actions. This is not anything significant and cannot be used as evidence of any sort; it does nothing to prove that we have enough admins. It could mean anything. As for the bots, I don't think we can do much at all in the way of deletions. Deletions require discretion. Even seemingly simple CSDs, such as G7, require an admin to look at the tag. After all, imagine the damage vandals could do if they could get article deleted by some unintelligent bot by simply adding the G7 tag to articles. And finally, in conclusion, I would also note that we are trying to reduce the backlogs by obtaining more active admins. Simply showing that the number of actions/backlogs/other things have been stable for x years really means nothing. --Biblioworm 18:57, 5 October 2015 (UTC)

• Re: Anchoring being the voter's problem. If you wanted an unbiased appraisal from the community about whether RfA is producing enough admins, then the means by which you started this RfC to assess that have failed you. That's why I was pointing out the anchoring bias. If this is (as I think it is) part of a building effort towards RfA reform, then you're starting out on shaky ground. This can't be used as a basis for anything. You are welcome to disagree. Our discussion is devolving, so I've nothing to say on the other points. I wanted to post further about the anchoring only to highlight that using it to support anything is problematic. Anyway, all the best. --Hammersoft (talk) 19:56, 5 October 2015 (UTC)

• I should also note that data framing is very important, so you must be cautious whenever anyone says that there was some massive decrease (or increase) in admin actions. For instance, using this image, I could truthfully say, "Between June 2012 and February 2015, the number of admin action rose by almost 700%." (While ignoring that it was a very brief spike that was responsible for it. I also forgot to mention above that admin actions alone is not an important statistic. There could be many reasons for a decrease in admin actions other than a sufficiency of admins. The important part is the backlogs: are the backlogs improving in any significant way as a result of these admin actions? --Biblioworm 19:08, 5 October 2015 (UTC)
  • Except that Hammersoft made an appropriate year-on-year comparison, not June to February like you're implying. Samsara 20:13, 5 October 2015 (UTC)
    • My point was that the time span used is very important. For instance, the long-term data is more important than the short-term data for anything. Scientists look at long-term trends rather than short-term fluctuations when writing research papers. So, the long-term 10-year data since 2005 actually shows an increase in admin actions as opposed to the shorter-term data, which shows a decrease that is completely canceled out with 10 percent to spare by the more important long-term data. So, by your logic (e.g., number of admin actions correlates with our need of them), the long-term data supports my position. --Biblioworm 20:25, 5 October 2015 (UTC)
• To clarify, I forgot to mention in my previous comment (20:25, 5 October 2015) that I was referencing the long-term data for the top 30 active admins. I will write a final, organized summary of my objections to the evidence for those who do not want to read through the lengthy and rather messy discussion above. Two data points show an increase in workload per admin. These two data points are the long-term data for the top 30 active admins since 2005 (to hence be referenced as the "top 30" data) and the short-term data since 2010, which shows the average workload per admin. The long-term top 30 data shows an average 29% increase in admin actions per each admin in the top 30 activity group. The short-term data for shows an average of a 10% increase in admin actions per individual admin. Therefore, these two data points actually show an increase in burden for each individual administrator. This actually proves my point. This data is conflicted by the short-term data since 2010 for the top 30 group, which shows a 19% decrease in admin actions per individual admin in this group. However, this data point is shorter-term (and as I mentioned above, long-term data is almost always preferred over short-term data in studies) and fails by a healthy 10% to cancel out the longer-term data for the top 30 group. And even if it did, the other short-term data point showed that every average admin (not just those in the top 30 group) performs 10% more admin actions than they did in 2010. This point covers a wider spectrum of users and is therefore more relevant. Finally, even though there has been a 1% decrease simply in the amount of admin actions performed, this does not prove anything. First of all, it is an extremely negligible change, and finally, we should not be as concerned about the actual amount of admin actions performed as we should be concerned about the individual burden per admin and how the backlogs are affected as a result of these actions. So, overall, the facts are still in favor of Wikipedia lacking sufficient admins, and this data does nothing to prove otherwise. (Some points are actually in favor of the position opposite of that it trying to prove.) --Biblioworm 13:49, 6 October 2015 (UTC)

• Actually the proper conclusion is that the burden on individual administrators has increased, and effort should be placed in the most efficient place to reduce the burden. Given the amount of automated admin actions in that time that has otherwise reduced the growth of human-admin actions, the logical thing to do would be to explore how more automation can reduce the need for the human component. 'Throwing more people at it' is rarely a good solution to any problem. Work smarter not harder. Only in death does duty end (talk) 15:07, 6 October 2015 (UTC)

Yes. Really, we already know "more admins" is not the answer. If it were, the present admins would be actually doing the work. It is a rather useless idea to make, say, 100 more admins today in the hopes that a handful of them might do the work. Address the work, or address the lack of interest by admins to do the work, first. Alanscottwalker (talk) 15:24, 6 October 2015 (UTC)

• Alan, as I mentioned above, the facts (see my reply to Axl) show that over 80% of users who pass RfA this year would be considered active. I seem to recall doing this for last year at one point, as well, and the results were similar. The data certainly seems to support the idea that RfA motivates most users who pass to use the tools. AdminStats shows that we have over a hundred admins that perform many hundreds or thousands of actions within a month or two. Is this dedicated work, or isn't it? And I addressed the possibility of bringing back inactive admins in my op-ed. I pointed out that many became inactive for reasons beyond our control, and might not care or even notice if we beg them to come back. In response to your comment on automation, Only in death, I discussed just recently how humans will always have to perform many tasks here. Almost all deletions, protections, and blocks (other than open proxies) require discretion. It's a plain fantasy to ever think that the idea of a bot that handles most deletes/protects/blocks would ever be accepted by the community. And Hammersoft's data is rather curious in that the individual burden per admin has generally increased despite a 169% increase in bot admin actions. What would we do if simply making more bots gave the same result? It apparently hasn't made any considerable impact in the past... --Biblioworm 17:05, 6 October 2015 (UTC)

So your reading of statistics is that we will appoint 4 in 5, who do the work, and over time that will become 1000 to 1 who do not do the work. And you get the temporary 4 out of 5 who do the work in the present system, but there is certainly no reason to think we will get that good a temporary yield, under another system. Alanscottwalker (talk) 18:51, 6 October 2015 (UTC)

Proposals for Modifying the Promotion Threshold

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

---

Assuming, as I do, that Biblioworm is correct that Wikipedia is not promoting enough new admins to replace those who become inactive over time, the community must consider several proposals for how to promote more admins. These are:

1. Do nothing and let Wikipedia die. I have no objection to letting Wikipedia die, but assuming that other people do not want this outcome, please consider the following alternative proposals.
2. Reduce the promotion threshold for all RFAs. I know, WP:PEREN. But see WP:CCC and consider the change in circumstances from 400+ promotions in 2017 to only about 20 for 2015.
3. Reduce the promotion threshold for some RFAs. One way to accomplish this is to say as follows: If a candidate's RFA would pass by the existing rules, then it passes under the new rules with no delay. However, if a candidate gets between 50 and 75 percent support, then the RFA result is placed "on hold." The RFA will then pass if fewer than three (3) new admins (who had not previously been admins) are appointed by RFA in the next 30 days. Exception: an RFA will be moved from "on hold" to "failed" if the candidate is subjected to a legitimate block of their user account while the RFA is on hold (even if the block is only for a few hours).
4. Same as the preceding option, except that a runoff voting method can be used. Thus, if in a defined period (say, a calendar month), an RFA is placed "on hold" and fewer than three candidates passed by the normal 75 percent heuristic, the next best candidates will be promoted from the "on hold" group, from the highest percentage down to the next highest percentage pass rate, until at least three admins are selected for the month. This prevents the absurdity of Option 3 above whereby one candidate with 51 percent support gets the sysoop flag while another candidate who ran a month later with a 70 percent support rate won't get the flag. Or to be more precise, it allows the absurdity to happen but only under different circumstances.

The point of Options 3 and 4 is to challenge the community to ask, if candidate X with 68 percent support is not good enough for you, do you have somebody better to put forward? And if you don't, would you prefer to let the project die as per Option 1 above? Agent 73124 (talk) 02:31, 8 October 2015 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

While using an undisclosed alternate account is not allowed in internal discussion, I suppose there is nothing to stop people from responding to this post if they want. The poster can participate if they log into their regular account. HighInBC 02:38, 8 October 2015 (UTC)