Wikipedia talk:Good article nominations#New function for GA bot

Recently I decided I wanted to try for the first time to get an article to GA status. Reading through the criteria and surrounding material and discussions and things, I started to feel confused about something I remain confused about even after attempting an article submission. Basically, I've read, both in the archives of this talk page and in the "Reviewing good articles" guide, both the perspective that "there is no minimum length for GAs" and the perspective that "some articles can never practically be GAs because the needed sources don't exist". "Reviewing good articles" essentially says both, in fact:

{{Blockquote|text=The good article criteria are achievable in almost any article…}}

and

{{Blockquote|text=Not every article can be a Good article. If the references to improve an article to Good article standards simply do not exist, then you should not overlook that part of the criteria.}}

If "the good article criteria are achievable in almost any article", to me that intrinsically implies that length alone shouldn't be a factor because many articles will remain fairly short even if they cover the available sources comprehensively (simply because many article subjects are too obscure to have many sources available). In turn, I would think that implies that the criterion of broadness should be evaluated based on the available sources—like, what the "main topics" are would be determined by what the available sources treat as the main topics, what it means to cover them well should be evaluated based on how much detail the sources go into, etc. That way, just about any article would be able to pass the broadness criterion as long as it really did cover the available sources well, whether there are many or few.

If "not every article can be a Good article" because the article should be failed "if the references to improve [it] to Good article standards simply do not exist," I would think that implies instead that the criteria for broadness would be based on something else than the available sources—something I think would, in practice, have to involve some kind of minimum article length or level of detail independent of the available sources (e.g. a list of required sections for the topic area and a standard of minimum depth each must go into for the article to be considered sufficiently broad). With this approach a large majority of articles would not and could not ever be GAs—only those articles with a large, wide ranging pool of available sources, which I would say is a select few based on the many times I've clicked the "Random article" link.

These perspectives seem to me to directly conflict, and yet they apparently coexist. Even though I've seen it said in many different places that "there is no minimum article length for GAs," most of the articles currently up for nomination do seem more on the side of satisfying the latter criterion, and the article I tried submitting was quickly failed on that same basis (lacking the needed sources to ever practically be a GA). On the other hand, I know there are some very short GAs, although I've also seen people express disapproval of that too, and suggestions like "stub-length articles should never be GAs" etc. Still, people do often say "there is no minimum length for GAs" and of course the GA criteria don't directly say there is. As someone totally new to the GA process, I feel deeply confused about this and unsure how to navigate it; I've worked on several articles that I feel are quite polished or at least close to, where I've gone to pretty great lengths to try to ensure that I've considered the vast majority of relevant sources anyone could reasonably find, and yet the resulting article is shorter than what it appears the average GA is simply because the subject just hasn't received that much in-depth coverage in RS (e.g. sometimes I like to press "Random article" and then just try to take whatever I get as far it can go). I don't want to waste anyone's time by submitting articles like that if there really is wide agreement that such articles simply can never be GAs, but I've seen people say that GAs can theoretically be of any length so many times that if there really is also wide agreement that many articles are just ruled out by having few available sources, people are finding some way to harmonize those perspectives that hasn't occurred to me. 🍉◜⠢◞ↂ🄜𝚎sₒᶜa𝚛🅟ම𛱘‎🥑《 𔑪‎talk〗⇤ 09:22, 7 June 2025 (UTC)

:You raise a good point and I think the project should probably have a discussion about rewording this criteria, if it's confusing and/or contradictory. I'm personally of the opinion that stub-class articles shouldn't ever be nominated for GA, but I'm not sure how much higher I'd set the lower bound for GA length. It would be interesting to know the statistics on our current good articles, so we could see what the average length is and what the shortest and longest articles are. --Grnrchst (talk) 09:46, 7 June 2025 (UTC)

::Statistics on the longest and shortest can be found at Wikipedia:Database reports/Good articles by size:

::*Ohio State Route 778 is the shortest by words (179). I believe WP:OVERSECTION applies.

::*2024 Men's T20 World Cup Super 8 stage is identified as the shortest by prose size (1020). The article is a list of cricket matches.

::*Fidel Castro is the longest by words (19267) and prose size (124439). It is currently tagged as being too long.

::*Pilot (Devious Maids) is the median by prose size (11272). It has 1885 words.

::*The average GA has a prose size of 15418 and 2511 words. A few articles are around this, such as Arthropleura and Maryland Route 313.

::The shortest articles seem to be covering roads, Olympics articles and some sports players. A lot of oversectioning.

::Rollinginhisgrave (talk | contributions) 10:10, 7 June 2025 (UTC)

:::Haha, I was starting into the process of basically writing exactly this but you got there first, thanks for doing this ^^ The article I tried submitting is 646 words/3,847 prose size, so it's on the short side for a GA candidate but there are about 2,000 shorter existing GAs (which is part of why I thought it might be acceptable). 🍉◜⠢◞ↂ🄜𝚎sₒᶜa𝚛🅟ම𛱘‎🥑《 𔑪‎talk〗⇤ 10:20, 7 June 2025 (UTC)

:::Seems like we have a few dozen stub-length articles at GA. I'm surprised that Ohio road made the cut, given it's entirely cited to state government reports without any secondary sourcing. GA criteria aside, I'm not sure it even meets notability standards. --Grnrchst (talk) 10:32, 7 June 2025 (UTC)

:::I've just gone through the articles that are >15,000 words long, as per our article size guidelines they should be trimmed at this length. I do think most of these articles are of a large enough scope to justify a >9,000 length, but at GA they really should be following the style guidelines. --Grnrchst (talk) 11:05, 7 June 2025 (UTC)

:::{{yo|Rollinginhisgrave}} The Talk:2024 Men's T20 World Cup Super 8 stage/GA1 is pretty clearly AI generated. Worth giving a once over?--Launchballer 09:36, 11 June 2025 (UTC)

::::Geez, and a sockpuppet to boot. Talk:Golden Temple/GA1 needs to be reversed as well. CMD (talk) 10:01, 11 June 2025 (UTC)

:For reference, the GA nominee in question is: Mr. Nutz: Hoppin' Mad. I think the review is lacking and strays from the criteria as written. The review focusses on three points: length, coverage, and media presentation. The article at 646 words/3847 bytes is sufficiently long to be considered for GA. It is not a perma-stub. The broad criterion is determined by available sources, not by editor expectations. If a section cannot be written from sources, then whether or not you might expect one is irrelevant. I'm not sure whether 'gameplay' might be excepted from inline citations like WP:PLOTREF which allows editors to write a plot section without citing sources because the work is the source itself. You can use that exception to write a fuller 'story' (maybe 'synopsis') section, at least. Any video-game oriented editors might be better positioned to provide input on this particular point. Media should be presented if possible. An article may attain GA criteria without a single piece of visual media if none is available. Considering the article length, without using a gallery, the article is already densely packed with the media presented. The reviewer might benefit from reading the essay WP:GACRNOT (note that the GACR aren't a guideline or policy) particularly regarding the broadness and appropriately illustrated criteria. Mr rnddude (talk) 10:19, 7 June 2025 (UTC)

::I see that there are 12 reviews for this game. Generally, per MOS:VG rules, each review should also be present in the prose of the "Reception" section, and not only in the reception table. Vacant0 (talk • contribs) 10:40, 7 June 2025 (UTC)

:::They are, right? Sometimes in brief but I think I made sure to give each a summary—is there one I left out? If several reviews said essentially the same points I would group them together and note that, but that's also recommended by the MOS. 🍉◜⠢◞ↂ🄜𝚎sₒᶜa𝚛🅟ම𛱘‎🥑《 𔑪‎talk〗⇤ 10:44, 7 June 2025 (UTC)

::::The prose appears to be missing: Mega Fun, Amiga Joker, Mega Zone, and Games World. Mr rnddude (talk) 11:13, 7 June 2025 (UTC)

:::::Oh! Thank you for checking, I'll incorporate those. I wrote that section kind of a long time ago now, so I don't quite remember what was going on at the time; I definitely intended to cover them all, but maybe I was working on that section when I was too sleepy or something and lost track of what I'd covered partway through. 🍉◜⠢◞ↂ🄜𝚎sₒᶜa𝚛🅟ම𛱘‎🥑《 𔑪‎talk〗⇤ 11:26, 7 June 2025 (UTC)

::Oh, thanks for the advice re WP:PLOTREF, I could definitely flesh out the plot summary on that basis. I kind of remember coming across that at some point now that you mention it, but I don't think I had seen it at the time I was writing that story section, and at this point I had kind of forgotten about it again. The WikiProject Video Games MOS section on gameplay says that the gameplay section should be sourced as typical, though. 🍉◜⠢◞ↂ🄜𝚎sₒᶜa𝚛🅟ම𛱘‎🥑《 𔑪‎talk〗⇤ 10:41, 7 June 2025 (UTC)

::Something else to note, maybe…WP:GACRNOT implies that the Video Games MOS should be disregarded entirely for purposes of evaluating game articles for GA status, but I get the impression based on my experience so far that many regular game article assessors would howl at that. 🍉◜⠢◞ↂ🄜𝚎sₒᶜa𝚛🅟ම𛱘‎🥑《 𔑪‎talk〗⇤ 11:07, 7 June 2025 (UTC)

:::The GACR only mandate a very small amount of compliance with the MOS, see GA1b. It isn't prohibited to ask for improvements concerning other sections of MOS, but it isn't a requirement to follow them. Mr rnddude (talk) 11:13, 7 June 2025 (UTC)

::I took a look at Wikipedia:Good articles/Video games (1990-1994) and found a few GAs of similar length and depth as the nomination: Captain Novolin, Navy SEALs (video game) (sub-700 words), Painter (video game) (~ 670 words long and attained GA in May 2025), and Somari among several others that are 800-900 words long. These were brief checks, as the point was to establish precedent for shorter articles in the same topic area. A few of these shorter articles combine story and gameplay in a single section, such as The Simpsons: Bart's House of Weirdness. I noticed that they were consistent in sourcing gameplay and the MOS link above confirmed that this is necessary, this may make combining the two sections difficult or undesirable. Mr rnddude (talk) 11:13, 7 June 2025 (UTC)

:::Thank you, those are great points of reference. 🍉◜⠢◞ↂ🄜𝚎sₒᶜa𝚛🅟ම𛱘‎🥑《 𔑪‎talk〗⇤ 11:28, 7 June 2025 (UTC)

:There are differing opinions as this is somewhat of an art rather than a science. Regarding "I would think that implies that the criterion of broadness should be evaluated based on the available sources", that is I believe generally true, but there are exceptions where new sources might be expected that don't exist. For example, media which is not yet released. CMD (talk) 11:56, 7 June 2025 (UTC)

::I'm interested in how you see GA's broadness requirement compares with the broadness standard discussed in Wikipedia:Articles for deletion/Ira Brad Matetsky (2nd nomination)'s close. Rollinginhisgrave (talk | contributions) 12:27, 7 June 2025 (UTC)

:::The GACR are content curation considerations, and subordinate to policies and guidelines such as those relating to notability and BLPs. We've had GAs deleted/redirected before. CMD (talk) 12:46, 7 June 2025 (UTC)

:My position is that if there's enough sourcing to confidently establish notability, then that's enough to create a GA. If there's not enough to create a GA, then the subject isn't notable. Thebiguglyalien (talk) 🛸 15:57, 7 June 2025 (UTC)

::We have notability criteria that are not based on depth of coverage in secondary sourcing. But I would avoid nominating or passing an article whose notability is based only on those criteria and that does not also have GNG-worthy sourcing. —David Eppstein (talk) 18:52, 7 June 2025 (UTC)

::I was thinking this, but then I remembered that GA criteria includes that the article is broad enough to comprehensively cover the topic. We can have enough sources to establish a topic's notability, without having suitable RS for certain levels of detail that would be needed to meet the broadness criteria. That's where "too few sources" can come into play. Kingsif (talk) 20:40, 15 June 2025 (UTC)

:::Well if the notability-proving sources don't care enough to cover it then I don't think it would ever be a piece of information that would be required. Comprehensiveness is FA, not GA. Broadness is not equivalent. PARAKANYAA (talk) 20:02, 19 June 2025 (UTC)

::::In the sense of the topic coverage would be incomplete, not the FA level of comprehensiveness. Kingsif (talk) 14:18, 20 June 2025 (UTC)

Is it possible for this to be deleted and added back to the nominations needing review list? Pretty obvious at this point that the reviewer, NAUME GOU, is not going to do this review/concerns of disruptive editing have been raised numerous times at their user talk. Sarsenet•he/they•(talk) 05:42, 13 June 2025 (UTC)

:I agree that the reviewer seems far too inexpirenced for this. He has also been pinged and not responded, despite have edited since. I know we are supposed to assume good faith, but their edit history and inexpirence do not fill me with enough confidence to ping them again before taking this step (which I normally would). I have nominated the review page for deletion, and when that is done will change the nomination template to reset the nomination and reinsert the nomination in the queue. SSSB (talk) 16:47, 14 June 2025 (UTC)

::{{done}} SSSB (talk) 17:19, 14 June 2025 (UTC)

:::Thanks! courtesy ping @DarkKholi, just letting you know this was taken care of and sent back to the queue. Sarsenet•he/they•(talk) 05:14, 17 June 2025 (UTC)

::::Thank you both, @SSSB and @Sarsenet DarkKholi (talk) 11:35, 17 June 2025 (UTC)

Wikipedia:Good article reassessment/Addition/1

{{ping|jacobolus}} has posted the following [https://en.wikipedia.org/w/index.php?title=Wikipedia%3AGood_article_reassessment%2FAddition%2F1&diff=1295630757&oldid=1295630481]: "Judging from their behavior, people starting these processes generally seem not to realize that they have more collegial alternatives available, not involving a short time limit or an implicit threat, so it's valuable to clearly explain it to make sure they get the message." This was after they posted [https://en.wikipedia.org/w/index.php?title=Wikipedia%3AGood_article_reassessment%2FAddition%2F1&diff=1295628564&oldid=1295627983] "Leaving the "do this or I'll take the green badge away" threats until after you run into page where other editors are either (a) completely disengaged and unwilling to discuss concrete and specific concerns, or (b) outright refuse to make changes that seem clearly required by good article criteria."

I have asked [https://en.wikipedia.org/w/index.php?title=Wikipedia%3AGood_article_reassessment%2FAddition%2F1&diff=1295629272&oldid=1295628564 here] that comments about the GAR process happen on this page, and comments about my conduct happen here or on the appropriate noticeboard. The comments above, and others at the Addition GAR, do not give me confidence that this will happen, so I'm initiating it.

I hope there are some ideas on how to avoid lengthy discussions about GAR process on individual GAR pages. I also do not think jacobolus's comments are creating a welcoming, collaborative environment to make improvements to the article: instead, it brings an adversarial relationship between the reviewers and the editors making changes. I hope other editors can comment on the interaction at the Addition GAR here and suggest new paths for a better environment so that the GAR can be reserved for making improvements to the article. Z1720 (talk) 00:52, 15 June 2025 (UTC)

:Mountain out of a molehill, from both sides. Move along. ~~ AirshipJungleman29 (talk) 00:59, 15 June 2025 (UTC)

:I just don't understand why some editors like to start a formal process, usually with a vague handwave of a criticism, something like "this article is too long" or "this article includes uncited statements" as their first step whenever they find an article they have problems with. I urge anyone tempted to ever kick off one of these processes to always instead lead by starting a talk page discussion with as complete and specific a list of criticisms as they can muster, actually putting in the work to engage positively with other Wikipedians who are watching the page. It's even better if they make some proactive effort to fix some of those issues, asking for help in the case where the task seems too large for one person to quickly knock down. If a local talk page discussion doesn't immediately work, linking the discussion somewhere like WT:WPM can be a good second step, to bring more eyeballs along.

:The biggest problem with processes like this is that they are dramatically asymmetrical: it takes very little effort to start the process as a critic, but potentially unbounded amounts of work to satisfy the critic's criteria, especially if further critics are attracted to a discussion, since those criteria are typically vague and underspecified, and usually revolve more around ticking off boxes on a checklist than doing a careful editorial review of articles. The critic has effectively no skin in the game, and can make many quick drive-by criticisms and move on without consequence, whereas anyone interested in keeping the little green or gold badges (admittedly a kind of pointless goal) is more or less told to "fix this or else ...".

: The same advice applies to many other kinds of formal processes here. For example, it's nearly always better to try an article talk page conversation, possibly followed by a user talk page conversation, to resolve a revert war instead of immediately escalating to some noticeboard or other. The formal processes work better if treated as a back-up to basic conversation when it fails, rather than a basic frame bounding conversations from the start.

:Thanks @Z1720 for making a more specific list of problems you found. That is useful work. The GAR wrapper for it seems counterproductive and unnecessary though, and in the future I recommend leading with the specific list, as an ordinary discussion topic. –jacobolus (t) 02:59, 15 June 2025 (UTC)

:One editor has just recently been topic banned for doing exactly this, and I'm quite certain that the only reason Jacobolus didn't get a topic ban as well was because of dumb luck, that the one instance where they weren't heavily involved in the disruption was the one that ended with ANI. Jacobolus, if you think that non-GA articles classified as GAs should receive some special privilege or first claim on editors' time because they at one point in their history passed a GAN review, then you lack a fundamental understanding of the process and need to stop engaging with it until you fix that. Thebiguglyalien (talk) 🛸 03:20, 15 June 2025 (UTC)

:*Most talk page discussions and GARs get no responses, as witnessed by the current list at WP:GAR. For a reviewer to list every concern in the initial statement would be time-consuming for no benefit if there is no response. I do not add citation needed templates to articles unprompted anymore because I get accused of WP:REFBOMBING or trying to prove a point. I am willing to add citation needed templates when asked, as stated in the opening statement of the Addition GAR. Editors can also use this script to highlight potential uncited statements in an article. If editors want to add additional processes before an article is nominated at GAR, they are welcome to propose them here, but not at a GAR. Z1720 (talk) 03:30, 15 June 2025 (UTC)

:*:Most "good article" nominations get no responses, as is witnessed by the even much longer list at Wikipedia:Good article nominations. More generally, most anything one person tries to do at Wikipedia typically gets few responses because other people are instead working on something else of more personal interest. But talk page discussions (and bold edits) remain the best tools we have. If you don't get a response for a while about some concern related to a mathematics related article, I recommend pinging WT:WPM, which is fairly active and pretty good at dragging at least a few editors to any relevant discussion that seems stalled or in conflict.

:*:@Thebiguglyalien my impression from reading a very large number of Wikipedia pages about a wide variety of topics is that the vast majority of the "very good" articles on the site, especially if judged holistically for quality, do not have a badge of any kind, and most articles which do have badges, including ones granted badges very recently, are alright but not amazing and usually have significant flaws, not particularly distinguishable as a group from, say, "B class" tagged articles except for the coincidence of having once gone through an arbitrary formal process. If someone really wanted to be strict and literal about the various checklist items probably 80% of them could be demoted for one reason or another.

:*:As far as I can tell the entire concept of "good articles" was originally created because the gold star reviews were too slow and onerous with a backed up queue, and GA was supposed to be a quicker and more effective process for getting more articles reviewed and given basic badges as a signal to readers that those articles are considered alright and a motivator for author–editors to put pages through some kind of peer review. However, the current GAN and GAR process is as, if not more, troublesome than the FA processes used to be, and now as a reader there is no obviously discernible difference between pages with gold star badges and little green circle badges and not really any obvious difference between review quality or criteria when reading archived reviews of gold vs. green badges, which reviews are highly variable in care and usefulness. The number of old "good articles" getting prodded for demotion seems about as great if not greater than the number of new good articles getting reviewed for promotion, while the number of subject-interested and somewhat experienced editors has not increased much if at all over the years. It seems likely any promoted article will eventually be demoted again, more or less at random.

:*:{{tq|i=yes|"I do not add citation needed templates to articles unprompted anymore"}} – I also wouldn't recommend leading by plastering articles with tons of templates. But a list of concrete and specific criticisms on the talk page is pretty much always helpful, even if other editors disagree about them. –jacobolus (t) 04:46, 15 June 2025 (UTC)

:*::"probably 80% of them could be demoted for one reason or another" seems a good reason to ease the process, not throw up additional roadblocks. (As does the comment about potential onerousness.) CMD (talk) 05:42, 15 June 2025 (UTC)

:*:::I guess it just depends what your goals are. I think the main goal should be improving the articles and directing effort toward facilitating improvement of the articles. If your goal instead is to randomly cycle badges and direct as much effort as possible toward toward badge management, then sure. –jacobolus (t) 16:56, 15 June 2025 (UTC)

:*::::The assertion that Z1720's efforts are random badge cycling is incivil, and incorrect to boot. Believe it or not, the article classification system is designed to facilitate article improvement, and to the extent ratings do anything, having the system be more accurate facilitates broader assessments and can direct improvements. None of this prevents any editor noting areas that articles, whether in the 80% or 20%, can be improved, or carrying out such improvements if they are interested. CMD (talk) 17:12, 15 June 2025 (UTC)

Can someone look at Talk:Halimah Yacob, either GANReviewTool or ChristieBot has misfired. CMD (talk) 18:11, 15 June 2025 (UTC)

:I was the reviewer. Sorry if I messed something up. Firefangledfeathers (talk / contribs) 18:18, 15 June 2025 (UTC)

::Ah no, I've figured it out. CMD (talk) 18:34, 15 June 2025 (UTC)

:::Thanks for the help! I hereby call for your Wikipedia salary to be doubled. Firefangledfeathers (talk / contribs) 18:37, 15 June 2025 (UTC)

::::Most of my salary is paid in tithe to ChristieBot. CMD (talk) 18:41, 15 June 2025 (UTC)

:::::Hah. Which never passes any of it on to me. CMD, what went wrong exactly? It led to something odd happening in the database which I'd like to fix, but I don't see exactly what happened. Mike Christie (talk - contribs - library) 02:11, 16 June 2025 (UTC)

::::::What happened was that the nominator originally requested that the nomination be reset as FireFangledFeathers stopped editing in November 2024. However, when FireFangledFeathers came back to editing, the original GAN was used leading to some technical bugs. It is a very edge case, although ChristieBot is still unhappy. CMD (talk) 03:18, 16 June 2025 (UTC)

:::::::Thanks; I think it's now fixed. The bot was confused because having seen a nomination with page 2, and seeing that the article was promoted, it could not assume the nomination was withdrawn so it was looking for a reviewer for that page number. I removed the references to page 2 for that article from the database and it should be fine now. Mike Christie (talk - contribs - library) 11:26, 16 June 2025 (UTC)

Hello,

I'm a new reviewer, and I'm not sure if I should quick fail one of the articles I've started reviewing. It has deprecated sources and it is relatively short, but I am not sure if I am required to quick fail it or not because of these factors. The article is Princess Si of Anding. Can somebody please help? 🌷Reverosie🌷^★talk★ 20:17, 15 June 2025 (UTC)

:{{u|Reverosie}}, decisions on passing or failing are always up to the reviewer's discretion. Which deprecated sources does the article use? ~~ AirshipJungleman29 (talk) 20:28, 15 June 2025 (UTC)

::I'm not actually sure. Wikipedia just told me that there were deprecated sources when I transferred the article to my sandbox, but not which ones were bad. I went searching for them, but had no luck, hence my request for help. 🌷Reverosie🌷^★talk★ 20:30, 15 June 2025 (UTC)

:::I think it's referring to Encyclopedia Britannica, which should be assessed per use, and Abebooks, which can include user-generated information. FWIW, I would be quickfailing that article for struggling to distinguish fact and fiction and using a storytelling tone throughout. If the nominator is a dedicated editor, they can turn that around quickly, but it might be something you'd want to ask for a peer review to comb through first. Kingsif (talk) 20:33, 15 June 2025 (UTC)

::::Thank you so much! I'm thinking that I might have to do that. I really don't want to quickfail an article, but this seems like it needs too much work for me not to, so I'm deeply considering it 🌷Reverosie🌷^★talk★ 20:36, 15 June 2025 (UTC)

::::I've failed the article and let the nominator know that I'd be happy to review it once more when it is ready 🌷Reverosie🌷^★talk★ 21:12, 15 June 2025 (UTC)

:::::Hi Reverosie, I see the article has been nominated again less than a day after quickfail - this is usually not enough of a refractory period for improvement or to reflect on the previous review comments. A quick glance at the current state of the article, it still has poor tone and now has unreferenced parts and unclear references. I haven't looked at the content enough to tell if the historicity is clearer, but given a new section called "verified facts" has been added, I get the impression that little thought has been given to improving the presentation of information throughout the article with this as a backdoor 'fix'. Feel free to copy these comments over to the new review, but I would advise asking the nominator to seriously go over the article, with an eye on MOS if needed, and wait at least a week before re-nominating. Kingsif (talk) 21:43, 16 June 2025 (UTC)

::::::Thank you for posting this!!! I was again questioning what to do, but didn't feel inclined to ask again. This is very helpful! 🌷Reverosie🌷^★talk★ 23:34, 16 June 2025 (UTC)

I've completed my first GA review. If any of the experienced reviewers has a few minutes to spare, I'd appreciate a quick review of my review, and any pointers for future ones. Thank you :) YFB ¿ 20:35, 15 June 2025 (UTC)

:I'm not nearly the most experienced reviewer there is, but with nearly ten under my belt, I'll say that this is a very nicely done review! I really like how you split up your comments. Keep up the good work reviewing articles! 🌷Reverosie🌷^★talk★ 20:38, 15 June 2025 (UTC)

::Thank you - that was quick! YFB ¿ 20:53, 15 June 2025 (UTC)

:That's a well-structured, well-explained review, especially considering a tougher topic. Your handling of GA criteria combined with the relevant other policies and guidelines was particularly good, and your interaction with the nominator and another commenter was a positive. I'll note that your review felt confident, and I think it warrants that confidence, which I hope you carry into future reviews :) Kingsif (talk) 21:51, 16 June 2025 (UTC)

::Thank you for taking a look and for this really encouraging feedback. YFB ¿ 23:16, 16 June 2025 (UTC)

Hello,

I've never been in a review circle before, but I noticed that four nominees are currently in the pool. How long does it usually take for the circle to open once this happens?

Thank you! 🌷Reverosie🌷^★talk★ 21:12, 17 June 2025 (UTC)

:I've added a new nomination to the pool after my original article was taken up by a reviewer, and word on this? 🌷Reverosie🌷^★talk★ 22:13, 19 June 2025 (UTC)

::Since we recently deprecated the coord position for GA review circles, any experienced editor can follow the directions to set one up. If you feel confident following the directions, you can do it yourself. ♠PMC♠ (talk) 22:24, 19 June 2025 (UTC)

:::Oh! Thank you so much! I'm quite new to wikipedia, so I'm not sure how well I could do it myself, but I might try! Thank you again 🌷Reverosie🌷^★talk★ 22:30, 19 June 2025 (UTC)

:Another follow-up: Would anybody be willing to open a review circle for the four articles there? I don't feel confident enough in my abilities to do it myself. Thank you! 🌷Reverosie🌷^★talk★ 23:03, 19 June 2025 (UTC)

Almost 900 nominations. 800 waiting to be reviewed. Still growing each day.

Backlog drives not having the same impact they used to. Review circles and mentorship doing good work, but not at a scale to make a huge difference.

Articles waiting almost a year to get reviewed.

I feel like something has to change, and QPQ seems to be the only thing that can make a difference at this point. Similar to WP:DYK, we could find a threshold so that new editors don't get dissuaded and to ensure reviewers have the experience of a certain number of GAs before doing their own reviews. Heck, we could only require QPQs for super GA writers, or for editors that have a huge difference in the number of GAs versus the number of reviews.

Many may state (which I will admit I have said): there's no rush; they'll get reviewed when they get reviewed; good articles still help Wikipedia, even if they aren't GAs; the GA icon doesn't mean that much; etc. But I think at this point we have a broken process, one that is not serving the purpose it is meant to.

QPQ has its challenges, obviously, related to quality of the review. But we have that problem today, even without QPQ. Imagine if we say all QA nominators with over 25 GAs under their belt are required to complete a QPQ review for each subsequent nomination. We could have a technical solution that flags this for reviewers, similar to DYK, noting that in order for this specific nomination to be approved, the nominator has to complete a GA review. At the very least, we only burden editors who have a fairly in-depth knowledge of GAN (how many people actually have more than 25 GAs?). « Gonzo fan2007 (talk) @ 17:42, 19 June 2025 (UTC)

:This isn't an endorsement of QPQ, but I believe we should lower our standards for what constitutes an acceptable GA review. Not in the sense of allowing quick passes, but by actively discouraging FAC or PR level reviews, or really anything that contradicts WP:GANOT. We also need to be quicker to fail nominations that aren't super close to meeting the criteria, but that would be much easier to do if we weren't condemning the failed nomination to another 6+ months of waiting in the queue. Thebiguglyalien (talk) 🛸 17:49, 19 June 2025 (UTC)

::I support both of these comments, {{U|Thebiguglyalien}}. « Gonzo fan2007 (talk) @ 18:27, 19 June 2025 (UTC)

:::I'll throw out one more, in that I'd be much more willing to review if it weren't for the time and energy investment expected for a spot check, which is much larger in a relative sense than every other aspect of the review, and in my opinion the most tedious. I get why they're there, but all the same they're the main thing that discourages me from reviewing. Thebiguglyalien (talk) 🛸 22:10, 19 June 2025 (UTC)

::I agree here; I think we may need to enforce returning an article to the queue if it is passed without spot checks, but beyond this I think a QPQ is a good idea. Generalissima (talk) (it/she) 21:06, 19 June 2025 (UTC)

:I would fully support instituting a qpq for editors with >X GAs. If you aren’t reviewing your fill of articles, you shouldn’t get the privilege of participating in GAN. Seems fair enough to me. Eddie891 ^Talk _Work 18:31, 19 June 2025 (UTC)

::Agreed. Making a GA nomination is, by definition, making a request for another editor's time and energy. If you are willing to ask this of someone, you should be willing to commit to returning the favor to the community. If that's too much to ask a nominator, then it's too much for the nominator to have asked of someone else. ♠PMC♠ (talk) 18:57, 19 June 2025 (UTC)

:::{{+1}} Vacant0 (talk • contribs) 19:16, 19 June 2025 (UTC)

::Alternatively, editors with more nominations than reviews. ~~ AirshipJungleman29 (talk) 19:39, 19 June 2025 (UTC)

:::In the end, this would be everyone. If I have never written a GA, then my first nomination is a "freebie". However, every subsequent nomination would require a QPQ, because technically my ratio would be 0-1, then 1-2, then 2-3, etc. I also don't want someone inexperienced with GA writing being forced to do a review they are ready to do. « Gonzo fan2007 (talk) @ 19:46, 19 June 2025 (UTC)

::::Sure, but if someone who's contributed several times more reviews than nominations in the past wants to nominate, I'd like to let them. ~~ AirshipJungleman29 (talk) 21:38, 19 June 2025 (UTC)

:::::No comment on whether or not this is the right idea, but copying the DKY five freebie limit seems a simply possibility for this option. It could even be extended to perpetuity, eg. requiring no more than five more GAs than reviews. CMD (talk) 02:59, 20 June 2025 (UTC)

• Support QPQ: The large backlog means new editors have to wait months for a review. This delays their ability to get feedback and improve their article-writing skills and discourages them from editing because they do not get the validation that someone cares about their contributions to the site. The process would be more effective if those with lots of GAs could only nominate more articles if they had a positive review-to-GA ratio. Numbers can be discussed after, but I think a general consensus of a QPQ is good to establish now. Z1720 (talk) 19:23, 19 June 2025 (UTC)

:*For context, there are currently 451 nominations where the nominator has a negative review-to-GA ratio. This number does not include nominators with 0 GAs. Z1720 (talk) 19:31, 19 June 2025 (UTC)

::*{{U|Z1720}} are you able to calculate how many (1) unique nominators there are and (2) how many of those nominators have 25 or more GAs? « Gonzo fan2007 (talk) @ 19:36, 19 June 2025 (UTC)

:::*{{re|Gonzo_fan2007}} There were 159 out of the 451 nominations with a negative review-to-GA ratio. For editors with 25 or more GAs with a negative review-to-GA ratio, there are 37 nominators with 162 nominations. If anyone wants to play around with the numbers themselves, User:ChristieBot/SortableGANoms has a chart of all the GANs: I copied the chart into a Google Sheets document, then sorted the table by various parameters. Since Sheets numbers the rows, I can figure out various statistics quickly. Z1720 (talk) 19:50, 19 June 2025 (UTC)

::I don't think it should be retroactively applied; requiring a positive review-to-GA ratio would mean several very prolific editors would be punished unduly and forced to review hundreds of articles before ever nominating another of their own. I think that'd be a very unfair expectation to suddenly put upon people.

::I think the best thing to do would be to give everyone with a negative review-to-GA ratio a fresh start (ie, zero QPQs, and just the one freebie) while people with a positive ratio should get to keep their current number of hypothetical QPQs. Generalissima (talk) (it/she) 21:09, 19 June 2025 (UTC)

:::That feels like a reasonable compromise. ♠PMC♠ (talk) 21:22, 19 June 2025 (UTC)

:::I do not think applying it retroactively it is punishing prolific editors. Instead, it is requiring them to help reduce the backlog that they contributed to. I also think resetting some editors ratios while leaving others in place might get complicated, as two groups of editors would have different stats (some would have total reviews complete, while others would have reviews complete from a certain time). I would prefer to have a time period for editors to get in a positive ratio (maybe six months?) and then start the positive ratio or QPQ requirement. Z1720 (talk) 21:24, 19 June 2025 (UTC)

::::I think it would be completely unfeasible for editors like BeanieFan (negative 204 QPQs) Chiswick Chap (negative 263 QPQs), Parsecboy (negative 536 QPQs) or EpicGenius (negative 538 QPQs) to do the needed number of reviews in any reasonable amount of time whatsoever. We would be arbitrarily barring some of our most prolific editors from nominating articles for several years at the very least. Generalissima (talk) (it/she) 21:32, 19 June 2025 (UTC)

:::::{{Ping|Generalissima|Premeditated Chaos|Z1720}} just to be clear, no one is proposing that you need to get back to even or a positive ratio in order to nominate an article. I think the proposal would be more "if you have a negative ratio today, then for every subsequent nom you need to do a QPQ". That said, I would not propose that as the rule. Its easier and more straightforward to just require a QPQ for any editor over a certain amount of total GAs/noms. This number is already automatically tracked for nominators and reviewers. Based on the data above, this would give us maybe an extra 150 reviews from the 40 editors or so who are prolific GA writers. And it would be a long term fix, as the QPQ requirement would not go away when someone reaches a positive ratio. As an example, I have written 47 GAs and have 150 reviews. I don't think I should be immune to the QPQ requirement and the likelihood is that with 47 GAs I know what is required for a good review. Taken with {{U|Thebiguglyalien}}'s comments above about sticking more to the intent of GAN, I think this would really help alleviate the backlog. I think our goal should be the reduction of the timeframe of reviews, not the total amount of nominations. Ideally, nominators shouldn't wait longer than a few months (the general timeframe that most other article review processes take). « Gonzo fan2007 (talk) @ 21:40, 19 June 2025 (UTC)

::::::As much as I wish those prolific editors had done more reviews in the past, and as much as the scale-balancing Libra in me would like to say "no, make up the deficit, get reviewing!", I think Generalissima's proposal is the best way to move forward. It's unreasonable to expect people to do several hundred reviews now when QPQ wasn't a firm expectation when they were making those nominations. (It's certainly always been a courtesy expectation; I was taught to do two reviews for every one nomination I made, but it's never been enforced). We'll get more people on board if we're willing to bend on this and not force people to do a ton of work just to get back to square one. ♠PMC♠ (talk) 21:45, 19 June 2025 (UTC)

::::::Just jumping in here, but there is an interesting Reward Board offer regarding the use of DYK QPQ "credits" as incentives for contritbuting. If an individual absolutely does not want to do GA reviews to build their personal QPQ bank, such tasks like that could be a way to improve the site without needing to do reviews (esp. since some editors would already have an existing surplus of these credits). Leafy46 (talk) 21:47, 19 June 2025 (UTC)

:::::::The whole point is specifically to reduce the staggering GA backlog. We already don't have enough people who review more than they nominate, so I'm not sure this would help very much with that. ♠PMC♠ (talk) 21:53, 19 June 2025 (UTC)

::::::::{{U|Premeditated Chaos}} as I mentioned, no one is proposing to force editors to catch-up in order to nominate articles. The QPQ would merely start at a certain point moving forward, requiring any additional nominations to have an associated QPQ. « Gonzo fan2007 (talk) @ 22:02, 19 June 2025 (UTC)

:::::::::How is that different from Generalissima's proposal, then? ♠PMC♠ (talk) 22:07, 19 June 2025 (UTC)

::::::::::It's not, I'm trying to clarify to everyone that no one is proposing that. I replied above to the three of you, including Generalissima, to clarify that I wasn't proposing this, because the comments made it seem like I was. That said, what I am proposing is to require QPQs after a set amount of GAs. This allows new or casual editors to avoid QPQ for the most part, while ensuring those who are required to do a review are capable of doing it positively. « Gonzo fan2007 (talk) @ 22:22, 19 June 2025 (UTC)

::::::::People may feel more incentivized to review articles and build their bank of QPQs, so they have more bargaining power towards other individuals who are more inclined to contribute through writing than through reviewing. The same amount of work would be getting done (i.e. articles would be getting reviewed), and people who like to do reviews could still see contributions in areas they want to see them in, without necessarily needing to do that work themselves. Ultimately, though, it's just an extension of what QPQs could entail. Leafy46 (talk) 22:12, 19 June 2025 (UTC)

:::::::::These are all very good points, but they raise questions. QPQ may lead to shorter articles receiving massive prioritization while longer articles are left to sit. This has grim implications for many vital and high-viewership articles. 🌷Reverosie🌷^★talk★ 22:28, 19 June 2025 (UTC)

::::::::::Perhaps this is a different issue entirely; More short articles being reviewed is not a bad thing. Regardless of what we do, long articles will likely have longer wait times. Vital and high-viewership articles, however, should have indicators in my opinion 🌷Reverosie🌷^★talk★ 22:36, 19 June 2025 (UTC)

::::::::::Short articles already get prioritized while longer, more complicated, or more esoteric articles sit, so I can't imagine QPQ would make it worse. ♠PMC♠ (talk) 22:44, 19 June 2025 (UTC)

:::::::::::I agree with this entirely 🌷Reverosie🌷^★talk★ 22:46, 19 June 2025 (UTC)

• What I would support would be a tiered cap on concurrent nominations based on review-to-nominations ratio. My concern with a straight QPQ is that we're going to see a marked decrease in review quality if we do that + some individuals are simply better at writing content than reviewing other people's work. And if you aren't reviewing but are only nominating a few at a time, then if there's people who consider your work valuable enough to review then that's OK to me. Where I think the process bogs down is when you have bulk nominators who don't review. As of [https://en.wikipedia.org/w/index.php?title=Wikipedia:Good_article_nominations/Report&oldid=1296276828 this version] of the GAN report, the 10 editors with the most nominations have 171 nominations open. Running them through Mike Christie's GA stats tool, these editors have Reviews - successful GA nominations of +1, -538, -263, -7, +87, -6, +1, -4, -204, and +2. Of the ones who are triple digits in the negative, one has never completed a GA review. I don't think we should be blocking somebody from nominating entirely just because they haven't been reviewing, but to me this is like a community garden. If you're going to be consuming a disproportionate amount of the end benefits, you should at least by trying to put something back into the system. Rate limiting will avoid having a situation where someone cannot participate in the system for lack of reviewing, but will limit the concurrent amount of benefit the large negative balances will get out of the system at once. Hog Farm _Talk 22:27, 19 June 2025 (UTC)

::{{U|Hog Farm}} would you elaborate how this would work in practice? Like if someone was 200 GAs/20 reviews, compared to someone who is 50 GAs/50 reviews and someone who is 20 GAs/200 reviews, as an example. « Gonzo fan2007 (talk) @ 22:32, 19 June 2025 (UTC)

::: Just spitballing something here - I haven't looked into the statistics or anything yet. If you have over a certain minimum sample size of review/nominations and you're above a certain nominations-reviews ratio, you can have up to a certain number of nominations open at once. Someone with a lower ratio would have a lower number of articles they could nominate at once. We'd also probably want to have some sort of cap for those under the sample size - if you've never had a GA promoted before, you really don't need to have a ton open. I know from personal experience of my first 4 GAs that were ever promoted, I've since gone back and completely rewritten two, have gone back to make substantial improvements to another, and sent the other one to GAR where it was eventually kept. Just off of gut numbers, which could easily be improved by anyone with more time in the data, I'd say maybe in the 12-15 range for 200/20, 7-10 for the 20/20 and 50/50 (I'd treat them both the same; they've put in what they took out), and then something like 3-5 for 20/200. Nothing overly drastic. Although I'm personally colored by my bias that I'm not sure that anyone can really give sufficient attention to 20+ GA nominations at the same time. The biggest sticking point I can see is where somebody is up against their cap, has to do a ton of reviews to get to the next tier, and all of their current nominations have been sitting around for months. Hog Farm _Talk 22:43, 19 June 2025 (UTC)

::Ooh - I like this, that's a really good compromise. It doesn't shut people off completely, it avoids some of the problems of a strict QPQ, but it encourages reviewing a lot more. Generalissima (talk) (it/she) 22:33, 19 June 2025 (UTC)

::I like this idea quite a bit. It encourages more reviews without shutting people out entirely. It also doesn't force people to review as much as they nominate, which could lead to less people leaving. 🌷Reverosie🌷^★talk★ 22:35, 19 June 2025 (UTC)

:::However, what about nominators who don't review articles because English is not their first language? 🌷Reverosie🌷^★talk★ 22:40, 19 June 2025 (UTC)

:::: That's why I think the cap should always be at least a couple. I hope this isn't too blunt, but there is the concern that if you have enough struggles with learning English that you can't review, your nominations probably need more work as well, so it would make sense to not have a large number of those open at once. Hog Farm _Talk 22:43, 19 June 2025 (UTC)

:::::That makes a lot of sense 🌷Reverosie🌷^★talk★ 22:47, 19 June 2025 (UTC)

::::::{{U|Hog Farm}} my only concern is that all this does is effectively hide the backlog. Any restriction on nominations will inevitably lead to nominators queuing noms and nominating immediately once another is complete. QPQs actually require clearing out the backlog, instead of hiding it. « Gonzo fan2007 (talk) @ 23:05, 19 June 2025 (UTC)

:::::::Yeah, we already hide nominations when there's more than 20 from one person, this would just be a shadowy extension of that practice, and does not encourage people to actively review content made by others. ♠PMC♠ (talk) 23:20, 19 June 2025 (UTC)

:::::::Well, is that really true? If I have only 3 nominations up, it might take months before someone reviews one of them. When I put up 20 nominations, the odds that at least one of them will be reviewed in a timely manner goes way up - so if I were limited to only having 3 nominations at a time, I'd be able to get new GAs at a much slower pace than if I was able to put up as many as I wanted, even if I had the same "queue". Generalissima (talk) (it/she) 23:25, 19 June 2025 (UTC)

::::::::I don't find this argument convincing; I think Gonzo has the right of it here. Given how long GANs already sit, I don't think "you can let more noms sit at once" is a strong enough incentive to review that it will actually dent the backlog. You're a bit of an outlier here in terms of your high output and varied topics. If I don't want to review X topic, I might still review one of your others on Y or Z, so for you, having more nom slots probably does up your chances of someone picking up your review.

::::::::On the other hand, most prolific editors have a more limited area that they tend to write in, so having a greater number of noms doesn't necessarily make them more likely to get reviewed. If I won't review architecture, Epicgenius is out of luck whether he has 1 nomination or 100. What's the incentive for someone like him to do a bunch of reviews just so they can let 20 noms sit instead of 10? Even if this worked for the people with a high number of simultaneous nominations, there are far fewer people who have 10+ or even 5+ active nominations at one time than there are people who have 1 or 2, so this is not at all an encouragement for those people to increase reviewing. ♠PMC♠ (talk) 23:59, 19 June 2025 (UTC)

• :We should explore rate limitation in general (and probably should not have reversed the successful initiative to shift default order). I believe the last proposal was X articles per author per topic (ie. per lv2 header at GAN). This was a good idea to limit backlog growth while also facilitating reviewers who may work mostly/only in specific topics. The hiding the backlog arguments frame GAN as a mechanical system, but HogFarm is absolutely right that this is a "community garden", given we are managing limited resources and are all volunteers. A bit of selectivity in what you plant in your allotment does not seem very onerous. CMD (talk) 03:12, 20 June 2025 (UTC)
• ::In what way does this encourage reviewing, as opposed to simply throttling nominations? ♠PMC♠ (talk) 03:35, 20 June 2025 (UTC)
• :::Possibly in the way HogFarm mentions, but a more core point is perhaps that this sort of question purposefully sidesteps the question of community responsibility being raised. CMD (talk) 03:54, 20 June 2025 (UTC)
• ::::I agree with Hog Farm that GAN is a community garden, always have. You may have missed my first comment in this discussion, in which I specifically say "Making a GA nomination is, by definition, making a request for another editor's time and energy. If you are willing to ask this of someone, you should be willing to commit to returning the favor to the community."
• ::::However, I simply cannot see how simply throttling simultaneous nominations encourages most editors to do more reviews. Since I do not believe it would encourage more reviews, I cannot see how it encourages the common good. My reply to Generalissima above lays out my argument to this effect. I think this discussion would be more productive if you presented yours, as opposed to accusing me of "purposefully sidestepping". ♠PMC♠ (talk) 04:23, 20 June 2025 (UTC)
• :::::The question sidesteps the issue because it narrowly focuses on the generation of new reviews as the only metric to examine. There is more to the system than a simple I/O process. It should not be taken for granted that a small number of editors requesting disproportionate amounts of the community's time/efforts is normal. WP:FAC is limited to one nomination by default. CMD (talk) 05:36, 20 June 2025 (UTC)
• ::::::The goal should be to encourage editors to make as many articles as possible meet the standards of being a good article. In an impossible ideal world, all Wikipedia articles should meet those standards. Artificially throttling prolific editors does the opposite of that. And if the system is not scaling now, then pushing people away from the system will not help it scale because those same pushed-away people will also be discouraged from doing reviews. Fewer nominations and proportionately fewer reviewers will just leave us with the same problem of nominations waiting too long for reviews.
• ::::::If, on the other hand, what one wants to encourage is a system where badges are precious and editors squabble over who can get one, then throttling nominations seems a step in that direction.
• ::::::What we should be doing instead is somehow encouraging more reviews. I have qualms about a QPQ system (experience with DYK QPQs shows that it incentivizes cookie-cutter checkbox reviews instead of depth of reviewing, and a QPQ system will only work for balanced editors who like both nominating and reviewing; it will push away some editors who are comfortable nominating but for whom the confrontational nature of reviewing is a blocker) but at least it has the clear goal of encouraging more reviews. —David Eppstein (talk) 05:47, 20 June 2025 (UTC)
• :::::::Well put. ♠PMC♠ (talk) 05:58, 20 June 2025 (UTC)
• :::::::Agree IAWW (talk) 09:55, 20 June 2025 (UTC)
• ::::::We're taking it for granted that a GAN review is one editor giving time to another editor, as opposed to two editors giving time to an article. Thebiguglyalien (talk) 🛸 06:24, 20 June 2025 (UTC)
• :::::::No we are not, we are in fact directly challenged by the fact that the time for the second editor's investment is not keeping up with demand. The related claim that "Artificially throttling prolific editors does the opposite of that" in unfounded, especially as it does not grapple with the plain fact that this system exists at FAC, and supported by "those same pushed-away people will also be discouraged from doing reviews" which somehow ignores the stats HogFarm pointed to. The claims also ignore those who are turned away by the slow and backlogged system. CMD (talk) 06:32, 20 June 2025 (UTC)
• I hate "stick" methods for making people do reviews. I tend to review articles I am interested in or during backlog drives. In the first case, I get to engage in depth with an interesting topic where the spot checks take no time at all because I read the sources anyway. In the second case, I get the satisfaction of being part of a joint effort to kill the backlog and some silly recognition like barnstars. As a nominator, I hope for reviewers interested in the article who are willing to find ways to improve it. I am worried I will get more quick passes and fewer interested reviews out of QPQ reviews. A general QPQ requirement would not make me do additional reviews (I already do about two per GA) but I do not think it would improve anything. Overall, some current issues at GA seem to be:
• Lots of content by the same nominators in relatively narrow topic areas with not enough reviewers in these areas
• The recent backlog drive did not work as well as some others have in the past (I don't know why, I didn't have time so did not participate)
• Some general disagreement on how reviews should be performed.
• At the same time, interesting stuff often gets picked up quickly (I think James Cook was nominated and reviewed in under a week). I think my general question is how can we make it more fun? —Kusma (talk) 07:04, 20 June 2025 (UTC)
• :It would be great if this could be done, but I can't think of a way to make reviewing so much more "fun" that it would actually have an impact on the backlog. IAWW (talk) 10:09, 20 June 2025 (UTC)

:I see QPQ as the only way to make a real difference to the backlog, which I think is at crisis levels. Once the backlog is down <200 or so, we could remove QPQ again. I would be more in favour of everyone starting on a blank slate with QPQs, and I think the only exceptions to the requirement should be new nominators (<5 noms). I'm not really a fan of rate limiting solutions because I think they will have the annoying effect of "hiding" the backlog, are not as scalable, and are more complicated. IAWW (talk) 10:06, 20 June 2025 (UTC)

::We could also just advertise review circles better, which are a bit like QPQ: agree to review an article, get your own article reviewed now instead of in three months. Works great if you want to know when you'll be busy with GA reviews from both sides. —Kusma (talk) 13:23, 20 June 2025 (UTC)

:::"Advertising" could mean a talk page message to all nominators who have not yet participated in the circles, perhaps? Mike Christie (talk - contribs - library) 13:34, 20 June 2025 (UTC)

::::Sounds good to me. Review circles mean that you can choose whether you want your nomination to be affected by the backlog. I do not think a lengthy backlog of unreviewed nominations by people who never review is a major issue. —Kusma (talk) 13:42, 20 June 2025 (UTC)

• As a user with a three figure gap of more reviews than GAs, I would not support QPQs here, for a variety of reasons. Most obvious is that we’d be pushing users who can’t do a good review to not only do so, but to do multiple as quickly as they can. But even if we build in failsafes for that, I’m inclined to think it’s not ‘fair’ in the way being proposed. By this I mean, I think in most cases we’d be asking the nominator to do twice the work of the reviewer, if not three times. Very rarely are GANs quick passes, and by nominating a user has taken on the responsibility to collaborate in the review. We’re all quite annoyed with drive by noms that don’t take part in the review process. It may not be as heavy as the reviewer that often, but it’s still committing their time and effort for the ultimate benefit of Wikipedia. If they did a preliminary check that the article met GA requirements beforehand, that’s another pseudo review, so to ask for a QPQ on top does seem unfair. Can we look at each nom as a reviewer and nominator working together - I see Epicgenius called out above for a three figure deficit in reviews but I’ve been reviewer for multiple of those noms and the participation really could be tantamount to a review. The other big reason that comes to mind is that GA will still lean on users who just willingly review more than they nominate, and that comes (I believe) from such users understanding and valuing the purpose - IMO we are unlikely to produce many more of such editors if there are obligations and requirements in the process, it may seem more cold than collaborative. Kingsif (talk) 14:10, 20 June 2025 (UTC)
• :{{tq|I think in most cases we’d be asking the nominator to do twice the work of the reviewer, if not three times}}. I just don't agree that this is the case, or even that it's a bad thing. If we're assuming that the GA reviewer is also a GA nominator – and most are – they are also doing this work, just elsewhere, with their own article. In a perfect world, it's a virtuous circle - you do one, you review one. Everybody shares the load for the common good.
• :{{tq|GA will still lean on users who just willingly review more than they nominate}}. This is not working now, it has never worked for as long as GA has existed, and it will continue to not work in the future, because there are not enough of these people to cover the amount of people who want to write without doing the work of reviewing. This is the whole reason people are proposing QPQ in the first place. Continuing to believe that outlier reviewers will save us, without significant changes to the process, is nothing more than willful blindness. ♠PMC♠ (talk) 14:45, 20 June 2025 (UTC)
• ::On the first part, I do agree - I later wrote {{tq| Can we look at each nom as a reviewer and nominator working together}}, which sums up my view on that. My argument is that a nominator actively participating in the review is effectively QPQ and it doesn’t matter which side of the equation (nominator/reviewer) a user is, the workload is shared already. I accept that if the reviewer is a nominator elsewhere, asking for QPQ will maintain this balance, but it is still demanding more work roughly over the same time, risking burnout.
• ::Om the second, I don’t think “outlier reviewers” will save GA, I don’t even think the bulk of reviews are completed by users like myself. But I don’t think making GAR a chore will benefit editor retention, and won’t encourage users to just believe in the project. Kingsif (talk) 20:16, 20 June 2025 (UTC)
• :Quite frankly, if you can't give a good review after tens of GA nominations, you probably shouldn't be nominating articles anyways- imo most users with lots of GAs are capable of good reviews, but just don't conduct them. Eddie891 ^Talk _Work 14:46, 20 June 2025 (UTC)
• ::I disagree, to a point. Not every A student makes a good teacher, likewise, understanding the GA criteria enough to get an article through as nominator does not necessarily mean someone can translate that understanding to a review. Maybe if nothing or little needs to be improved, and I do think most people are able to learn how to do a good review, but I’ve taken over enough to know that it’s not a natural skill for everyone. Kingsif (talk) 20:21, 20 June 2025 (UTC)
• I stand by what I said the last two times (2023, 2024) this has been proposed: The potential benefits do not outweigh the drawbacks of potentially (1) discouraging high-quality nominations and (2) encouraging low-quality reviews. I'll also reiterate what I said in response to another 2024 proposal: I am in principle decidedly in favour of incentivizing reviewing, just not like this. I had no strong opinions on either changing the sort order to be by review ratio or changing it back, but both[https://en.wikipedia.org/wiki/Wikipedia_talk:Good_Article_proposal_drive_2023#Proposal_9:_Change_sort_order_of_GAN_page_to_prioritize_frequent_reviewers][https://en.wikipedia.org/wiki/Wikipedia_talk:Good_article_nominations/Archive_31#RfC_on_the_sort_order_of_WP:GAN] had fairly strong consensus in favour at the time. The sort order by review ratio did incentivize me to review more than I otherwise would have, although this mainly took the form of encouraging me to pick up clearly-premature/unprepared/deficient nominations to close as unsuccessful rather than ignoring them as I otherwise might have—which may or may not be the kind of behaviour we want to encourage. I would at any rate not be opposed to changing the sort order to incentivize reviewing. TompaDompa (talk) 15:03, 20 June 2025 (UTC)
• :If none of these ideas work to cut the backlog (as they haven't so far), at what point would you support QPQ? What if the backlog reaches 2000 noms and over a year to get a review? IAWW (talk) 16:46, 20 June 2025 (UTC)
• ::Just chiming in that I would still prefer rate limits to QPQs. Force less work rather than more. QPQ, even if newbies are exempt for a period, can still feel like a barrier to participation. Fewer new users involved long-term, and more burnout is what we will get. I just won’t support asking the same pool of users (generally) to do extra work when (per above) IMO being an active nominator is already doing their part. Kingsif (talk) 20:33, 20 June 2025 (UTC)
• ::I would rather accept longer waiting times than lower quality. If we really want to reduce the size of the backlog at all costs, we could simply impose a moratorium on new nominations until the backlog is reduced to, say, 200. That would be a pretty drastic move, but it would (temporarily) reduce the size of the (visible) backlog. TompaDompa (talk) 23:25, 20 June 2025 (UTC)
• :::Artificially throttling nominations is not a functional solution, it just conceals the problem. ♠PMC♠ (talk) 23:31, 20 June 2025 (UTC)
• ::::I agree. Requiring QPQ would be artificially throttling nominations and as such not a functional solution but a way to conceal the problem. TompaDompa (talk) 23:41, 20 June 2025 (UTC)
• :::::I would appreciate if you could not twist my words. That is not what I said and you know it. I was responding to your suggestion that "we could simply impose a moratorium on new nominations until the backlog is reduced to, say, 200", which is a suggestion to throttle nominations. Requiring QPQ is not a system of throttling nominations, it is a system that asks people to put in what they are taking out. ♠PMC♠ (talk) 23:59, 20 June 2025 (UTC)
• ::::::I apologize. I do indeed know that was not what you meant and it was not my intention to twist your words. My intention was to make the point that requiring QPQ would also, in a way, throttle nominations. If editors have articles that they want to nominate but because of some reason that has to do with the process do not do so (yet), their nominations have in effect been throttled. For instance: if an editor has 20 articles ready to be nominated but can only nominate 5 of them at the moment because that's the limit imposed by their current QPQ status, 15 nominations have been throttled. The effect is not in actuality reducing the size of the backlog by 15, but artificially hiding part of it (15, to be precise) out of sight. If those articles are later nominated, QPQ has throttled them and the quantitative effect on the backlog related to those articles (both in terms of number of nominations and time from nomination to review) is purely cosmetic—or in other words false. If those articles are not later nominated, QPQ has had the effect of discouraging nominations that otherwise would have been made. I do not view either of those as a beneficial outcome, and I would not characterize it as a functional solution. TompaDompa (talk) 01:03, 21 June 2025 (UTC)
• :::::::The current backlog also throttles nominations. I don't generally tend to GAN because of the long times and because I'd be adding to a problem. I've talked to people who simply don't bother because of the time too. Theoretical discussions of throttling miss that it already exists, plenty of stuff is already out of sight. CMD (talk) 03:04, 21 June 2025 (UTC)
• The GA reviews and nominations by editor in 2025 can be seen in [https://docs.google.com/spreadsheets/d/1C2MKT7onWz2Cvbpx1yfONo_-nJb94F6o7F166uoqWpk/ this Google Sheet]. Sheet #2 applies QPQ at different amounts. ~35 editors would be affected if the QPQ is required after 5 nominations. This would impact ~1/3 of the deficit. Some editors mentioned above who have large gaps in the nom/review ratio, such as Chiswick Chap, have reviewed more than they have nominated this year. Rollinginhisgrave (talk | contributions) 16:48, 20 June 2025 (UTC)

Just to review a few key points:

• Past changes have not significantly helped: review circles, more backlog drives, changing the order of reviews, etc have not had an appreciably positive impact on the backlog.
• Throttling nominations does not improve the backlog: throttling nominations only forces editors to keep articles off the nomination page; this is done at WP:FAC and WP:FLC because editors are less likely to have that many noms waiting and each of these processes require multiple editors to review each nomination.
• A negative ratio of reviews versus noms is the cause of the problem: this is a good problem to have! This means we are collectively writing good content, and understandably, reviewing is difficult and time consuming. That said, if one editor nominates 100 articles and reviews 0, then the result is clearly that someone else (or multiple others) will need to review more articles than they nominate. In order for the process to work, this has to happen.

All that said, this then becomes a question of fairness. If we are going to have a GA process then what do we have to do in order to promote a fair and functional process. QPQs are not perfect, but incorporating QPQs would at least address the problem at hand. The likely sacrifice is quality of reviews, but those can be handled fairly easily. This is also why QPQs should only be required for experienced GA writers. The thought process being "if you are going to reap the benefits of this process, then you need to invest into the process."

A general idea of how this would work would be QPQs only being required for editors with X number of GAs (let's say 25). Some of this will require the honor system, but we likely could change the nomination template to include how many GAs a nominator has, and if it is over the magic number, then a big, bold text says "this nomination requires a QPQ". There would be a place where the nominator would have to link to the review that they completed. Just like at WP:DYK, a reviewer would then need to check if the QPQ has been completed before proceeding with reviewing the nomination. As stated above, this would be applied to everyone equally, regardless of past number of GAs or reviews (as long as they are above the threshold). The threshold would help prevent new editors who don't understand GA process from having to review articles and likely providing less than thorough reviews.

This would almost immediately slow down the growth of the backlog and provide backlog drives an opportunity to make a dent in the current deficit. It would also create a reciprocal process and we could always still utilize the talk page to flag deficient reviews (which already happens today). If someone has already written 25 GAs, they are much less likely to game the system or abuse this process, as intrinsically they are more invested in the community and seeing this process work. This also doesn't prevent editors from writing good content, it again just necessitates investing in a process if you want to take advantage of the benefits of the process. « Gonzo fan2007 (talk) @ 16:28, 20 June 2025 (UTC)

:Could QPQ be seen as a temporary measure until the backlog gets below 200? Then we can just wait another 5 years before requiring it again. IAWW (talk) 16:48, 20 June 2025 (UTC)

:Review circles may have had little numerical effect on the backlog, but they mean the backlog is not mandatory: everyone can choose to skip the backlog queue. If people choose not to, maybe the backlog isn't really as much of a problem? —Kusma (talk) 16:49, 20 June 2025 (UTC)

::Alternately, it means review circles are not the solution. Personally I avoid them because I like to pick my reviews based on interest rather than have them be assigned. I'm sure I can't be the only one. ♠PMC♠ (talk) 17:04, 20 June 2025 (UTC)

:::You can say what general topic areas you would like to avoid, and you can say no if you are assigned something really unsuitable. But yes, interest is a general issue. I would review more if there were more articles to review that are interesting to me. And I get bored if I do more than one or two radio or TV stations or NYC buildings or Tolkien scholars per year, even if I enjoy working with all of their nominators. —Kusma (talk) 17:49, 20 June 2025 (UTC)

::Also I think the long backlog's biggest negative effect is on new nominators, who often don't feel comfortable reviewing other people's work before they have a GA under their belt. These people won't use review circles. IAWW (talk) 17:11, 20 June 2025 (UTC)

:::I think we should highlight new nominators and encourage people to prioritise reviewing their articles. —Kusma (talk) 17:27, 20 June 2025 (UTC)

::::Let's take this to the extreme. Let's say we have 10,000 active nominations and it takes on average 3 years for a review to be completed. Is this a process worth having? I struggle with the benefits of this GAN process if it doesn't actually function as it should. So if everyone agrees GAN process is going to stay, we need to figure out some way to ensure that reviews are being completed in a timely manner, which is my mind is just a few months, not years. I have 3 times the reviews as I do GAs and my oldest GAN is from December 2024, and its a short article (only 900 words). The process is broken, we need something to fix it. « Gonzo fan2007 (talk) @ 17:34, 20 June 2025 (UTC)

:::::The process is not working great, but it not at all like the strawman you describe (which again has nothing to do with my suggestion to be nice to new nominators, so I do not know why it is a response to my comment). —Kusma (talk) 17:39, 20 June 2025 (UTC)

::::::It was more a response to your earlier comment {{tq|If people choose not to, maybe the backlog isn't really as much of a problem?}} I agree that we should be nice to new nominators, definitely. I just don't see that as a huge problem. What I see is a backlog that is growing and nothing being done to address it. So I'm trying to describe a scenario that I imagined almost everyone would agree would be problematic. I would hope everyone would agree that nominating an article in 2022 and having to wait until now for it to be reviewed would be problematic. At our current pace, we will get there without any changes. « Gonzo fan2007 (talk) @ 18:06, 20 June 2025 (UTC)

:::::::Usually our backlog drives work well to eliminate the backlog of oldest articles, which are worth more points. I have not seen noms older than half a year survive a backlog drive where reviewing these was incentivised. I did not have time to participate in the most recent backlog drive, which had the different aim of encouraging new reviewers. I expect the next standard backlog drive will ensure all of the oldest articles get reviewed again. —Kusma (talk) 18:23, 20 June 2025 (UTC)

::::::::Maybe, or maybe everyone is sick of backlog drives and we get diminishing returns again. In any case, if we have to rely on backlog drives to get things done, it suggests we don't actually have a functioning process, and maybe we should admit that. ♠PMC♠ (talk) 19:07, 20 June 2025 (UTC)

:::::::::It is quite demoralising participating in a backlog drive and actually having the backlog increase for half of it, and then barely decrease overall. That's what happened last drive. IAWW (talk) 19:48, 20 June 2025 (UTC)

:::::If even half of everyone who nominates an article has done the prerequisite checks, that article should already be improved to a decent standard, and GA has already served a purpose before a review has taken place. I like GA (over FA) as a process because I think it strikes a good balance between quality and quantity - a sign off that the article can be considered reliable without such an extensive depth and process that limits creating/editing elsewhere. The process isn’t broken. Kingsif (talk) 20:28, 20 June 2025 (UTC)

• There is another cause for the backlog that has not been considered yet: this year's WikiCup has produced unusually low excess of GA reviews, see [https://wikicup.toolforge.org]. More WikiCup points might help with the backlog, or other gamification that I am not clever enough to design. —Kusma (talk) 17:36, 20 June 2025 (UTC)
• :@Kusma, that's very interesting. Any idea why? And are other elements of WikiCup also skewed in some way? I've noticed that the requirements to place at all in each round are higher than normal, but I don't really follow the wikicup and hadn't noticed this thing about the GA reviews. -- asilvering (talk) 23:59, 20 June 2025 (UTC)

:Make reviewing faster and easier for the reviewer. Change the process, change the criteria, change the expectations, whatever. More people will review, and reviewers will be able to get more reviews done in the same amount of time. I've largely stopped reviewing because as soon as I start, I've just committed to a few hours of work. Nothing will change until we make reviewing easy. Thebiguglyalien (talk) 🛸 19:36, 20 June 2025 (UTC)

::Reviewing is always going to involve some decent time commitment. "Nothing will change until we make reviewing easy." - you think QPQ wouldn't work? IAWW (talk) 19:53, 20 June 2025 (UTC)

:::I suppose QPQ will make reviewing easier in the sense that most of us are going to cut more corners in our reviews. Maybe that will end up redefining standards anyway, albeit in a less focused way. Thebiguglyalien (talk) 🛸 20:24, 20 June 2025 (UTC)

:I'm not sure any of the three key points is fully correct. We can't be sure of the first, because we can't really A/B test it. Don't discount that things could be worse! The second is a contradiction, an article not on the page is not on the backlog, and it assumes that the nominations are equal to all possible nominations under a different system. On the third, it's worth keeping in mind the long view, in that we have historically a roughly equal rate of noms and reviews as each nom comes with one review, with the current backlog representing the rounding error difference. We can't have a positive ratio, and will always rely on some reviewing more than they nominate under any system that gives slack to newer participants (be that through regular volunteering or a drive). It's also worth noting that reviews to noms is not a one-way interaction, noms do not occur totally independently of the system and it shouldn't be assumed that the existing backlog has no impact on other noms. Do we have stats on how the divergence between noms and reviews is progressing? How has the median age of noms changed over time? I also lack confidence the bring reviews here to reassess process is working, as we have no metrics on that. Nobody has found the time to sort of Talk:2024 Men's T20 World Cup Super 8 stage/GA1 yet, as an anecdote. If Rollinginhisgrave is correct, we'd still have 66% of the backlog in a QPQ situation, so that is still only one part of a system that will still require drives or similar. CMD (talk) 20:04, 20 June 2025 (UTC)

::You can look back in the history of the page or the past backlog drives. We can be sure of the first. In the last 3 years, the total number of nominations have almost doubled, as with the total number of unreviewed noms. Regarding the second point, you are just talking semantics. Yes, if we change rules to remove a backlog, then we don't have a backlog. A fairly straightforward statement that misses the point though. We are trying to find a solution to move reviews more quickly, not hide the problem. No one is looking for a positive ratio. What we are looking for it to avoid articles waiting almost a year to get reviewed. That's the problem, which comes from people nominating more articles than get reviewed. It does not come out equal, as we wouldn't have a growing backlog if it did. Regarding the QPQ, a 1/3 reduction is noms and a 1/3 reduction in wait time is beneficial, even if we don't fully solve the problem. « Gonzo fan2007 (talk) @ 20:19, 20 June 2025 (UTC)

:::I for one would rather wait 50% longer for an in-depth review than get a checkbox review quickly. —David Eppstein (talk) 20:22, 20 June 2025 (UTC)

::::Yet you really can't control who your reviewer ends up being, even today. I've had GAs of mine reviewed without comment and passed, and had GAs reviewed as if it was a WP:FAC. At it's core though, WP:GA is a checkmark: i.e. its a verification tool that someone has looked at the article and said "this is good" per these criteria. At its core, its an inherently subjective assessment. FAC or even peer review would likely be better tools for most editors if they want extremely high quality reviews. « Gonzo fan2007 (talk) @ 20:27, 20 June 2025 (UTC)

:::::Perhaps somewhere (I admittedly don't know where) we could spell out that GAN is specifically for basic reviews of articles that already more or less meet the listed criteria, while FAC is for detailed reviews of already high-quality articles and PR is for flexible reviews of any kind. GAN is trying to be a bit of everything at once. Thebiguglyalien (talk) 🛸 21:51, 20 June 2025 (UTC)

:::No, {{u|Chipmunkdavis}} is correct. We cannot be sure that previous actions have not helped, because we do not and fundamentally cannot know what the situation would have been like if those actions had not been taken. What we can say for sure is that they have not been sufficient (for some given value of having a sufficient effect). TompaDompa (talk) 23:39, 20 June 2025 (UTC)

:How many times has the idea of QPQ at GAN been brought up at this point yet rejected? Anyways, I don't like the idea still. I see zero reason to punish other editors' content output like this, and GAN and FAC effectively have a de-facto QPQ anyways (reviewing more than you nominate = people are probably more likely to review your nominations). Or at least, I personally take however much an editor reviews into consideration on when I choose to review, and I don't expect other editors to review my stuff if I slack off in reviews. But that doesn't mean make my thoughts into bureaucracy.

:Not everyone is good at reviewing articles, and that's okay. Users should not be forced to do so. That's how you get half-baked or not-baked-at-all reviews. Ones done not because they want to, but because they have to. And over time, I could see it turning into a cycle that could heavily tank their motivation. I know it would kill mine even if I already try to keep my review count over my GA count (kinda falling off from that lately but I'm working on it). DYK is something which makes sense to have a QPQ - after all, it's an opportunity for content to be featured on the main page. That obviously needs some sort of regulation. But for GAN, I don't see it. The backlog is temporary, but the consequences that would be inflicted onto people would outlive it.

:I could also argue over the consequences this would have on newer editors but I guess that's one thing this discussion has an idea on how to handle per the bullet points. Those points aren't enough for me to view this idea positively, though. λ NegativeMP1 20:52, 20 June 2025 (UTC)

::{{tq|The backlog is temporary}} - is it? Reductions in the backlog are temporary, due to backlog drives. Aside from that, the backlog has always continued to grow, because more people want to nominate than want to bother to review. It's simple math. ♠PMC♠ (talk) 20:59, 20 June 2025 (UTC)

:::With the point I was attempting to get across, yes. I probably should've worded it as something along the lines of it being fixable through other methods, or will eventually not be viewed as that bad. Arguably, there will always be a backlog - whether its 100, 500, 900, or so on and so forth. But my point remains the same.

:::{{xt|more people want to nominate than want to bother to review}} - So? I still don't see a reason to force editors into reviewing. I still stand by what I said. λ NegativeMP1 21:06, 20 June 2025 (UTC)

::::You don't think it's a disservice to the people who want to nominate that they sit waiting for half a year or more? Does that help improve the encyclopedia? ♠PMC♠ (talk) 21:21, 20 June 2025 (UTC)

::::{{tq|So?}} I mean, when I started writing GAs, it never took longer than a few months for a review. Today, 6 months to a year is common. I'm feeling discouraged with the process, and I've been editing for 20 years. I can only imagine how it feels for a newer editor, or one who hasn't gone through this process before. Again, if we value this process, then it should work fairly well. Imho, this many noms and this long for reviews is beginning to break the process, thus rendering it either pointless or counterproductive. « Gonzo fan2007 (talk) @ 21:23, 20 June 2025 (UTC)

:::::I never really denied that how it is right now isn't discouraging in of itself. I can easily understand that long wait times are very frustrating. But is forcing reviews going to be any less discouraging? That's the argument I'm making right now. λ NegativeMP1 21:59, 20 June 2025 (UTC)

::::::Yes! Because at least things will be getting done! ♠PMC♠ (talk) 23:10, 20 June 2025 (UTC)

:::::::At the detriment of other editors...? I'd rather have seemingly no progress than bad progress. I'm really not seeing it, sorry. λ NegativeMP1 04:03, 21 June 2025 (UTC)

::::::::If I have anything further to say about this discussion and the thought about it getting implemented, then I think a message should be sent out to most, if not all active editors at GAN through their talk pages asking for their opinions if this discussion ever turns into a serious proposal / RFC. There should not be, under any circumstances, serious changes made or even proposed to the GAN process without a notice given to its participants. λ NegativeMP1 04:23, 21 June 2025 (UTC)

::::::::Is it really to the detriment of other editors? It redistributes an existing burden (annoyingly long wait times for nominations) by asking people to pitch in a little more to help make the process easier for everyone. I think that's in keeping with the spirit of the encyclopedia. And besides, if someone doesn't want to review nominations, they can just choose to not to nominate them at GAN - maybe they'll save them for FAC nominations, or just keep to writing quality articles without needing the fancy little green circles; that doesn't really impact one's on-wiki experience. Generalissima (talk) (it/she) 04:25, 21 June 2025 (UTC)

:::::::::This whole thing reducing participation at GAN is exactly one of my concerns, though. As much as collaboration is within the "spirit of the encyclopedia", I would say that forcing other editors into things, or gatekeeping a process, is not. I already made the point that "GAN and FAC effectively have a de-facto QPQ anyways", in that if you want your nominations to be reviewed quicker, then you review others (or at least, again, ones review-to-nomination ratio impacts my thoughts on whether to review theirs). That incentivizes users to give back, but it doesn't force them to do so if they aren't comfortable with doing reviews. And in my opinion, that is how it should be done. Bureaucracy is not needed in the GAN process. This is, at best, a double edged sword. But the negatives simply outweigh the positives to me. λ NegativeMP1 04:37, 21 June 2025 (UTC)

:I'm in favor of a QPQ, but I agree with TBUA that one of the main reasons the backlog grows is that GAN reviews are quite difficult. While I think there's been a bit of prose creep (I've seen quite a few people do FAC-level prose reviews even when this isn't requested), but I think the biggest clog is spot-checks, which are much more strict than FAC! For reference, FAC only strictly requires a spot check on a user's first nomination; otherwise, you just need to review source formatting. Now, I know why this was implemented for GAN, but I wonder if there isn't a middle ground between "require thorough spot checks on every single GAN" and "never check anyone's work if they seem reliable". Maybe a user's first nomination would require a spot-check, and then again for every Xth nomination after that?

:Ooh, or: Maybe have spot-checking be a separate task from normal reviewing that one can do to gain GAN QPQs; so you have the normal reviewers looking over the typical GA criteria stuff, doing a source review without a spot-check, and then sometimes you'd have a spot-checker butt in to see if you're doing a good job. This would recognize that spot-checking is pretty objectively more difficult than the rest of the review process (especially because it requires accessing sources to begin with), so would give people incentive to do it instead of prose reviews if they're not the type to do good prose reviews. Generalissima (talk) (it/she) 03:50, 21 June 2025 (UTC)

:I thought more about how this would work in practice and I think I have some ideas for implementation:

:*In addition to regular reviews, which would be the same as normal except for not requiring spot checks, users could choose to join a currently open review and spot check it, or just do a spot check alongside their review. Articles would not need a spot check, but if one gets opened on your nomination, you're still required to fix any found errors to get your nom promoted.

:*Each spot check would give a QPQ, equivalent to that from other reviews. This means that if you do a review AND a spot check on the same nom, you can "double dip" and get two QPQs from the same article.

:* Instead of the current 'Nominations/reviews ratio' displayed next to users on the nomination page, you would have a "nominations since last spot-check" counter, to encourage people to give users regular spot checks if they nom a bunch.

::*Perhaps if a spot check reveals a significant amount of issues, you'd have your next nominations flagged and required to receive spot checks to pass until you've gotten some quantity of clean passes. This way, Copyvios Georg wouldn't be able to just renominate poor-integrity articles over and over again until someone doesn't spot-check.

:*I think this would function to a) give people who enjoy spot-checks an alternative to prose-reviewing, b) make it way easier to do a bunch of reviews quickly (since spot-checks are generally the most time-intensive portion of the process), and c) still serve as Coldwellproofing, because no one is going to go that long without a spot check, and you'd have no control over which of your articles get checked. Generalissima (talk) (it/she) 04:10, 21 June 2025 (UTC)

::Broadly like the idea of lightening-but-not-eliminating spot checks, and more generally to bring the effort level below FAC reviews which are currently not as hard as GAN to do. If spot checks were separate, I can say I'd do them rarely if ever; I'm wondering how common that sentiment is, or whether finding a spot-check-reviewer would weigh a nom down more than they are currently. Besides that, I'd be far more hesitant to point out possible spot check issues if it meant "branding" the editor I'm working with as needing more frequent spot checks. One more thought is that checking sources can coincide with the rest of the review; it's often a read-through of the article that indicates where things might need to be checked the most. It's unclear how much of this a prose reviewer should be doing, and whether it should count toward spot checks if they're separated. Thebiguglyalien (talk) 🛸 04:35, 21 June 2025 (UTC)

:::I don't really understand the "spot checks are hard" thing. Unlike image reviews, they do not require any specialist knowledge. And you usually have to engage with the sources anyway when you check for broadness. So looking at sources is inseparable from reviewing the content. —Kusma (talk) 05:05, 21 June 2025 (UTC)

::::From personal experience: spot checks are absolutely the most monotonous and unfun part of reviewing for me. I really hate doing them and often procrastinate on them, and I think I'd do quite a few more reviews if I didn't have to. I've heard quite a number of peers express similar sentiments. Generalissima (talk) (it/she) 06:43, 21 June 2025 (UTC)

:::I think a big problem with the introduction of spot-checks has been that it's unclear how deep such spot-checks need to be, which has led (in part) to their recent intensification of workload. I've seen some reviews checking the majority of sources, and others checking only a couple sentences. The instructions say we should take {{tq|a sample of the sources}}, but gives no indication as to the sample size. --Grnrchst (talk) 09:11, 21 June 2025 (UTC)

::Changes should not be premised on there being an alternative to prose reviewing, as prose reviewing should be one of the lightest aspects of GAN. If it's understandable and without glaring errors, it's a pass on prose. Spot checks are the most time-intensive portion of the process because they are the core of the whole process, which at its heart is assuring that the article in question is a good example of WP:V. The spot checks also inform GACR3 on broadness and GACR4 on neutrality as you get some idea of the source positions. CMD (talk) 04:47, 21 June 2025 (UTC)

:::Prose checks being light is how it should function, but not how it currently functions. For a lot of reviewers, it appears to be the main part of their review. Thebiguglyalien (talk) 🛸 04:54, 21 June 2025 (UTC)

::::There's no restriction to reviewers adding additional prose points if they want (so long as they don't hold up the nomination over them), and it's natural that a reviewer might come across prose issues as they read. More prose review than necessary doesn't mean the process isn't functioning, so long as the other five criteria are also looked at. The point I was making is that we shouldn't create the expectation of prose reviewers, as heavy prose reviewing is very much an optional step (as [https://en.wikipedia.org/w/index.php?title=Wikipedia:Good_article_criteria&diff=prev&oldid=1288630286 PMC added to the GACR last month]). CMD (talk) 06:35, 21 June 2025 (UTC)

:Been reading this for a bit, and while I'm not sure QPQ is what's needed (it would certainly be a dramatic solution), I think something clearly needs to change. Speaking personally, it has been the glacially slow pace of reviews that has kept me from submitting more articles to the process. I am continuously mindful of the backlog and I don't want to contribute to it getting worse, so often when I improve articles, I never end up bringing them to GA. I've now even got a section of my user page for dozens of near-GA-level articles that I never nominated (and might not ever).

:People say that some of the proposed measures will be discouraging to article nominators, or lead to worse reviews (based on a complete absence of evidence, because we haven't even tried doing this to see if it works)... but the existing state of things is already discouraging to nominators. It is restricting the pace of article improvement by ensuring nominators have to wait up to a year for a review of an article or subject space they might have moved on from since nominating. This ends up being a vicious cycle, in which a long numerical backlog of nominations creates an even longer invisible backlog of articles people never end up nominating.

:QPQ might not be somewhere we want to go for a permanent change, but I wonder if it might be more acceptable as a temporary measure imposed when the backlog gets too long. At DYK, when the backlog gets too long, experienced nominators are required to review 2 other nominations (rather than the usual 1) to help bring it down, and this regularly helps to solve the problem. We could trial imposing QPQ measures when the backlog gets over a certain length, or even when the backlog is growing instead of shrinking, and see if it contributes to a more lively project culture. --Grnrchst (talk) 08:44, 21 June 2025 (UTC)

::The difference between limited-time QPQ and a backlog drive is that QPQ forces unwilling people to do reviews, while backlog drives encourage willing people to do more reviews by clever gamification. I would rather have my articles reviewed by someone who enjoys reviewing than by someone forced to do it; not sure whether I'd be motivated to nominate anything under a QPQ requirement. I do admit getting a bit impatient about my oldest nom, so I put it up at WP:GARC (and reviewed another article) to speed things along. —Kusma (talk) 09:25, 21 June 2025 (UTC)

Pulling out a comment from IAWW above, since I don't think it got much attention. I am not sure yet if I would support QPQ at all, but IAWW's suggestion was to apply it only above a certain level of backlog, and I think that deserves another look. To make this more specific, suppose we say that if the backlog is more than 750, then any new nominations are hidden (perhaps expandable, as the current hidden sections; perhaps not, so totally invisible) unless the nominator has at least a breakeven review ratio since a given date? Would that be a compromise between those who dislike QPQ and those who think it's now necessary? Mike Christie (talk - contribs - library) 11:53, 21 June 2025 (UTC)

:I would support this, on the condition that the QPQ requirement does not apply to new nominators. I think we should go further and say that the backlog should be significantly lower (300? 400?). I also think that they should not be in collapsible sections, but should instead be completely hidden. SSSB (talk) 12:01, 21 June 2025 (UTC)

::Yes I would agree 750 is too high – I think 200–300 is ideal. I also think they should be completely hidden and it should not apply to new nominators (<5 GAs). Though, what if these hidden nominations still grow faster than the backlog gets cut? I could see this happening as users with large review deficits continue to just write and nominate. IAWW (talk) 12:07, 21 June 2025 (UTC)

:::Which is why "nominator has at least a breakeven review ratio since a given date?" (My emphasis) is so important. It wouldn't be a poor ratio from the start of time (that wouldn't be fair). It would be a poor ratio from when this "rule" comes into effect. Therefore, there would still be incentive for those editors who may have a poor ratio now. SSSB (talk) 12:13, 21 June 2025 (UTC)

::::Sorry I should have worded that better, I meant that I think people with large review deficits will continue to operate at a large deficit despite the changes, so I think the hidden backlog may still increase faster. IAWW (talk) 12:37, 21 June 2025 (UTC)

::::In terms of fairness, counting the ratio since date X rather than ratio overall also means that somebody who has a good overall ratio but is currently nominating more than they are reviewing gets counted the same as somebody who has always nominated more than they have reviewed, which doesn't seem fair. This might be the case if, for instance, they have a habit of alternately reviewing and nominating in periods (say, reviewing 20 nominations, then nominating 10 articles, and so on). I think the most fair approach would be to calculate both ratios and use whichever one is more favourable. Opinions may differ on how well QPQ works for DYK, but a key part of why it has any chance of working at all there is that QPQs do not expire. TompaDompa (talk) 17:12, 21 June 2025 (UTC)

:::::Counting all old reviews isn't fair either (GA criteria have evolved). Looking at something from the last 3 months to 2 years probably works. —Kusma (talk) 18:08, 21 June 2025 (UTC)

::::::Would it be overcomplicated to say that we can do all time if the all time is better than 1:1, and only do a time frame if they don't have 1:1 or better all time. SSSB (talk) 18:47, 21 June 2025 (UTC)

:I think this is a great compromise. I personally don't think it will be strong enough to reduce the backlog, but I think it's way better than doing nothing, and compromise is clearly needed here. IAWW (talk) 12:02, 21 June 2025 (UTC)

:Hiding is better than disallowing; this is far preferable to mandatory QPQ. Those who feel like they really can't review (for whatever reasons) could just wait it out and maybe get their noms reviewed anyway. The numbers may all need some tweaking (perhaps after a few months). —Kusma (talk) 12:14, 21 June 2025 (UTC)

:If ChristieBot could handle the necessary calculations, that would be an excellent solution in my view. Perhaps we could work to simultaneously rework Wikipedia:Good article reviewing guide to replace WP:RGA and WP:GANI, and make the expectations on nominators and reviewers clearer? ~~ AirshipJungleman29 (talk) 12:31, 21 June 2025 (UTC)

::I don't immediately see anything that ChristieBot couldn't handle, but let's see what the specifics look like if this idea gains more traction. The "start date", and the backlog trigger size, could both be hardcoded in ChristieBot, but they could also be set on a protected page somewhere so any admin or template editor could update them. Mike Christie (talk - contribs - library) 14:21, 21 June 2025 (UTC)

:Sounds like a good compromise to me; at the very least its something we can try out and see if it improves the situation. Generalissima (talk) (it/she) 17:42, 21 June 2025 (UTC)

:I'm late to this whole conversation but I do want to voice my support for this idea. IntentionallyDense ^(Contribs) 18:53, 21 June 2025 (UTC)

:I feel like this is even worse. Lowering the bar to 200-300 nominations and requiring QPQ until we reach then is basically an unwinnable battle. As I already said: "{{xt|Not everyone is good at reviewing articles, and that's okay. Users should not be forced to do so. That's how you get half-baked or not-baked-at-all reviews. Ones done not because they want to, but because they have to. And over time, I could see it turning into a cycle that could heavily tank their motivation}}". Now combine this with how it would pressure editors more if we ever do get below that threshold. Sure, it might not be required to review at that point, but what if editors are too scared to nominate out of fear of reaching that threshold? What if they fear that, if the threshold is ever hit again, they'll be blamed? λ NegativeMP1 19:11, 21 June 2025 (UTC)

::I think social pressure to review does more good in reducing the backlog and all the associated benefits than it does harm in affecting the motivation of deficit reviewers. What do you think is more harmful: decreasing the motivation of experienced deficit reviewers, or having a new nominator not become a regular GA-quality contributor because their first nomination takes over a year to be reviewed? IAWW (talk) 19:33, 21 June 2025 (UTC)

:::I'm sure a hypothetical new nominator would rather wait for a review than receive one from someone that's not reviewing because they want to, but because they're forced to. Let alone risk getting one that's low quality and could possibly even go to GAR, which would over-complicate everything for them. The possible consequences would likely drive away a new editor more than the long wait times. Also, I'm 99.9% sure that doing this would cause both long wait times and decreased editor motivation. If an editor is really uncomfortable with reviewing, like I have said many times already, more likely than not they just... won't participate in GAN at all anymore. Preferably, neither problem should be happening, but I'd rather not cause a second problem simply because a first problem exists. λ NegativeMP1 19:47, 21 June 2025 (UTC)

::::By this argument, you are against any form of social pressure to review, which is quite an extreme view to be 99.9% sure about. I personally don't think social pressure or QPQ will cause experienced editors to not review properly very much. I think most experienced editors want the encyclopedia, and therefore the processes in it, to succeed. IAWW (talk) 20:01, 21 June 2025 (UTC)

:::::I am not against any form of social pressure to review, though. To quote what I've also said prior: {{xt|"GAN and FAC effectively have a de-facto QPQ anyways", in that if you want your nominations to be reviewed quicker, then you review others (or at least, again, ones review-to-nomination ratio impacts my thoughts on whether to review theirs). That incentivizes users to give back, but it doesn't force them to do so if they aren't comfortable with doing reviews. And in my opinion, that is how it should be done.}} What is being proposed right now is not just mere social pressure, it is forcing. λ NegativeMP1 20:04, 21 June 2025 (UTC)

:::::If I'm expected to review when I don't want to, my reviews are going to be more cursory than if it was something I was interested in doing. The question is whether that's a bad thing, given that good enough is the whole point of this process. Thebiguglyalien (talk) 🛸 20:12, 21 June 2025 (UTC)

:If a number is selected, I'd support a higher number that we can meaningfully achieve like 700. Then if this rule proves successful in a permanent reduction, we can lower it in the future to 600, and so on. Baby steps. Regarding ratios or any other rules the nominator needs to be aware of, the simpler the better. Every additional layer of complexity will decrease the number of people who feel encouraged to participate. Thebiguglyalien (talk) 🛸 19:36, 21 June 2025 (UTC)

:I am not categorically opposed to trying this as a one-off experiment to evaluate, but if we do that I think it is necessary to define in advance a date at which the experiment is ended regardless of the outcome. I would suggest that this date be no later than 3 months after the start of the experiment. TompaDompa (talk) 20:21, 21 June 2025 (UTC)

There are enough positive comments above for me to think this might get support, so I've thought a bit more about how the algorithm could work. Here are some thoughts, just to make this specific -- this is not a proposal, just a way to make it clear how this could work.

One thing to consider is what to include in the backlog size that triggers this rule. Say we set that to 700, per TBUA's suggestion above. Call the nominations that might be rendered invisible by this rule the "hideable" noms, and the others the "visible" noms. There are a couple of odd side effects that we probably don't want, and which we'd have to decide how to avoid. First suppose that the backlog size trigger only includes visible noms. Then if the number of hideable noms is, say, 100, whenever the backlog of visible nominations falls to 700, all 100 invisible nominations will suddenly appear, expanding the number of nominations to 800. Alternatively if the trigger looks at the total number of hidden plus visible nominations, then if there are 100 hidden nominations the number of visible nominations will have to fall to 600 before any hidden nominations appear, and the backlog will jump to 700. I think the way to resolve this would be to use the total (hidden plus visible) to trigger the rule, but to allow up to 700 to display, hiding only as many "invisible" nominations as necessary to bring the listed total down to 700 (or whatever the trigger number is). If there were more than 700 visible nominations, all of them would display, and no invisible ones would display.

The parameters could be:

• Size of the backlog that triggers the rule. I would suggest this includes articles under review, as well; otherwise simply flagging an intention to review, but not completing the review, would reduce the count, which I don't think is desirable.
• Day from which to start counting review/GA ratios. I think this would be set to the day we implement this, and would never change -- it wouldn't roll forward if we get below the backlog for a while.
• Flag to indicate whether we count all-time GA ratios and choose the most favorable, as suggested above.
• Ratio minimum to count as "visible" -- discussed as 1.0 above, but we could set this to e.g. 0.95 if we wanted to
• We could set a number of GAs below which a nomination is not subject to invisibility. For example, we could exempt new nominators by setting this to 0, or we could set it to 1 to excuse nominators with 1 GA from having to review.

My own suggestions, if we were to try this, would be to make the numbers as generous as possible while we see if it works. I would suggest setting the trigger size to about the same as the actual backlog at the time the rule goes into effect -- i.e. if we turned this on today I'd suggest 875 or 900. I think we should include whichever is most favourable -- all time review ratio or ratio since the rule went into effect. I think we should allow ratios slightly below 1.0, and I think nominators with only 1 GA, or no GAs, should be exempted from invisibility. Mike Christie (talk - contribs - library) 20:51, 21 June 2025 (UTC)

:I think it reasonable to allow up to 3 GAs until editors are expected to review (that is, they would be expected to start reviewing once they reach 3 GAs), since I recall that being about the time it took for me to get a decent hang of it. I suggest setting the permissible ratio at 0.90 or 0.95. TompaDompa (talk) 21:20, 21 June 2025 (UTC)

• April–May 2020: unreviewed 605->238
• October 2020: unreviewed 552->285
• March 2021: unreviewed 579->273
• July 2021: unreviewed 464->265
• January 2022: unreviewed 462->165
• June 2022: unreviewed 520->366
• August 2023: unreviewed 638->198
• March 2024: unreviewed 655->399
• July 2024 ?
• October 2024 ?
• January 2025: unreviewed 653->468, old 178->60
• May 2025 ?

I think there is evidence that generic backlog drives work and (used to) regularly get us below 200 (I don't actually think we should aim below 100), although recent drives have been somewhat less successful. I don't quite know what to think about the focused ones, I couldn't immediately see the full data. —Kusma (talk) 06:33, 21 June 2025 (UTC)

:Having participated in the most recent 3 backlog drives and helping organize some as well, I will say that from my viewpoint, the January backlog drive was the most successful of the recent 3. There was only 3 of us reviewing submissions and while one of the coordinators would check every morning, i still was reviewing probably up to 10 reviews every single night. I think this was a combo of the wikicup and the way that the drive worked points wise. I think the backlog drives focused on older nominations are extremely valuable (as are the newbie focused ones imo) and should possibly be increased in frequency (as in replacing the one unthemed one with the one focused on old noms).

:However I will also say that we need more people willing to coordinate these drives. By chance, this month both me and another coord had some life stuff that started going down about halfway through. it was quite overwhelming for me personally, although I can't speak for anyone else, i definitely felt like we could have had a more successful drive if we had more active coords (not that the coords I worked weren't amazing, life just happens sometimes). ideally we should have 3-4 active users as coordinators, that way if 1 or 2 need to take a step back it doesn't fall into one users hands. I'll tag {{u|Vacant0}} in case he has anything to add here. IntentionallyDense ^(Contribs) 19:03, 21 June 2025 (UTC)

::Hi @IntentionallyDense, I never got to thank you for playing such a big part in organising last month's drive. It's much appreciated. I am happy to co-ord drives in the future. IAWW (talk) 19:10, 21 June 2025 (UTC)

:::I was actually planning on reaching out to you about that as I see how driven you are. Ideally there should be enough coords that the coords can also participate in the drive itself. IntentionallyDense ^(Contribs) 19:14, 21 June 2025 (UTC)

::One of the concerns about doing more frequent backlog drives when it came up in the big discussion in 2023 was that we don't want to exhaust the backlog drive participants. So January being the most successful one is the system working as intended. (Whether the system needs to be changed is its own question, but it is functioning as designed.) I don't think increasing frequency is a good idea since we're already at three per year. More co-ords is an easy fix. Maybe what would help more than simply "more co-ords" is to appoint a backlog drive co-ordinator - as in, someone who is expected to do the general co-ordination, make sure things are set up on time, make sure each individual drive has enough co-ordinators, etc? -- asilvering (talk) 19:34, 21 June 2025 (UTC)

:::I see @It is a wonderful world says above that they found the most recent backlog drive dispiriting, because the backlog went up. Certainly, that's dispiriting. But if we continue to agree that three full backlog drives a year is too much on reviewers (and I think it is, myself), that's going to keep happening, probably. So if we keep/tweak this system, we probably need to also get better at managing expectations. I tried to make it clear with the names of the drives when I was co-ordinating them, naming them things like "backlog elimination drive", "newbie drive", etc, but we can do better than that if we rename the whole concept something like "Reviewing drives", and then name the individual drives something like "Backlog reduction drive", "New reviewers drive", "Old nominations review drive", etc etc. Those three are the ones I came up with for 2024 - but there's no reason we couldn't do more smaller "flash drives" in and around those, or switch to some other system like a bi-yearly "Big backlog drive" and smaller "flash drives" spaced throughout to target something specific (eg, all the history nominations, newbie reviewers, all the 90+ day nominations, etc). -- asilvering (talk) 19:48, 21 June 2025 (UTC)

• [https://wikicup.toolforge.org/index.php?year=2020 2020]: 556 GA, 892 reviews
• [https://wikicup.toolforge.org/index.php?year=2021 2021]: 470 GA, 650 reviews
• [https://wikicup.toolforge.org/index.php?year=2022 2022]: 349 GA, 683 reviews
• [https://wikicup.toolforge.org/index.php?year=2023 2023]: 397 GA, 545 reviews
• [https://wikicup.toolforge.org/index.php?year=2024 2024]: 408 GA, 623 reviews
• [https://wikicup.toolforge.org/index.php?year=2025 2025 (incomplete)]: 460 GA, 481 reviews

So the WikiCup's usual beneficial effect on the backlog (to the tune of about 200/year) has not materialised yet this year. There are discussions about higher scores for GA reviews in the Cup; perhaps that would have an effect on the backlog. —Kusma (talk) 07:07, 21 June 2025 (UTC)

:This is very interesting - both because of the GA review issue, but also because the numbers are on track with the usual ones, even though we're only halfway through the year. So this seems to me to be quite likely to be a major driver of our backlog problems. Wonder what happened in 2020 and how we can steal that energy back. -- asilvering (talk) 19:39, 21 June 2025 (UTC)

There seems to be agreement that an increase in reviews would, ceteris paribus, be a good thing. Other than that, it is not clear to me that we even desire the same outcome—that is, it is not clear that we are looking at the same metric(s) to assess how things are going. Moreover, I think we have to decide what we want to happen if reviewing remains constant—that is, if we assume that we cannot positively affect reviewing habits (which we should of course nevertheless keep trying to do), what is the next best option? And thirdly, what costs are we willing to incur in trying to improve the situation—or put another way, what downsides would we find acceptable (and equally importantly, which would we find unacceptable)?

Possible metrics:

1. Size of the backlog (number of nominations)
2. Time from nomination until review (be it median, mean, or some other measure)
3. Throughput (reviews per unit of time)
4. Pass-specific throughput (new GAs per unit of time)

Possible second choices if reviewing habits remain fixed:

1. Having a sizeable backlog (the status quo)
2. Reducing the number of nominations across the board
3. Reducing the number of nominations from prolific nominators (who consume a large portion of the available reviewing resources)
4. Reducing the number of nominations from nominators who do not review much (and thus "take out more than they put in")
5. Reducing the number of nominations from nominators who do review much (and whose time would thus arguably be better spent reviewing than nominating)
6. Reducing the number of nominations from less experienced editors (whose nominations might be expected to be of lower quality)
7. Reducing the number of nominations in specific topics (in case some topics contribute more to the backlog than others)

Potential downsides of actions we might take:

1. Decreased nominator-side participation (fewer editors nominating articles)
2. Decreased reviewer-side participation (fewer editors reviewing nominations)
3. Decreased nominator-side engagement (each editor nominating fewer articles)
4. Decreased reviewer-side engagement (each editor reviewing fewer nominations)
5. Decreased quality of reviews

Feel free to add anything I have missed. TompaDompa (talk) 18:21, 21 June 2025 (UTC)

:@TompaDompa, thanks for laying this out like this. I think agreeing on the metrics we care about the most is more likely to get us somewhere useful than our usual generalized concerns about the backlog. -- asilvering (talk) 19:36, 21 June 2025 (UTC)

::In my opinion, the size of the backlog is almost completely irrelevant. If we have a persistent backlog of 2000 articles, but each nomination is reviewed within a month, that would be an extremely efficient and desirable situation. Time from nomination until review is the big problem. -- asilvering (talk) 19:50, 21 June 2025 (UTC)

:::This is the correct answer. Its never been about the amount of throughput - arguably more nominations is better. If theres a crazy high backlog, but it gets done really quickly, then theres no issue. Im not even a fan of the word "backlog", as it suggests that we should be in a state where there are no open nominations.

:::What we need is a quicker time from nom to review across the board. Lee Vilenski ^{(talk • contribs)} 19:54, 21 June 2025 (UTC)

:::{{ec}} I entirely agree about the size of the backlog being largely irrelevant; I think the main reason it is brought up is that it is both easy to quantify and highly visible. I'm not entirely sure about time from nomination until review being the big problem, however: if it instead took 100 days (so, slightly more than 3 months) to review each nomination, but 150 nominations were reviewed each day, we would have a persistent backlog of 15000 articles but roughly twice the throughput of your example (30 days until review multiplied by 70 reviews per day would equal a persistent backlog of 2100 articles). It is not clear to me that this would be worse. TompaDompa (talk) 20:04, 21 June 2025 (UTC)

::::@TompaDompa, in my opinion, that is worse, since the overall, absolute numbers - the size of the backlog, the number of reviews per day, etc - are all things that would change simply based on how many people were involved in the project. From the perspective of the people participating in the project, time to review is what you're subjectively feeling. I think that subjective feeling is more important to target than the size of any of the numbers. -- asilvering (talk) 21:03, 21 June 2025 (UTC)

:::::Alright. That means that the goal of WP:GAN is not really about the articles (whether article improvement, quality control, or recognition of quality content) but rather editor enjoyment, I suppose. TompaDompa (talk) 21:10, 21 June 2025 (UTC)

::::::Well, improving Wikipedia relies on sustaining editor motivation. GA seems to me to function primarily as a driver of motivation: nominators are motivated to improve articles by the 'reward' (such as it is) of getting the article to GA status (recognition of their production of quality content). Obviously the goal of motivating the nominator is article improvement, and the goal of the review is to ensure that recognition reflects meaningful improvement in article quality. When the reward/recognition is excessively delayed, the link between editor effort (getting articles ready for review) and reward is weakened, i.e. GAN as a motivator becomes less effective. So yes, wait time really matters. YFB ¿ 21:35, 21 June 2025 (UTC)

:::::::Sure. The purpose of everything we do here is ultimately improving the encyclopedia. Sometimes that is very direct (content creation, content improvement, and so on), and sometimes it is indirect (editor retention, editor encouragement, and so on). The question here is where on that sliding scale the GAN process falls: is it meant to in itself improve the encyclopedia, or is it meant to have accomplish that indirectly through the effect it has on the editors? I think the DYK process fairly clearly falls way more toward the editor side than the content side, whereas the FAC process leans substantially more towards the content side. I view the GAN process as being more content-focused than editor-focused, but I may be in the minority in that regard. TompaDompa (talk) 21:47, 21 June 2025 (UTC)

::::::::The sliding scale metaphor is a reasonable one, and it seems intuitive that GAN is more editor-focused than FAC on that scale, since GA is explicitly framed as a more achievable milestone than FA (and for some editors it's the interim reward that provides a motivational boost to go all the way to FA). When GA works well it does directly improve the encyclopedia. But when it has a very long wait time it is not working well, just like all the other real-world contexts in which long queue lengths result in dissatisfied humans. YFB ¿ 22:00, 21 June 2025 (UTC)

::::::No, @TompaDompa, I'd disagree with that statement as written. But when it comes to "what metric matters when we're talking about reducing the backlog", yes, it's the subjective feelings of the editors involved that matter, imo. I don't think anyone would be complaining about the backlog if it were simply large. The problem is that it feels hopelessly delayed. -- asilvering (talk) 21:51, 21 June 2025 (UTC)

:I personally think cutting the "Time from nomination until review" by "Reducing the number of nominations from nominators who do not review much" is the biggest net positive change we could make without changing reviewer habits. IAWW (talk) 19:46, 21 June 2025 (UTC)

:Time until review is the problem, total number of nominations awaiting review is both a cause and effect of this. More nominations needing review means it may take longer and longer to get to each one, and taking longer to get to them means that more and more nominations will be waiting as they come in. Under the current system, either can be used as a valid metric. Thebiguglyalien (talk) 🛸 20:10, 21 June 2025 (UTC)

::I think it is important to note that if throughput changes, the effective interchangeability of those metrics breaks. I don't know if we have data on throughput over time. TompaDompa (talk) 20:16, 21 June 2025 (UTC)

:::We do -- I have over 67,000 records in a historical table that includes all the GAs I've been able to parse. It includes nomination, review, and outcome timestamps. I can make the data available if you are interested. Mike Christie (talk - contribs - library) 20:55, 21 June 2025 (UTC)

::::I am. I think you are probably better positioned to select the most relevant metric, but I suppose something like number of finished reviews per month might be enlightening (and not subject to too much noise in the data)? It may also be interesting to see how many of those are passed versus failed. TompaDompa (talk) 21:14, 21 June 2025 (UTC)

::::I'd be interested in that dataset too @Mike Christie, and yes pass/fail outcome for each row would be beneficial if available. YFB ¿ 22:09, 21 June 2025 (UTC)

Just as a heads up, the Wikipedia:Developing Countries WikiContest is being planned to run from 1 July to 30 September. The scoring rules are not available yet, but if it is like last year then GA reviews will be eligible for points. We had a couple of issues with new editors reviews during last year's competition, but the coordinators took this into account then and I'm sure they remember, and I'm sure they'll be helpful if we find a new reviewer needs to add a bit more to the reviews.{{pb}}Another aspect is that of course if anyone is reviewing relevant articles here, they can double dip for DCWC participation. It helps fill out the overall combined improvements. Perhaps we could also consider separately giving some sort of GA review barnstar to those who use the opportunity to work on reviews. CMD (talk) 04:21, 20 June 2025 (UTC)

:The new scoring has come out, and they've increased the points for a review to 10 (from 5 last year). I created a huge list of every GAN that seemed to meet the inclusion criteria, hopefully it gets picked up. CMD (talk) 07:12, 21 June 2025 (UTC)

There's a new pilot program open at Wikipedia:Resource support pilot, where editors can submit requests for the WMF to buy sources for them. I encourage folks to check it out, and notify any WikiProjects and editors that may be interested. Apologies if you've seen this elsewhere already. Toadspike [Talk] 07:25, 20 June 2025 (UTC)

:man, I wish this existed before I bought those cue sports encyclopedias. Lee Vilenski ^{(talk • contribs)} 21:03, 21 June 2025 (UTC)