Wikipedia talk:Good article nominations

MainCriteriaInstructionsNominationsFAQDiscussionReassessmentReport

Reviewing initiativesJune backlog driveMentorshipReview circlesPledges

Good article nominations

This is the discussion page for good article nominations (GAN), good article reassessment (GAR), and the good articles process in general. To ask a question or start a discussion about the good article nomination process, click the Add topic link above. Please check and see if your question may already be answered; click the link to the FAQ above or search the archives below. If you are here to discuss concerns with a specific review, please consider discussing things with the reviewer first before posting here.

ChristieBot error reporting -- where to look

I've recently made some changes to the way ChristieBot handles errors. Since I know there are some editors who help out by fixing various kinds of GAN error, here's a summary of how it is now supposed to work, and what can usefully be monitored. My hope is that this is a lot simpler and easier for editors to monitor than it used to be.

There are five different ways that errors can show up:

• Individual GAN entries on the main GAN page. Some issues will appear directly next to a nomination entry (in red text). These come from the {{GANentry}} template and may be further tweaked in the future. These usually indicate something editors can fix on that specific nomination.
• Article talk page. Thanks to Prhartcom, who has enhanced the {{GA nominee}} template, if there are errors in the GA nomination template they will now show up on the article talk page in bold red text. These are typically formatting or parameter issues that the nominator (or any editor) can correct. Note that you can also watchlist Category:GAN error to see these, since anything that creates an error in the GA nomination template will put the talk page that category.
• GAN errors page. User:ChristieBot/GAN errors will list all current nomination/template issues. These will be similar to the problems identified on the GAN page and the article talk page, per the above. If you're interested in helping fix GAN errors this is a good page to watch. For example, it will list nominations with missing or invalid parameters, and nominations in an inconsistent state. Note that the "Errors" section on the GAN page has been eliminated. All editor-facing errors should be listed on here; and if the edit summmaries on the main GAN page start with "Errors listed!", that will be a link to this page.
• ChristieBot bug messages page. User:ChristieBot/Bug messages will list internal bot errors (unexpected failures, database issues, etc.). These are not generally fixable by editors, but may be useful for awareness or reporting. Most editors will not find it useful to watch this.
• Operational status page. User:ChristieBot/Operational status will show a summary of the operational state. Sometimes there are temporary system issues that are not bugs or GAN errors; this page will (when possible) show the status. This page is transcluded onto ChristieBot's user page. This may be worth watching if you want to know, for example, why ChristieBot has not run in a few hours; this might tell you why.

This separation is intended to make it clearer which issues might benefit from editor attention and which are informational only. If you see any errors that don't fit this pattern, or if there are any questions, let me know; it's entirely possible I have not yet converted all the messages to fit this plan. Mike Christie (talk - contribs - library) 01:23, 28 March 2026 (UTC)

Mike Christie, thank you for improving the error handling and reporting in ChristieBot and honored to work with you on the GAN templates. OK, good plan: Watchlist Category:GAN error, User:ChristieBot/GAN errors, User:ChristieBot/Bug messages, User:ChristieBot/Operational status, as well as vigilance for any red text appearing on the main GAN page or on the article talk page. Prhartcom (talk) 21:08, 30 April 2026 (UTC)

2012 emo killings in Iraq

2012 emo killings in Iraq was assessed as a Good Article ... and made its way to AfD. It was considered notable - the title is. It was nominated for Did You Know where huge errors were found. I changed the lede to reflect that the article misrepresented and confused different events under the wrong titles. Its bad. I was assured that it was going to be merged... but nothing has happened. I am concerned that it says its a Good Article and it is not even close. No idea how to create a reassessment/major edit.... but this comment might help. 2012 emo killings in Iraq is not a Good Article. Please help. Victuallers (talk) 11:27, 16 August 2026 (UTC)

I sent it to GAR for you: Wikipedia:Good article reassessment/2012 emo killings in Iraq/1 Bgsu98 (Talk) 12:17, 16 August 2026 (UTC)

What should ChristieBot do if the GAN page number goes down instead of up

I'm working on some related bugs, and have a question. On Aktion AB, the review with page=1 was closed by the bot because the nominator incremented the page number by 1 (per 4a in the GAN instructions). So review 1 was failed by the bot automatically, and review 2 was sitting on the talk page as an open nomination. Then this edit was performed. This took the open GAN nomination with page=2 and changed the page number back to 1, presumably because the nominator was under the impression that the original review was going to take the review back up again. But the bot now thinks that the current review number is 2, so this caused the bot to believe there were two active nominations. (To be clear, I'm not blaming the nominator; their actions were logical, but the bot currently doesn't know what to do when this happens.)

The bot could fairly easily decide to fail review 2, because there is now no nomination for it -- that's what the bot does if you go from an active page 1 review to an open page 2 nomination; it just silently assumes page 1 is a fail. (And in fact as noted above the bot did exactly that for page 1.) In principle it doesn't care that the number has gone down instead of up. If it had done that, though, the nominator would have received a fail message on their talk page for page 2, which would probably have baffled them. And what should the bot do with the new "1" nomination, which it has previously silently marked as a fail in the database? I don't really want to implement a "reopen old review" function. Given how the bot is constructed, the easiest thing for me to do is probably to make the bot report an error and stop processing that nomination in any way till the nomination is fixed. But what does "fixed" mean here? Does it mean "change the nomination number to 3"? (Any answer that requires me to clean up the database's inconsistent records before the nomination can proceed is obviously bad since I'm not always quickly available.) Mike Christie (talk - contribs - library) 13:11, 16 August 2026 (UTC)

As the nominator who did this (because, after over a month since they last edited, I had despaired of my reviewer returning and, per instructions, closed GA1 and relisted the review, only for them to return two days later and continue using GA1 even though I'd told them to open GA2 if they came back ... apparently I'd forgotten to add the "failed nomination" archiving template, I think because it's not in the instructions for what to do when you think a nomination has been abandoned) I think that maybe if we were able to add some field to the GAN template that told the bot not to autofail the nomination, that might help you and us best.

I had had another situation with this nom where this happened that might be worth seeing if we could do something about: A couple of months ago, an issue raised during the review led me to decide to rename the article to be consistent with how we name articles about similar subjects. This resulted in a message on my talk page that the nomination had failed because the nom (apparently) no longer linked to the original page. I disregarded it and the nomination remained active, but given the likelihood that this has happened and will happen again perhaps some way should be found for the bot to able to take that into account. Daniel Case (talk) 17:34, 16 August 2026 (UTC)

The new parameter idea is interesting, but even if we update the instructions there will still be cases where people make the same edits you did, so the bot still needs to know what to do. At the moment the best idea I have is that it autofails the page=2 review, with a reviewer of "None", and then puts up an error message saying "Nomination for page = 1 but that review has already concluded -- use page=3 for a new review", or something like that. Perhaps the autofails in these cases should also add the FailedGA template, to make it clearer what is happening. Mike Christie (talk - contribs - library) 19:00, 16 August 2026 (UTC)

The main issue with this error is that it is inscrutable unless you have access to the bot. I wonder if there's a clearer error message that can be created. CMD (talk) 22:47, 16 August 2026 (UTC)

Agreed -- it's not at all clear what happened, just from the error message. I can fix that easily enough, but at the moment it requires me to clean out the database before the error goes away, and I'd like to change the behaviour as well as the message, so the database never gets screwed up. If I don't hear a better idea I'm going to go ahead and implement the fix I outlined above, probably some time this week. Mike Christie (talk - contribs - library) 23:01, 16 August 2026 (UTC)

Update: I have implemented some changes that I think will reduce some of the sources of errors we're seeing. The bot is now going to do the following:

• If the page parameter in the GA nominee template is changed, without terminating the review, the bot will determine if the new page number is valid -- that is, is it equal to 1 more than the highest number for this page in the database?
  1. If it is valid, the bot will fail the previous review, and add the FailedGA template. It will update its database so that it knows the new review page is the current one. It will write FailedGA to the talk page. The talk page of the nominator will get a "failed" message. The nomination will remain open on the GAN page, with the new page number -- note that this will keep the old timestamp, which could be seen as jumping the queue. If that's not acceptable, I could generate a warning message to WT:GAN that the nomination has a nomination timestamp preceding the close timestamp of the previous nomination. This would be a one-time message to let GAN regulars know that someone might be jumping the queue.
  2. If it is not valid, the bot will post an error and do nothing else at all -- it won't remove the nomination (under the old page) from the GAN page, and it won't fail it. It will still be possible to click on "Start review" under the old page number, but the bot won't do anything with any actions -- it won't transclude the review or remove it from the talk page or record it as passed or failed. When the page number is set back to either the current page, or incremented by 1 from the current page, it will carry on from that point.

Daniel, the bot should no longer notify you of a fail if you rename an article mid-GAN review. I made that fix earlier this year; it's hard to test properly but I think it's fixed. If you (or anyone) sees that happening again, let me know. Mike Christie (talk - contribs - library) 01:43, 18 August 2026 (UTC)

LLM reviews by Metalicat

Hi. There's an ongoing discussion over at Talk:Pat O'Keeffe#AI generated text over how yesterday's TFA was a violation of Wikipedia's LLM policies. Specifically, while looking into the edit history of fairly new editor Metalicat, I stumbled across a pattern of behavior reflecting that they used an LLM to essentially rubber-stamp a bunch of GA reviews. I'm requesting that, because GA reviews are performed by a single editor, these be systematically undone and renominated. While I recognize these have been up for several months, I think we all understand that doesn't mean these articles have been given proper scrutiny since.

To that end, I'm going to list a timeline of GA reviews, but peer reviews and FAC reviews will be included where temporally relevant. This relies heavily on the frequency and quantity of reviews to show 1) the pattern of misbehavior beyond a shadow of a doubt and 2) which articles were affected as it pertains to GA. These reviews individually have hallmarks of LLM usage, but I think the timeline already trivially proves this. (All of this is in 2026.) I want the reader to keep in mind both how disparate these subjects are and that the editor joined 28 December 2025 (with their very first edits being used to create what would become a FA).

• At 08:11 on 11 February, they add a +1912 peer review to Leave It All to Me.
• 31 minutes later, they perform a +2591 FAC review of Hurricane Joaquin.
• At 04:24 on 12 February, they publish a +4744 GA review on Carl Lawson (sprinter).
• 86 minutes later, they add a +5148 peer review to El Alma al Aire.
• At 10:52 same day, they publish a +4413 GA review to LRAShM (missile).
• Less than two hours later, they publish a +19,972 GA review to Israel Keyes.
• The next day, 13 February, at 06:37, they publish a +9266 GA review for 1976 San Diego Chargers season.
• 33 minutes later, they publish a +26,153 GA review of Jimmy White.
• On 16 February at 05:33, they publish a +5480 GA review of Hedycarya arborea. There were multiple edits in the hour before this, including a +2325 to the Jimmy White "review".
• At 04:52 on 25 February, they publish a +6339 GA review for 1990 United States Senate election in Rhode Island. Just 31 minutes before this, they had been editing Arthur Frederick Bettinson, and they would go on to drop a +10,767 to the Arthur Frederick Bettinson article 35 minutes after they published this GA review.
• On 3 March, they published a +5426 review for Milan Heimerl. This one would be completely inexcusable by some bizarre "Well clearly they were working on it" excuse, because their 28 February, 1 March, and 2 March were utterly packed with inhuman levels of content added to (especially) Pat O'Keeffe.
• On 11 March at 16:19, they publish a GA review for 2023 Capita data breach.
• On 17 March at 16:03, they publish a +4887 GA review for CIH (computer virus).
• On 24 March at 09:13, they publish a +7437 GA review for Urtica ferox. At this point, they've been heavily editing printing press from 21 March to 23 March.
• 22 April at 02:03, they publish a +1013 GA review for Chapter Three: The Case of the Missing Lifeguard. (This was followed by a +4755.)
• On 7 August at 16:08, they publish a +8613 GA review for 2026 Punjab Kings season. 46 minutes later, they publish a "source spot check" that's +4996. Then 12 minutes later, they publish "first pass complete" which is +4005.

In sum, the following GA reviews were hollow nonsense by someone who had zero idea what they were doing:

Carl Lawson (sprinter) • LRAShM (missile) • Israel Keyes • 1976 San Diego Chargers season • Jimmy White • Hedycarya arborea • 1990 United States Senate election in Rhode Island • Milan Heimerl • 2023 Capita data breach • CIH (computer virus) • Urtica ferox • Chapter Three: The Case of the Missing Lifeguard • 2026 Punjab Kings season

Lastly, articles they've submitted for GA review should be revoked for blatantly violating the policy against LLM usage. I might be missing some, so please double-check me. I don't think it should be on any editor to affirmatively nominate these "reviewed" articles for reassessment, because there was no assessment in the first place. TheTechnician27 (Talk page) 14:43, 21 August 2026 (UTC)

It doesn't seem like the user has admitted to LLM usage on any(?) of these. If you take Talk:LRAShM (missile)/GA1, the fact they did this four hours after other things doesn't really implicate anything, whether or not the work earlier that day used LLMs. Parts of that review feel LLM-written to me, but sometimes it's quite hard to tell the difference between a GA reviewer with no understanding of the GA process and a review written by AI. Personally, unless they admit to AI use or someone more clever than me can "prove" it for a given review, I'd probably prefer GAR. That's also because some of these are quite old and most reviews that are pulled are done so soon after it is completed.

Again, Talk:Israel Keyes/GA1 looks LLM-written to me but it could just be the style of the reviewer, especially if they're not an experienced editor. (in solidarity), JacobTheRox^{(talk|contributions)} 15:26, 21 August 2026 (UTC)

@JacobTheRox, in the discussion where I suggested that TheTechnician27 post to WT:GAN, I suggested that WP:GAR would be a better venue for reevaluating the status of GAs that have held that status for quite a while. There isn't really precedent for immediately invalidating GAs that have been approved so long ago, unless it's an instance where the nominator was at fault (e.g. WP:DCGAR). – Epicgenius (talk) 15:30, 21 August 2026 (UTC)

Looking further at Israel Keyes, I had to move the note to a proper section and two external links remain in the body, which could be seen as needing to be removed for GA. I have concerns about the possible and ruled-out victims section; firstly, this seems like a cobbled-together list where no single source gives a general overview that he could have committed those crimes (and thus that there are not more crimes he could have committed that have not been listed). Secondly, some of his links to these crimes seem tenunous and SYNTH-dependent. For example:

Hidden to save clutter but it's just me ripping apart a portion of the article

Julie Marie Harris, a 12-year-old Special Olympics medalist in skiing, disappeared on March 2, 1996, while waiting for a ride to a local church in Colville, Washington. Her remains were found in April 1997 in a wooded area nearby. Keyes, then 18, lived in the area at the time and was questioned about the case after his 2012 arrest. He neither confirmed nor denied involvement, stating only that he had "heard about that".

If he neither confirmed nor denied involvement, and the source doesn't say he was investigated, arrested, or charged for it, why is it being considered a crime he could have committed? He also has, presumably, at not point confirmed nor denied his involvement in the death of John Lennon.

Cassandra "Cassie" Emerson, 12, was reported missing in 1997 after the remains of her mother, Marlene Kay Emerson, were discovered in their burned-out trailer home in Colville, Washington. Cassie's remains were later found approximately 21 km (13 mi) away. Keyes admitted his first act of arson involved a trailer in Colville, and he was at one time considered a suspect in the Emerson deaths. However, in December 2024, Charles Tatom was arrested in connection with the murders.

The second source doesn't say he was a suspect, so it is unverified unless the first source is misplaced. Secondly, the fact someone was arrested in relation to the crime (1) doesn't mean they did it (2) doesn't mean he didn't and (3) is SYNTH.

Giovanna Tyler, a 29-year-old woman from Tacoma, Washington, was reported missing in 2004. Media coverage later discussed her case among those examined in possible connection with Keyes.

None of this is in the CBS News article.

Keyes has been suggested as a possible suspect in a series of unsolved 2007 crimes near the Town Center at Boca Raton in Florida. [...] However, authorities have not publicly confirmed Keyes was connected with those crimes.

If authorities haven't connected him, doesn't that make this a fringe theory? The two sources that aren't the inacessible book don't mention him either.

He has also been suggested as a suspect in the disappearance of James "Jimmy" Lamar Tidwell Jr., an electrician who disappeared in Mount Enterprise, Texas, on February 15, 2012. Authorities examined similarities between Tidwell's appearance and a bank robbery suspect in Azle, Texas, the following day, but no definitive link has been established.

References

I can't access the Weatherford Democrat article from the UK. However, once again, we are not told who is suggesting him as a suspect, and he was never arrested/charged with it, so this seems like a fringe take to me that is probably reliant on the book.

I'm not quite ready to put money on it, but I wouldn't be surprised if the inaccessible book in question doesn't provide any evidence that the authorities thought he committed many, or any, of these crimes. The reason I'm posting this here, rather than e.g. on the article talk page, is that it shows how poor these GA reviews must have been, and it's not a great look to have this article with the GA icon at the top. (in solidarity), JacobTheRox^{(talk|contributions)} 16:08, 21 August 2026 (UTC)

I might be missing something, but were the GA reviews generated without refinement by AI? That is what would be required to invalidate the reviews per WP:LLMTALK (I’m counting GA reviews as talk pages per WP:NOTBURO). If it can’t be proved that the reviews were LLM-generated without refinement, then I think it would be incorrect to reverse them unilaterally. It seems clear that the editor was using LLMs as a crutch, likely excessively, but while we do have regulations against LLM use they do have constraints embedded in them. ~~ AirshipJungleman29 (talk) 15:34, 21 August 2026 (UTC)

What kind of "refinement" would we be talking about here? This was a brand-new editor pumping these out sometimes in literally half an hour. Moreover, if you look at their history, there's strong evidence that they were doing absolutely no "refining" – e.g. randomly replacing en-dashes with the character 'â' at the very top of the edit summary, duplicating an article's prose twice in a row by copy-pasting their new version in (e.g. putting an entire new version of the article within the infobox), directly copy-pasting chunks of prose in, and otherwise giving the LLM free reign to do things they plainly couldn't understand like the change from "{{Reflist}} to {{Reflist|}}". These reviews were clearly just fed into an LLM and then copy-pasted by a brand-new editor with a snowball's chance in Hell of actually meaningfully changing what was generated – let alone knowing what they were doing in order to make "refinement" a meaningful metric. There is literally no way these were meaningfully reviewed by a human who could've possibly known they were doing, so I wouldn't even call this "undoing" the review; there was no review in the first place. Frankly, I think this alone calls into question the idea that just anybody should be allowed to perform GA reviews, given it's obvious nobody is auditing them. Edit: Moreover, I think your definition of trying to make these "talk page comments" is abjectly incorrect or at least completely violates the spirit of the guidelines, and I hope the guidelines are rewritten to properly clarify this. Nominations? Sure. The review itself? Absolutely not. That's supposed to be a substantive critique by a human. TheTechnician27 (Talk page) 16:44, 21 August 2026 (UTC)

(edit conflict) The refinement required by WP:LLMTALK to discard their reviews, of course. The timing seems to be the most significant factor in your evidence; normally that sort of cross-referencing is considered circumstantial in real life. I’m looking for clear evidence that the editor’s comments were divorced from reality to such an extent that it can only be produced by an LLM with no human looking at the text: either something obviously wrong that the editor claimed was correct, or vice versa. I don’t think your repetition of brand-new is particularly helpful; new editors may have edited previously as IPs, or just get up to speed quickly.

I had a quick look at Talk:Carl Lawson (sprinter)/GA1 and am struggling to match your claim there was no review in the first place: the comments seem precise and relevant. Are you saying that reverting all the changes Metalicat suggested in their review would improve Wikipedia? ~~ AirshipJungleman29 (talk) 16:55, 21 August 2026 (UTC)

I'm honestly a bit insulted that you can look at what they did from 11 February to 13 February and try to even contest that a human plausibly did that. Yes, I'm going to harp on about "brand-new editor", because IP addresses aren't running around performing GA reviews, and even an experienced editor isn't doing the kind of nonsense they did. The fact they were brand-new just makes it more plainly obvious.

"normally that sort of cross-referencing is considered circumstantial in real life" Yes, and? Multiple pieces of circumstantial evidence come together to remove doubt. Any lawyer, detective, etc. will tell you that most evidence is circumstantial, that most cases are decided on circumstantial evidence, and that your treating of "circumstantial" like a dirty word is nonsensical.

Your being the only one known to be with a victim at their house the night of the murder is circumstantial. Your fingerprints being on the knife is circumstantial. The victim's blood being on your clothing is, believe it or not, circumstantial – you could spin a tale that the victim had a nosebleed. Your Internet search the night prior for "how to dispose of a body" is circumstantial. The fact that you have a lengthy history of abusive text exchanges with the victim is circumstantial to the murder. Meanwhile, direct evidence is often less reliable than circumstantial evidence. A witness saying "I saw you stab the victim" is direct evidence, but on its own considerably less reliable than the circumstantial evidence I laid out above.

"Are you saying that reverting all the changes Metalicat suggested in their review would improve Wikipedia?" I'm saying that unless you think LLMs should be given carte blanche to perform GA reviews now, this was not ever properly reviewed against GA standards and needs to be done by a human. It sucks that nobody actually did the work when people thought they did, but it's better than the alternative, where we let people flood the zone with LLM output and then decide "Well, too bad, I guess". TheTechnician27 (Talk page) 17:19, 21 August 2026 (UTC)

But if testimony is your thing, I copy-pasted in about 10,000 characters from this edit to Pat O'Keeffe, GPTZero 4.9b gives it approximately an 87% chance of being generated ("We are highly confident this text was AI generated"). The chance for "mixed" is 13%, and the chance for "human" is 0%. I think that's less reliable than the circumstantial evidence, but I'm always willing to accommodate disagreement. TheTechnician27 (Talk page) 17:36, 21 August 2026 (UTC)

I don’t see why you would be insulted. We’re here talking about the potential LLM-generation of their GA reviews, not their contributions to Pat O'Keeffe, which as I have said seem likely to be based on LLM-generation.

If you want to gain consensus for a complete reversal of all the editor’s contributions, the place for that is WP:ANI. But currently, I don’t see proof that the GA reviews under discussion are LLM-generation without refinement, and, to be blunt, your theatrics aren’t really helping convince me. But I’m probably mistaken (and supposed to be on holiday!) so I’ll WP:DISENGAGE and let others take this forward. Best of luck, ~~ AirshipJungleman29 (talk) 18:20, 21 August 2026 (UTC)

User:AirshipJungleman29: Enjoy your holiday! Bgsu98 (Talk) 19:44, 21 August 2026 (UTC)

I think the answer, broadly, that you've got is that (1) the appropriate venue is GAR and (2) there is no consensus on whether the reviews themselves are satisfactory. What's important now is that the articles, in their current state, meet the GA critera, which as I have demonstrated is certainly not the case for one article. Please ping me if you start one for Israel Keyes so I can copy my comments over. (in solidarity), JacobTheRox^{(talk|contributions)} 19:19, 21 August 2026 (UTC)

Any GA articles you believe do not meet the criteria for GA should be sent to GAR. If you identify them, I'm happy (well, "happy" may not be the right word, but...) to initiate the GARs for you. Bgsu98 (Talk) 19:44, 21 August 2026 (UTC)

With some knowledge of taxonomy (albeit mostly arthropods), I'd be willing to throw myself at a reassessment of one of the plant articles (or maybe both). Noted above, my problem with the idea that we must raise specific objections to GA status when they weren't assessed to begin with (and the same reason why LLM prose isn't allowed on Wikipedia) is the bullshit asymmetry principle, where someone could flood the zone with garbage GA reviews which then take 10x as much effort to debunk. I appreciate you raising these for reassessment, Bgsu98, so I'll try to help out where I'm able. If I can at this point (I don't know the logistics of reassessment), I'd like to start a review of Urtica ferox. I really hope there are guidelines that further clarify this soon. TheTechnician27 (Talk page) 22:41, 21 August 2026 (UTC)

I think if you do the GA review right there on the reassessment page, that should be sufficient, and it would be logged in the article’s talk page history. I will go ahead and do the review of the Rhode Island election article. Bgsu98 (Talk) 22:47, 21 August 2026 (UTC)

Hi @AirshipJungleman29, I nominated Carl Lawson (sprinter) earlier this year. See my GAR-review (GARR?) here - although I didn't think it at the time, with a fresh pair of eyes I'm convinced that AI was used for at least part of the review. I don't think I've done a GAR so I'm not super familiar with the standards expected, but given there were shortcomings and to erase all doubt I would appreciate a reassessment. Thanks, --Habst (talk) 22:30, 21 August 2026 (UTC)

Usage of Russian "reliable sources" on English-language articles

Hi all, just wanted to ask a quick question: I have a couple sources for something I'm working on that have a green checkmark on Cite Unseen (for "generally reliable"), but they have "RU" on top of that checkmark. Clicking it links me to the Russian Wikipedia's reliable sources list, where it does confirm it's reliable; I just don't know if standards for source quality vary by language. The article is in English, and the other sources are all in English, but those sources do provide unique Russian-specific context I can't find anywhere else. Will including these sources affect a GAN? WALKKKKK (talk) 23:14, 22 August 2026 (UTC)

If there is no consensus for the reliaibility of a source on the English Wikipedia, the GAN reviewer will have to decide. I don't know how similar the Russian Wikipedia's source standards are but I doubt many other editors do either. Could you specify what the websites in questions are? (in solidarity), JacobTheRox^{(talk|contributions)} 23:24, 22 August 2026 (UTC)

The sources in question are ru:DTF and VGTimes, both "authoritative sources" for video games. The translated text is mostly criteria for what's valid. WALKKKKK (talk) 23:33, 22 August 2026 (UTC)

Looked more closely, and accidentally answered my own question: the DTF article isn't in their "editorial" section, and the VGtimes article isn't post-2020, which are the exceptions that would make them "authoritative" in the first place. Never mind then. WALKKKKK (talk) 23:35, 22 August 2026 (UTC)