Wikipedia talk:Counter-Vandalism Unit/Archive 7

Archive 1

←

Archive 5

Archive 6

Archive 7

Archive 8

Obviously self created and not-notable AfC drafts found in the process of counter-vandalism

What is the most useful thing I can do if a newly-created AfC submission draft comes up in the recent changes feed I'm watching, and it is obviously a self-created article (for example Draft:BaguetteCayden).

Should I just wait for these to go through the AfC process, or is there something I should be doing to flag these when I see them to take some workload off the AfC reviewers? Exor674 (talk) 02:05, 27 April 2019 (UTC)

Research and documentation - automated moderation

Hello, I do Wikipedia research at the University of Virginia. I am still sorting notes but I have this study to share -

• meta:University of Virginia/Automatic Detection of Online Abuse

In summary, the team used machine learning to examine Wikipedia users who have blocks, and used the analysis to try to predict which user accounts do not have blocks but are seemingly engaged in misconduct.

This research and practice raise all sorts of cultural and technological challenges. To advance the conversation on this I also started drafting documentation about the general concept at WP:Automated moderation. I am not seeking any particular response at this time but if anyone wants to develop our policy on human / bot interaction in moderation, then feel free. The social issues are more complicated than the technical ones. Thanks. Blue Rasberry (talk) 21:29, 26 July 2019 (UTC)

Number of recent changes watchers

Originally posted at WT:Recent changes patrol

Hi, I'm looking for a rough estimate of the number of editors watching recent changes at any given time. Even an order-of- magnitude estimate would be useful. looking ideas? T.Shafee(Evo&Evo)^talk 23:41, 29 August 2019 (UTC)

New to Wikipedia

Hi there,

While I’m new to Wikipedia, I have a particular subject matter on the the encyclopaedia which I want to start improving. I have spent a few attempts, the most recent being this evening, where I put one hours work into upgrading the article.

A contributor who has a profile for 13 years vandalised and undid all my edits even though I had not quite finished.

I would be considered a person of knowledge on the subject matter of this article in the real world.

I thought Wikipedia was interested in gaining knowledge and updates to articles. Is there something warning with the system, whereby a long term member of Wikipedia can vandalise a new contributors page because maybe there is a personal bias on the subject mater for example ? — Preceding unsigned comment added by Celtic Musician (talk • contribs) 22:41, 22 September 2019 (UTC)

Celtic Musician, welcome to Wikipedia! I assume this is about Mutt Lunker reverting your edits to Great Irish warpipes. Mutt has replied to you on their talk page twice, and they're entirely correct - while we appreciate that you are a subject matter expert, statements added to Wikipedia are expected to be backed by reliable sources. Is there a source other than oral history for your edits - perhaps a book? If you can cite that, then there is no problem with adding content. If you can't find a source, you can comment on the article's talk page (something like "I've heard (story) from my family but I can't find a reliable source for it, does anyone know of a source that could support it?")Also, please do not mark your edits as "minor" unless they are trivial changes like spelling or adding punctuation - when in doubt, don't mark it minor. creffett (talk) 23:06, 22 September 2019 (UTC)

Also, I forgot to mention - vandalism has a specific meaning on Wikipedia, and Mutt's actions were not vandalism. creffett (talk) 23:07, 22 September 2019 (UTC)

Bias

Having made a complaint of vandalism. Where the edits I was attempting to make were being undone by someone who is a long term member of Wikipedia, I take onboard totally the points made in references being needed. My point is, in this particular case , the policing of the edits was carried out by the actual original editor of the article who had many inaccurate statements with zero backup from real sources. I was not even been given time to finish my edits so frantic was this writer to undo my corrections. If Matt genuinely believed in the accuracy of Wikipedia instead of some idealogical belief he was protecting Wikipedia or his own article, he would have sworn this energy engaging and working with me to correct the errors on his original piece. Anyway , I’m done. Wikipedia can keep the inaccurate state article and continue to erode the original reason it was suppose to exist. I will just pursue a more trusted destination for the quality information and accuracy I can bring to that topic. Ultimately the resistance to allow improvements to Wikipedia articles without this kind of harassment only spells one trajectory for Wikipedia. Sad for the carry on. I donated a lot of funds to Wikipedia over the years also. When I believed it was a force for actually using the power of mass knowledge to improve as many encyclopaedic articles as possible.

Clearly someone in Wikipedia thinks Wikipedia is right enough now to stop improving or no longer make it easy for people to improve.

I have no time for this. I need to get back to the real world and life of now. Celtic Musician (talk) 01:06, 23 September 2019 (UTC)

Celtic Musician, given that Matt Lunker had edited the article a grand total of once before you edited it (to remove another unsourced claim over a year ago), I find your claims of bias and ownership somewhat hard to believe. If you're accusing me of bias and ownership, uh, I've never touched the article and don't really care one way or the other about bagpipe-like instruments (other than not liking hearing them at 6 AM when camping). If you want to post a dramatic statement and leave Wikipedia, go ahead, but you're making mountains out of molehills here. All Matt did was revert unsourced edits (as many people do to many articles), that happens all the time and has nothing to do with wanting to preserve a "factually inaccurate" article. Both Matt and I have given you very reasonable suggestions on how to proceed, so there's really not much else anyone can do here. creffett (talk) 02:00, 23 September 2019 (UTC)

User:Bonthefox3

Are this user really a member of this unit, despite the account was created on 2019-02-08 ? It seem his edits were fishy. Matthew hk (talk) 14:57, 8 October 2019 (UTC)

@Matthew hk: Probably not, but then there's not really any specific requirements to join CVU or display the userbox that I'm aware of. It looks like they copied some of the userboxen off of Ifnord's talk page. I'm more concerned about the NPR and account age boxes than the CVU box, since those actually are inaccurate. I dropped them a message, hopefully they take care of it. If they ignore it or go inactive again, then maybe some trimming is in order. ‑‑ElHef (Meep?) 16:43, 8 October 2019 (UTC)

Read the history of edits - removed changes 4 times since start of August. The end guys. No talking to ideology.

Reviews Ganging together to protect each other’s policing. Who polices the police? As the age old conundrum question of corruption goes. Best of luck guys. Celtic Musician (talk) 06:57, 23 September 2019 (UTC)

Too bad they don't read before they rollback.Skalle-Per Hedenhös (talk) 12:16, 11 October 2019 (UTC)

Patrolling vital articles

Hi, I was wondering if the members of CVU might be able to answer a question I have about Huggle. How do I filter the queue to just vital articles? Category:All Wikipedia vital articles is applied to talk pages, so if I add it to a Huggle filter, it seems to only include pages in the Talk: namespace and not in mainspace. Is there an "include corresponding namespace" switch somewhere I'm missing, or a script, or another tool that I could use? The only thing I can think of is to add the 38k pages to my watchlist, which doesn't seem like the best idea. Also, anybody know why we apparently don't have a mainspace hidden category for vital articles? Thanks for any suggestions! – Levivich 23:00, 23 November 2019 (UTC)

​This is an open task phab:T117122 which hasn't gotten any attention, sadly. Other projects create huge lists of articles: Wikipedia:WikiProject Medicine/Lists of pages/Articles in order to use Special:RecentChangesLinked (which it seems Huggle supports, but I don't have it myself). I can assemble this humongous list for you, do you want the associated talk pages as well? – Thjarkur (talk) 11:50, 2 December 2019 (UTC)

Thank you, Þjarkur, for the pointers and the offer to assemble the list. I can actually make the list myself but I very much appreciate your offering to take the time to do it. I'll bump the phab ticket in case any devs might have interest in adding the feature to Huggle–I think it'll be very useful for WikiProjects, not just the vital articles that I want to watch. I can't figure out a place to base a Huggle queue off of what-links-to-this-page. I'm not seeing an option for that in the settings? So I guess that leaves me with two choices: (1) create a page of all vital articles and then use Special:RecentChangesLinked to watch those articles, without using Huggle at all (which is fine, really, and probably what I'll do), or (2) create an alt account just for vandalism patrolling, add the 38k Vital articles to the watchlist of that account, and then use Huggle with the watchlist as the queue. The downside to both is that the list would need to be periodically updated, so I think the phab task is still the ticket for a long-term solution. In the meantime, #2 strikes me as more work than is justified just to be able to use Huggle to patrol, when Special:RecentChangesLinked will do the job. I wasn't aware of Special:RecentChangesLinked, so once more, thank you very much for helping me solve this problem! – Levivich 17:24, 2 December 2019 (UTC)

​Here you go: Special:RecentChangesLinked/Wikipedia:Vital articles/List of all articles. – Thjarkur (talk) 20:10, 2 December 2019 (UTC)

@Þjarkur: Awesome–thank you!!! – Levivich 20:41, 2 December 2019 (UTC)

Vandalism vs. Wandalism

As Wikipedia begins with a W, and 'wandalism' is more humourous, I think that vandalism had ought to be referred to as wandalism when it is in the context of Wikipedia. -- Beaneater00 (talk) 14:42, 10 December 2019 (UTC)

You're welcome to think whatever you'd like, but I've reverted your changes to the WP:CVU page. ‑‑ElHef (Meep?) 15:51, 10 December 2019 (UTC)

bots can accidentally hide vandalism and make it much harder to fix

If i understood my experience just now with Shrovetide correctly, we should require bot owners to check for vandalism before running them to prevent bots from making vandalism much harder to detect and fix. And why was it not possible to click on undo on this edit? There didn't seem to be any later edits that affected the parts changed by that vandalism. --Espoo (talk) 23:14, 21 December 2019 (UTC)

@Espoo: bots can't really analyze every edit that was ever made before they come along - though this is easier for human editors. Those recent bots did make parts of the article better than the last verions and they didn't make any parts of it worse. Compare that to your own last edit (Special:Diff/931884165), you made part of the article better with restoring that sentence, but you also made parts of it worse by weakening the accessibility of the references. You also removed content another editor added, and put in the wrong type of interwiki association. — xaosflux ^Talk 00:44, 22 December 2019 (UTC)

Espoo, the whole point of bots is that (most) operate autonomously without the operator having to do anything. If a bot operator were to manually review all of the bot's edits, then they might as well be making the edits themselves. creffett (talk) 00:45, 22 December 2019 (UTC)

But bots could be programmed to notice and at least warn their operators about the kind of easily noticeable and typical vandalism observed here: removing text and even individual letters and even producing nonexistent words and even without providing an edit summary!

And why wasn't the undo function available? (And my manual revert provided a valuable improvement and removed only trivial improvements that the bots can and will automatically redo.) And the bots made the article worse because they made it more difficult for editors to notice the vandalism and much more difficult to fix it. --Espoo (talk) 01:04, 22 December 2019 (UTC)

Espoo, no idea why undo wasn't available, but if we could make a bot that reliably flagged "easily noticeable and typical vandalism" with a very low false positive rate, we would (well, we have a few counter-vandalism bots, but none are perfect). If you think you can write one that does that, then please do so! As it is, the bots are doing exactly what they're programmed to do (since they're not designed to account for vandalism, and doing so would add a lot of complexity), and sometimes that might conflict with what a human thinks is best. In cases like this, your best bet would be to use something like Twinkle's "restore this version" functionality to restore the last good version. creffett (talk) 01:09, 22 December 2019 (UTC)

But for that i'd have to have first noticed the vandalism. Since it was "hidden"/obscured by the bots' trivial improvements, which resulted in a clearly less good and text with an incomprehensible sentence, there is the very real danger that bots will result in less vandalism being noticed and fixed by most editors!

Any idea why vandalism patrol bots didn't notice this vandalism?

I'll try to contact some bot owners to ask if it's difficult to make bots warn their users about such easily noticeable vandalism. --Espoo (talk) 01:24, 22 December 2019 (UTC)

@Espoo: bots only do a specific set of tasks, and no bot should ever be relied upon to make any future edit. No single bot will ever do "everything" either, so a bot that does things like say add additional reference links is very unlikely to also be an antivandalism bot. — xaosflux ^Talk 02:03, 22 December 2019 (UTC)

Also how much better or much worse a single edit makes an article is quite subjective, in your own edit an automatic review may show that you added some text (good) added an improper link (bad) removed some refences (bad) and removed a template (bad?). How should some good reviewing bot score that edit? — xaosflux ^Talk 02:07, 22 December 2019 (UTC)

Espoo, speaking as a professional software engineer: things that seem obvious and visible to you can be very difficult to program into a computer (see this comic for an example of something that's easy for a human but really hard for a computer). The answer you're almost certainly going to get from bot operators is "no": it's outside of what they designed the bot to do and it would be very complex to add a vandalism check before making a change. Also, you seem to have a misconception about how bots work. While some bots can be directed by a user to perform a task on a specific page, many of them are operating effectively unattended, and they just watch recent changes or use some other method to find pages to patrol (that is, there isn't a "user" to warn). creffett (talk) 02:12, 22 December 2019 (UTC)

I'll answer both of you at the same time. In order to not use up too much of our time and perhaps result in some change that actually improves WP, i'll try to summarize the situation and my comments, whose main points both of you misunderstood:

• Bots perform many helpful and tedious tasks that would otherwise not get done or would take editors' time away from more important things.
• Most vandalism is apparently spotted and fixed by editors without any kind of vandalism detection software.
• Most editors don't check an article's history for vandalism before editing, and most editors only check the last few edits for the cause of any vandalism they happen to notice in the article.
• When bots edit after vandalism has occurred, that vandalism is no longer noticed by most editors unless they happen to read an affected part of the article, and then they usually only notice and fix part of the vandalism and usually with less skill and knowledge than the "dedicated" writers of the text, who didn't just read it to inform themselves. So the more edits are carried out by bots (whose edits are overwhelmingly less important than normal edits), the more degradation of the content and readability of articles due to vandalism will remain unrepaired if bots don't even try to notice and warn their operators about even an obvious suspicious previous edit.
• I never said bots should check all previous edits or should be relied on to make future edits or that the operator should manually review all of the bot's edits or that the bots shouldn't operate unattended. On the contrary. I suggested that bots should warn their operators about the last one or two edits if they have certain features typical of vandalism and that the bots should then demand an OK before carrying out the edit so that the operator can quickly check if it's obvious vandalism before clicking OK. I also never said any bot should or would ever be able to do "everything", but if a bot that does less important things is unable to warn its operator that its edit will degrade the content and comprehensibility of WP by obscuring probable vandalism, then the bot is helping to degrade the content of WP in exchange for trivial improvements.
• Both dedicated anti-vandalism bots as well as other bots are apparently so far unable to notice and flag something as obvious as an edit without a summary that removes parts of words and sentences, and even when the result is a nonexistent word. This means that it is of prime importance that the undo function be improved so that it works even after new edits when these new edits haven't affected the vandalised sentences. --Espoo (talk) 11:37, 22 December 2019 (UTC)

Discussion on warning templates

 You are invited to join the discussion at Wikipedia talk:Template index/User talk namespace#How to handle editors chronically misusing the warning templates?. {{u|Sdkb}} ^talk 04:00, 28 April 2020 (UTC)

Idea for interface to make it easier for casual readers to revert vandalism

As good as we've gotten at finding and speedily reverting vandalism, there's still inevitably some period where it exists uncorrected in mainspace for readers to come across. I propose that we do something like the following for edits flagged as likely or very likely problematic by ClueBot:

1. All text added in the likely problematic edit is underlined for readers in, say, orange.
2. If a reader hovers over the added text, a small box (similar to a page preview) pops up, saying "our algorithms detected that this recently added text may have problems or be vandalism".
3. At the bottom of the mini pop-up box are two buttons; clicking one reverts the edit, and clicking the other removes the underline (but doesn't remove the flag for recent changes patrollers).
4. The underline doesn't appear to the user who made the edit (to fight abuse), and if no one interacts with it, it disappears after the page receives 10 new edits or 48 hours passes.

I see a few advantages to this:

• It helps readers more easily identify vandalism and, by making our vandalism-fighting more visible, reassures readers we are tackling the problem and discourages potential vandals.
• It helps vandalism get reverted more quickly.
• By engaging casual readers in a simple editing activity, it helps break down the mental barrier many readers have where they think of Wikipedia as something to be read and not edited. This could help encourage them to later start making more serious edits.

What do you all think of this? Is it desirable, and if so, what would need to happen on the technical side to implement it? Sdkb (talk) 06:18, 30 January 2020 (UTC)

Systems similar to this have been proposed, are you familiar with previous proposals at WP:Flagged revisions and WP:Deferred changes? Philosophically, you're taking things in kind of the opposite direction from most of those proposals - most of them seek to reduce the visibility of possible vandalism, rather than emphasize it. Personally I'm not sure applying an orange highlighter to problematic edits is such a great idea. It may discourage some potential vandals, but could easily encourage others. ‑‑ElHef (Meep?) 15:20, 30 January 2020 (UTC)

I have to agree with Elhef on this one. Galendalia (talk) 10:30, 30 April 2020 (UTC)

How do i join?

Hi... I just started yesterday, and the community help desk supplied me with a link to the help center in which I spotted a section for counter vandalism. I read a few articles regarding the topic and I found it to be pretty interesting. I'd like to get more involved with it. The only question is how do I join and how do I specifically use my user against internet vandalism. Juiceboxraider (talk) 01:20, 2 May 2020 (UTC)

Hi @Juiceboxraider: - Sorry about the long delay. The first thing you should do is understand what CVU exactly is. You should participate in the Counter Vandalism Unit Academy (CVUA). You can do this by following the instructions located at Wikipedia:Counter-Vandalism_Unit/Academy. Once you successfully complete this academy, you will understand what the CVU does and be given the tools to help you. I hope this helps. If you have any questions about this, please feel free to ping me. --Galendalia _{CVU Member} \ ^{Chat Me Up} 19:06, 19 May 2020 (UTC)

Is there a way to feed vandalism that failed automatic detection into a machine learning algorithm?

Lately there seems to be a decline in human-reverted vandalism and an increasing reliance on robots to revert obvious cases. I've sometimes caught vandalism that was left on pages unreverted for days even weeks whereas this would be unheard of years ago. I wonder what steps are being taken to allow the current robots on duty to learn from machine learning (speaking as someone who uses machine learning for scientific and medical research), and what are the current processes in place to "feed" undetected positives (false negatives) and false positives so robots can better learn from these mistakes? Yanping Nora Soong (talk) 20:13, 2 May 2020 (UTC)

@Yanping Nora Soong: User talk:ClueBot Commons would be a good place to follow up on this, for a bot that runs client-side; also you can look at mw:ORES for a API-based scoring system. — xaosflux ^Talk 22:48, 2 May 2020 (UTC)

Moved to User talk:ClueBot Commons

I like this thought! {{u|Sdkb}} ^talk 23:22, 19 May 2020 (UTC)

Unit member's Rzvas contribution

Unit,

19 February 2020 I corrected Wikipedia page of cosmonaut Aleksandr Samokutyayev. In the sentence "Samokutyaev and cosmonaut Maksim Surayev and cosmonaut performed a spacewalk outside the space station on 22 October 2014." I deleted words "and cosmonaut" which mean that EVA was conducted by three cosmonauts - Samokutyayev, Surayev, and some third - unnamed, unknown, or secret. This is plain error, evident for anybody with minimal knowledge of English language, just microscopical knowledge of the field, and that could be checked in 1 mouse click and 2-3 seconds of time.

User Rzvas, marked on user page as Counter-Vandalism Unit member, reverted this correction. The same date I asked user Rzvas at user's talk page to explain this reversion. Since then user did not answered or edited, so Wikipedia page Aleksandr Samokutyayev over 5 months bears ridiculous and disreputable error.

Not feeling that I have any obligations to correct such kind of user contribution to Wikipedia, I leave correction of this error to Counter-Vandalism Unit. Also I highly recommend the Unit to check all the other contributions of user Rzvas which could be of the same quality. — Preceding unsigned comment added by 188.35.130.139 (talk) 13:17, 24 July 2020 (UTC)

I have made the correction to the Aleksandr Samokutyayev article. I cannot speak for any other CVU members, but I have no plans to go back and review all of Rzvas's contributions. Please remember that Wikipedia is a volunteer service, and that everyone is free to respond (or not) in their own time. ‑‑ElHef (Meep?) 15:21, 24 July 2020 (UTC)

Request for Comment

Hi Counter-Vandalism Unit/Archive 7, there is an active RfC on m:WikiLoop/DoubleCheck/RfC:Levels_for_WikiLoop_DoubleCheck_Reviewers#Overview that we think you might be interested. Please join the discussion there. xinbenlv ^{Talk, Remember to "ping" me} 02:31, 15 August 2020 (UTC)

Can I help?

Hi, can I help fight vandalism? — Preceding unsigned comment added by Sir Lancelot of the Lake (talk • contribs)

@Sir Lancelot of the Lake: You don't need to ask for anyone's permission to fight vandalism. If you want to trained by a mentor, you can ask any of the mentors at Wikipedia:Counter-Vandalism_Unit/Academy/Instructors on their talk page that are willing to help you fight vandalism although Puddleglum seems to inactive so I don't recommend asking that one unless she becomes active again. Interstellarity (talk) 20:59, 21 September 2020 (UTC)

Sounds good! You've been very helpful! Sir Lancelot of the Lake (talk) 21:02, 21 September 2020 (UTC)

Village pump entry on undetected vandalism

Started a Village Pump proposal here about a more formal task force to target undetected vandalism on Wikipedia, which is my main focus of editing; it seems like the kind of thing people here might have thoughts on. Gnomingstuff (talk) 04:06, 5 November 2020 (UTC)

Specifically, I think there is a lot of focus on recent' changes -- which is good -- but almost none on edits that pass the initial sweep. So as a result, very obvious and often embarrassing vandalism can go undetected for years. If anything they can get harder to detect, as well-meaning people go in and wikify and clean up the vandals' edits, making them less obvious. I'm doing extremely low-tech "think of something people vandalize pages with, search the term, filter out likely false positive terms, go through each page individually" but there is probably a better way, even if it's just having more than one person do it. Gnomingstuff (talk) 04:11, 5 November 2020 (UTC)

how do I inform you of vandalism?

How do I actually tell you if I find anything? Bababeoy (talk) 22:49, 15 December 2020 (UTC)

@Bababeoy: The general idea of CVU is learning how to address vandalism yourself, so in most cases the goal would be for you to simply take care of it. However, if it's something you're not sure how to handle you can always raise it on this page and someone can take a look, or if it's particularly problematic you can raise it at WP:ANI. ‑‑ElHef (Meep?) 14:00, 16 December 2020 (UTC)

How can I become a trainer

How can I become one? TigerScientist (talk) 22:37, 14 December 2020 (UTC)

@TigerScientist: If you feel you're experienced and qualified to train new vandal-fighters, I'm not seeing any specific criteria that you would need to meet. That said, please be realistic on your level of expertise - it looks like when you were attempting to add yourself to the CVUA instructors list you were offering to train users on Rollback (a user right you do not have), Twinkle (which you have been using for less than two weeks) and RedWarn (which it appears you haven't used at all). My recommendation would be to get a few months of countervandalism experience under your belt, find out which tools you're most comfortable with using, work towards getting the Rollback right, and then give it a try. ‑‑ElHef (Meep?) 14:15, 16 December 2020 (UTC)

Ok thank you I will try to learn that. TigerScientist (talk) 16:11, 16 December 2020 (UTC)

Join

How do I join this WikiProject? Iconman1 (talk) 13:42, 21 December 2020 (UTC)

Counter-vandalism article inquiry

Hello,

I'm a journalist working on an article about counter-vandalism for Wikipedia Day. I'm hoping to speak to someone in a senior/administrative role in the CVU. Is there anyone here who might be willing to answer some questions about the project and counter-vandalism in general?

Thanks,

Ben — Preceding unsigned comment added by BenLindbergh (talk • contribs) 14:15, 11 January 2021 (UTC)

It's been a while

Anyone have a brief summary on advances in anti-vandalism in the past 10-13 years? I know the bots have taken a lot of the work humans do, but it seems like there's still a role for us. Specifically, do any of the tools have the ability to monitor a specific category, or pages in the remit of a particular WikiProject? Thanks, Abeg92contribs 20:42, 29 January 2021 (UTC)

Dormant little land mines, sitting there for years

I'd like to start a discussion to maybe exchange tools, tips, and tricks for finding and/or dealing with vandalism that sneaks through the various protections and ends up remaining in an article for a long time. I'm sure I'm not the only one to find vandalism in an article that has remained there for years. What, if anything, can we do to be more successful in ferreting these out?

My most recent example was from Historiography of the causes of World War I, where some college kid added his name, or his friend's name, into a list of famous historians of World War I. This happened in 2009, and was only just removed today. (Actually, the sequence was, that the user added it into the parent article, Causes of World War I in the #Historiography section, which was then split off in 2010 into the child Historiography article, where it remained until today.) It was first inserted in rev. 273335003 of 02:13, February 26, 2009 by 130.108.199.86, and then copied over in good faith by the creator of the new article on 21:08 May 23, 2010 in rev. 363797737.

What could we have done better to find this sooner, instead of after eleven years? What about going forward—there are surely other little unexploded land mines lurking out there, not attracting attention on anybody's watchlist because they've been out there for so long, and the regular editors on the article are so used to seeing them they don't see them; they might even have been inserted not far before a citation, so they might appear sourced. How do we unearth these? Thanks, Mathglot (talk) 06:46, 12 January 2021 (UTC)

I've been working on this for some time, and the short answer is, yes, there are hundreds if not thousands of cases of undetected vandalism. I know this is kind of my soapbox but these are particularly worrying because due to the nature of Wikipedia, if it sticks around enough, it gets picked up by sites that scrape Wikipedia, possibly Google, and even legitimate news/research sources. The longest I've found is over 13 years old.

The way I find some of these is very brute-force, and only works for "obvious" vandalism, but I have a list of keywords that I search for, usually combined by filtering out false-positive keywords (for instance "poop" but not "poop deck"). Some numbers are on the watchlist as well -- obviously the number 69 showing up in an article isn't always vandalism, but there's a better than average chance, same with 420, etc. Any slang or Internet memes, not just current but late '00s/early '10s as well. Phrases like "this wiki sucks" or whatnot. That kind of thing. You sort of have to think like a 14-year-old boy. Gnomingstuff (talk) 04:19, 13 January 2021 (UTC)

@Gnomingstuff: Is this a literal list that you can send to me? If so, I would like to have a copy. Sam at Megaputer (talk) 22:23, 29 January 2021 (UTC)

I don't have anything written down, but basically it's, like I said, stuff that a 14-year-old boy or an extremely online troll would find funny to edit Wikipedia with: swear words, misspellings of swear words, anything that has to do with poop, pee, farting, or other bodily functions, teenage slang, Internet memes, joke names ("Heywood Jablome," "Mike Hawk," etc.), joke numbers (any combination of 69, 420, 666, etc.), etc. Not quite along those lines, but still something I search for, are phrases found in common test edits ("___ was here," "hi," etc.), blatant statements of opinion ("this wiki is ____," "____ sucks," etc.), etc., although with these you have to be pretty heavy on the filters with these though since they catch a lot of templates. Gnomingstuff (talk) 01:39, 30 January 2021 (UTC)

Idea: Delaying newbie edits

I request your input/suggestions on Wikipedia:Delayed flagged revisions that proposes to add a time delay to IP/newbie edits; while also allowing reviewers/vandalism patrollers to approve those edits to go live immediately. It can be done by adding a time based auto-review feature to FlaggedRevs/PendingChanges. Delaying IP/newbie edits for few hours (~2 hours) would give a time window for patrolling those edits. All edits that are not undone by anyone would go live after the delay. Thus there will not be any backlog- Vis M (talk) 20:11, 16 February 2021 (UTC)

Want to be included

I wish to be included in the counter-vandalism projec Mrgamji (talk) 04:31, 26 April 2021 (UTC)

Add this userbox template to your page:

{{User wikipedia/CVU-Vandal Fighter}}

which renders as

This user fights vandalism.

It also automatically adds you as a member of the Counter-Vandalism Unit. --littleb2009 (talk page) 21:25, 17 May 2021 (UTC)

How to restore pink box?

I know I'm in the wrong place. An IP had no talk page but had a pink box saying the IP was blocked before I posted a message. Now there's no pink box or any indication the IP address is blocked unless one goes to the contributions page, which has a long list of reverted edits. Is there a way to show the IP address is blocked? And my message might be pointless anyway since it was about an old Help Desk question I just now saw.— Vchimpanzee • talk • contributions • 22:49, 7 June 2021 (UTC)

@Vchimpanzee: That pink box block notice appears on IP user pages and user talk pages when the page doesn't exist (a la the box showing deletion log entries, which shows on redlinks and not on bluelinks). It appears on the IP's contributions regardless. I'm not sure why it is the way it is, but as far as I know the only way to restore it on that particular page is to delete the page. ‑‑ElHef (Meep?) 23:55, 7 June 2021 (UTC)

Is there a way, or is there a need, to identify this IP address as a vandal? It turns out there are numerous IP addresses but many of them apparently committed vandalism.— Vchimpanzee • talk • contributions • 15:44, 8 June 2021 (UTC)

No need; the IP address hasn't done anything wrong. As you pointed out, the only contribution from that address has been an off-topic helpdesk question. Not really something we can help with, but certainly not vandalism. Even if they were vandalizing, it's a rangeblock on a broadband ISP which probably has dynamic IP addresses – the person using that address in April could be on a completely different address now. In any case, if someone needs to know about that block, there's still plenty of places to find out about it (block log, the box on the contributions screen, rangeblock finder, etc). ‑‑ElHef (Meep?) 17:53, 8 June 2021 (UTC)

Okay. My assumption was the person who asked the help desk question might have had a dynamic IP.— Vchimpanzee • talk • contributions • 19:33, 9 June 2021 (UTC)

Template:whois code needs fixing

Hello, hoping to draw some attention to the code withing Template:whois. I started a talk section on the template talk page, but haven't gotten a response yet (very inactive page as is). The template used to be used frequently, but it appears that the code has been broken for sometime due to URL changes and such. Its primary purpose was to alert IP users that their edits were being tracked to their IP addresses, in effort to prevent further vandalism by the user.
My attempts to remedy the code were unsuccessful; hoping that someone may be able to assist with this. Thanks! --PerpetuityGrat (talk) 22:18, 7 July 2021 (UTC)

@PerpetuityGrat: I added {{edit template-protected}} on your talk page section to get a template editor to take a look. If that doesn't work, WP:VPT would probably be a better place to ask. ‑‑ElHef (Meep?) 12:52, 8 July 2021 (UTC)

@ElHef: great thanks! --PerpetuityGrat (talk) 00:14, 9 July 2021 (UTC)

CVU Academy

Hello, I'm not sure who looks over this talk page but I just wanted to post a message to any editors instructing new editors about combatting vandalism. I often see newish editors whose vast majority of edits are reverts of IP editors' contributions. I don't see them reverting any user accounts even though, as an admin, I see plenty of user accounts committing vandalism and plenty of created user accounts get blocked for vandalism or disruptive editing. I'm wondering if the focus on IP editors is because they are unlikely to push back and complain about being reverted.

Having started my editing on Wikipedia as an IP editor, I don't think this focus solely on reverting IPs is fair and doesn't do much to combat vandalism by sockpuppets and other created accounts. Please take a balanced approach, don't discriminate and don't target IP editors as low-hanging fruit. Thank you for anyone who takes a moment to read this comment. Liz ^{Read! Talk!} 23:18, 8 July 2021 (UTC)

Liz, Yo! I agree with this, it is a bit unfair to be solely focus on IPs. As someone who has gone through the course it’s not all about IPs but it does give you a bias against IPs. I still find myself unconsciously checking IPs edits more than users with accounts. Thanks for the post, although I’m not an instructor, this will be really helpful for instructors to read. Best, Signed,The4lines |||| ^{(You Asked?) (What I've Done.)} 00:02, 9 July 2021 (UTC)

I have reverted probably thousands of instances of vandalism at this point, mostly longstanding vandalism. This means I start by finding the vandalism first, then checking to see who did it, i.e., I don't "target" anyone. And the fact of the matter is that most of it, at least the obvious crude variety, is done by IPs, or at least throwaway accounts with no edits besides vandalism. The ratio is something like 10 to 1, if not more. The reason is fairly intuitive: if a 15-year-old kid wants to change a bunch of numbers in article to 69 they're not going to go to the effort to make an account for that. I'm sorry you have the impression that IP editors are being "discriminated against," but this is how it is. Gnomingstuff (talk) 17:48, 17 August 2021 (UTC)

WP:DENY and user warnings

Other editors have already pointed out that the Counter-Vandalism Unit unintentionally encourages vandalism. Depictions of vandals as an enemy mob glorify vandalism. It does not help that we are role-playing as policemen and the military. I propose that we modify our vandalism warnings.

Compare:

This is your only warning; if you vandalize Wikipedia again, you may be blocked from editing without further notice.

STOP VANDALIZING WIKIPEDIA!!! MR ADMIN WILL BAN YOU!!!

What is really the effective difference? Our warnings amuse and encourage vandals. Seasoned LTAs and vandals read out the second example in their minds when we give them the real warning. I propose that for the uw-vandalism series of templates, we use minimal wording. We remove the stop sign template and boldface too:

If you vandalize Wikipedia again, you will be blocked from editing.

What do you think? wikinights _talk 22:24, 14 September 2021 (UTC)

Do users counter vandalism with anything but standard warning templates? -PerpetuityGrat (talk) 02:12, 15 September 2021 (UTC)

There is no time for patrollers to use anything else. Some anti-vandal tools may use their own custom warning templates. wikinights _talk 03:23, 15 September 2021 (UTC)

If you're dealing with a vandal with this mindset, getting any warning at all is giving them attention. I don't see that weakening the warning templates is going to help. If you honestly think you're in a WP:DENY situation, simply revert the vandalism and report them without warnings, explaining your reasoning. If it's severe and/or obvious enough for you to make that call, you should be able to make the case to an admin. If it's vandal who's not in this mindset, then I think a strong level 3 or 4 warning is a good thing, as it can serve its purpose (to discourage further bad edits). As a sidenote—if there's serious thought of trying to get this changed, probably a good idea to loop in the people at WP:UW. ‑‑ElHef (Meep?) 18:31, 15 September 2021 (UTC)

Practically, it is usually difficult to know when to DENY. There are two conditions:

• Has the vandal chosen to behave in a certain pattern that would make their identity obvious? Usually not. Most vandalism is simple and not unique, like content blanking or insertion of vulgar language. In most cases where DENY is ideal, it is impossible to know that.
• Is the reverting user familiar with the repeat offender's behavior? Usually not. Unless you RCP (recent changes patrol or patroller) often or like to read LTA pages, you are unaware that this account's behavior resembles a certain repeat offender, and thus you are unaware that they are a repeat offender at all. Remember that there are tens, if not hundreds of habitual vandals on enwiki at any given time. There are too many vandals for most RCPs to know most of them.

I think, then, that it is better to DENY all vandals by changing our warnings rather than rely on the inadequate knowledge of RCPs. wikinights _talk 04:16, 17 September 2021 (UTC)

I guess I'm a bit confused on your interpretation of DENY. To me, DENYing a vandal is revert/block/ignore with minimal-to-no warnings, which is something that's not reasonable for all vandals because there are still a great many opportunistic vandals who will stop once they realize someone is watching and that there are consequences to their actions.

As to the language in the warnings, which ones are you proposing to change? The lower-level warnings already have much softer language (level 1 in particular) and they're designed to escalate the further up it goes. You called out the 4im warning above, which should only be pulled out in extreme cases. When it's needed, it really should be a "Stop NOW or else" type of message. An attention-seeking vandal that's doing something that disruptive knows full well what's coming regardless of what message they get. ‑‑ElHef (Meep?) 15:55, 17 September 2021 (UTC)

• Because vandals usually end up getting indeed and leaving for more rewarding grounds I doubt that our templates precise wording matter int he grand scheme of things. Someone who is going to sock farm VOAs for a month straight or end up a LTA is doing it for other reasons apart from our provoking them with scary looking templates. ^Thanks,L3X1 ◊distænt write◊ 03:55, 18 September 2021 (UTC)

Dealing with a large number of vandalism edits

I recently stumbled upon 123.208.211.136, who has a prolific history of vandalizing and then usually but not always self-reverting, going back to August 2021. The contributions list has more than 500 entries on it! I've reported them at Wikipedia:Administrator intervention against vandalism and presumably they'll be blocked, but there are now a very large number of suspect edits, not all of which have been reverted. This includes changes to various numbers in articles by small amounts, which can be tedious to revert since it requires care to not disrupt subsequent constructive changes. And, given there are so many, it would be a big task to do entirely myself! Yet, if not somehow coordinated, I could see a lot of duplicated effort and missed reversions. Any advice on how to proceed? Smcpeak74 (talk) 15:04, 2 January 2022 (UTC)

@Smcpeak74: Late advice, but mass rollback helps a lot – AssumeGoodWraith (talk | contribs) 12:28, 26 January 2022 (UTC)

Vladimir Putin edits needed

Can Vladimir Putins page be edited appropriately to highlight that he is authoritarian, commuting an illegal war and invasion? Perhaps paint him in the way we wish Russians would see him- lying to his people, killing women and children and lying to the young men fighting for Russia who are also dying.

Seems like a move that could be helpful to improve the flow of correct information to Russia. 81.96.7.66 (talk) 10:12, 6 March 2022 (UTC)

This is not the proper place to make such a request. Since the article's talk page has been protected, your request should be placed at Wikipedia:Requests for page protection/Edit. That said, please note that any such request must be supported by reliable, published sources, and that edit requests that may be controversial must be supported by a consensus of Wikipedia's editors prior to being implemented. If you would like to see this information added to the article, your best course of action is to create a Wikipedia account and get it autoconfirmed so you can participate on the article talk page. ‑‑ElHef (Meep?) 16:11, 6 March 2022 (UTC)

{{CVU editnotice}}

Is Template:CVU editnotice (edit | talk | history | links | watch | logs) one of your templates? It is a recently created template by a new user with few edits and a redlink userpage. The content of the template does not seem to make sense in the context of the CVU. -- 65.92.246.142 (talk) 13:41, 11 March 2022 (UTC)

I suspect that User:Rockiemillion is going to propose to add it somewhere, you should ask them first. — xaosflux ^Talk 14:26, 11 March 2022 (UTC)

Teyora - Development first look!

Hi! I'm Ed6767, the original creator of RedWarn, now one of the most popular tools on the English Wikipedia that's been used by over 1,000 Wikimedians to make over 300,000 edits since mid-2020 that's been praised for its user friendliness and ease of use, but criticised for its limited functionality. I'm leaving this message as I think it may be of interest here - I left the RedWarn project in November to develop Teyora, my successor to RedWarn (alongside Chlod's UltraViolet). It's a new in development web app that uses some of the latest web technologies to create a highly extendable all in one editing tool with a focus on administration, counter vandalism and general patrolling - not to mention, it'll work on every Wikimedia project without any prior configuration and can be used by any user with at least auto-confirmed rights*. Now, I'm ready to give the Wikimedia community a first look at what I've been doing over the past six months and what to expect going forward.

You can check out the 20 minute first look at the in development version on YouTube here: https://www.youtube.com/watch?v=nzlpnzXdLP4.

There's lots more to expect too! Why not read the full details page at meta:Teyora and leave any feedback, comments or wishes at meta:Talk:Teyora (please leave any correspondence there to keep discussion centralised). If you're interested, you can leave your signature

*with basic features, advanced features require configuration. To prevent abuse, auto-confirmed users will be in a restricted mode until approved by an admin or via rollback rights.

All the best, ✨ Ed talk! ✨ 22:56, 15 April 2022 (UTC)

Huggle

Hi, after a few years hiatus I would like to go back to vandal patrol. I used to use Huggle extensively and I see that it's still around (possibly enhanced), but could someone advise if Huggle is currently the preferred tool for automated vandal fight - or if there are other tools I should also consider? Thanks. Húsönd 08:53, 14 April 2022 (UTC)

@Husond, Huggle is still one of the most-used tools. You could also try SWViewer, or just use Twinkle/Ultraviolet (or RedWarn) on Special:RecentChanges. ― Qwerfjkltalk 06:23, 12 July 2022 (UTC)

Twitter link

Category talk:Wikipedia notability guidelines is seeing spam due to a Twitter link. HLHJ (talk) 14:34, 2 August 2022 (UTC)

Discussion at Wikipedia:Village pump (proposals) § Requesting talk page access revocation via AIV instead of ANI

 You are invited to join the discussion at Wikipedia:Village pump (proposals) § Requesting talk page access revocation via AIV instead of ANI. 🐶 EpicPupper ^{(he/him | talk)} 21:50, 18 August 2022 (UTC)