Wikipedia:Over-hasty Speedy Deleters

Many people have had negative encounters with overly eager speedy deleters. The problem is that a careless, hasty, or overzealous CSD can be much worse than a common vandal. Vandalism is extremely easy to spot, can be removed in a few clicks, in very short order, and everything is completely visible. An overeager speedy deleting admin's damage, however, is completely invisible to the common user. The questions thus become:


 * Why do people have concerns about speedy deleters?
 * What do speedy deleters need to know or do to prevent this feeling?

The necessity of speedy deleters
Speedy deleters are a necessary part of the project. Without speedy deleters, the project would be overrun with vandalism and become impractical. If all the content editors suddenly quit working on Wikipedia, the site would still be useful for years to come. If all the new page patrollers/speedy deleters suddenly quit working on Wikipedia, Wikipedia would be worthless in a matter of days. Speedy deletion is thus an absolutely necessary part of the project—when done right!

Why the criteria are so strict
The policies and guidelines surrounding criteria for speedy deletion are pretty narrow. Since the criteria for speedy deletion allow a user to delete an article unilaterally, they have been narrowly constituted to ensure that only those articles with clear consensus to be deleted qualify. They are explicitly written so that in the case of doubt, the article should be prodded or added to Articles for Deletion. Thus, when considering the criteria for speedy deletion, one should always err on the side of caution. Unfortunately, the general sense is that most people who engage in them do not. This sentiment is probably the result of a few bad apples, but it taints the entire criteria for speedy deletion crowd.

Why do editors have to jump through the hoops?
Despite the narrow guidelines for Speedy Deletion, editors will often ignore these rules. The rationale that people will often give for ignoring the rules is they feel that the article would be deleted anyway. Thus we might as well delete the article today instead of wasting people's time at Articles for deletion or waiting for a 7-day proposed deletion (PROD) to expire. The problem with this reasoning is that it ignores one of the primary reasons we are here—to build the encyclopedia. When we speedily delete an article, we are telling somebody that their input is not needed or wanted. This will often turn a contributor off of Wikipedia. If, however, we give them the chance to redeem their article, they might learn something about our process and stick around. To quote Balloonman: "'One of the first articles I ever wrote was nominated for Speedy Deletion. I convinced the nominator not to delete the article, but to send it to Article for deletion. Today that article is a featured article.'"

While an article may be eligible for deletion, the community should work with different editors who are willing to prove the value of an article. Nobody would consider speedy deleting Twitter today, but that's exactly what was proposed back in 2007. It was, however, retained and was at one point a good article.

The problem with criteria for speedy deletion
The problem with criteria for speedy deletion is that it denies the author the chance to validate the article. Sometimes articles that we think are obvious deletes are not obvious in the eyes of the person who wrote the article. By giving the author a chance to prove the value of the article, we give the author an education on Wikipedia. We also give the author the sense that they had a chance to save the article, as compared to being the victim of a lone gun vigilante. If one or even two people tell you that your article is garbage and should be deleted, it comes across as bullying. If, however, you are taken to Articles for Deletion and a number of people collaborate on the verdict, it becomes more tolerable. It becomes a "fair" system.

The person who leaves the project because of a careless speedy deleter might have been the next SandyGeorgia, Tony1, or Moni3—all great featured article contributors. Perhaps the deleted article itself might become a featured article!

But shouldn't the author know?
Shouldn't the author know to read Your First Article before beginning? Your First Article clearly tells authors about Wikipedia's standards and guidelines, it tells them about notability, references, verifiability, etc. It also warns them that articles failing to meet WP's standards may be deleted. Even if the author did not know about Your First Article, there are other safeguards protecting articles from premature deletion. They can challenge the speedy deletion and an admin has to actually perform the deletion.

These are nice sounding platitudes, but don't really work in the real world. How many people, upon arriving at Wikipedia, knew to read Your First Article? Most newbies couldn't even find it if they had to—and even when it pops up when they create a page, how many people actually read it? Most people who write an article have experience as an IP before starting one from scratch. It's like the terms and conditions where people had to click, "I Agree" before starting an account—nobody does that? As for the safeguards of and an admin having to delete the article, these are only echoed by people unfamiliar with article building. Anybody who has written new articles has undoubtedly had their own run in with an over eager CSDer—some admins will delete articles without giving the author any warning. They simply delete.

Even editors with featured article/good article/did you knows to their credit have had articles speedily deleted. It cannot be argued that they are unfamiliar with all the policies and guidelines surrounding articles. Many people, while writing articles, will often save the article at various stages of development. On numerous occasions, they will save an article in draft form, only to go back to writing the article. But sometimes, before they have a chance to make another edit, their article will be tagged for deletion and then deleted. Even for a veteran editor, it can be very frustrating when somebody acts too quickly and doesn't give the creator a chance to add a hangon tag. Granted this is partially the fault of the system, as a criterion for deletion is by its nature a delete-on-sight beast, but it can be very frustrating to see articles tagged and deleted within 2 minutes of creation!

Plus, let's be frank, too many speedy deleters do not adhere to the policies and guidelines surrounding criteria for speedy deletion—and this includes admins!

Speedy deleters who do not understand the reality of the situation demonstrate the reason many people at Requests for adminship expect (or demand) significant article contribution. Without actually working on an article, it is impossible to really understand how the article writer feels.

The damage is unknown
The other problem with CSD is that the damage is not known. When a vandal wrecks an article, it will almost always be identified and fixed. The vandal will be blocked and potential collateral damage will be checked by a dozen users.

When an admin acts in a careless or hasty manner, deletes valid articles and upsets users, nobody will ever know. Occasionally a user will call the admin out, but this will usually be done via the user talk space... in other words, nobody will really know that a mistake was made. Even if, on the outside chance that the admin is taken to AN/ANI, the odds are that the case will be deemed an anomaly and written off. Heck, everybody makes mistakes right? The problem is, that there is no oversight of CSD. CSDers are a unique breed, you either like to work it or you don't. If you don't, you usually don't pay it any heed. There is no discussion and the actions of different admins are not compared with one another to ensure consistency. Generally, one doesn't know what articles any given admin has deleted or what CSD's an admin has rejected.

The stereotype
This leads to the stereotype of the typical CSDer/vandal fighter. The stereotype of the CSDer is the high school student who is flexing his (stereotypically male) muscles and asserting his authority, by deleting the work of others. Some have compared it to a role-playing game, where the CSDer gets experience points for every article he kills. This stereotype might be wrong, but until/unless people have better encounters with CSDers, the stereotype will be perpetuated. Again, it is only a small percentage of CSDers who spread this image, but it only takes a few to ruin it for all of the hard work of others.

The difference between "may" and "must"
One of the elemental mistakes when tagging for speedy deletion is that most people fail to see the difference between being allowed to delete the article and having to delete it. Speedy deletion allows admins to delete articles within the criteria – but at their own discretion. WP:CSD clearly states:

Just because a page does meet one or more of the criteria, it does not mean it has to be deleted. The reviewing administrator is still allowed to decide whether to delete the page in question and can always decline it. If the admin thinks the article should rather be PRODed or sent to AFD, then he/she may do so. Some people think that when an admin declines a valid speedy deletion request, the best course of action is to slap the tag back on, hoping some other admin will ignore the previous decline and delete it. Others think they have to complain to the admin in question about it, although his decision was well within policy. There are just two tags when an admin can only decide one thing, i.e. to delete: valid G10 and G12 taggings. In all other cases they may decline the request and not speedy delete the page.

Admins should check to see if the article has undergone admin shopping. If another admin has declined a speedy deletion, subsequent admins must as well (even if they disagree with the decision) unless it is a valid tag of one of the two mentioned above or if it is tagged with a different criterion that does fit.

At requests for adminship
When reviewing Speedy Deleters, people will look at their criteria for speedy deletion nominations. There are a few criteria that are closely examined:

G1

 * Patent nonsense. Pages consisting entirely of incoherent text or gibberish with no meaningful content or history. This excludes poor writing, partisan screeds, obscene remarks, vandalism, fictional material, coherent non-English material, poorly translated material, implausible theories, or hoaxes.

G1 is perhaps the most abused tag by careless CSDers. When somebody mistags articles G1, it indicates that they are lazy or too willing to delete. Per WP:Nonsense there are only two types of articles that fit G1:


 * Total nonsense, i.e., text or random characters that have no assignable meaning at all. This includes sequences such as "sdfgdsfkgdshgdkhgdsklhsklgroflmaolololol;;;'dsfgdfg", in which keys of the keyboard have been pressed with no regard for what is typed.
 * Content that, while apparently meaningful after a fashion, is so completely and irredeemably confused that no reasonable person can be expected to make any sense of it whatsoever.

The following examples are not G1:


 * Mwuggle is a term of endearment that combines a kiss (mwah) and a cuddle (cuggle) between two people in love. It is usually found at the end of a text message or e-mail to let a loved one know how much they are cared for. i.e. "Good night pettle. Mwuggle". The use of the word has grown steadily in recent years from its origins in Mullaghduff, Co.Donegal to Dublin and indeed into Edinburgh and other areas of Scotland.
 * The collection of water in the parking lot south of Palmer Hall on the University of Toledo campus in Toledo, OH. 
 * A day the is rumored to be the day after cyber Monday. In fact it is just dirty lies spread by business teachers which give there students something to do. 

That doesn't mean that they should not be deleted, just not per G1. Nor are the following: 1 2 3. A 2008 review of 25 articles deleted for G1 violations was done; the results can be found here. In only 6 (24%) of the cases deleted per G1, was G1 properly applied. 12 (48%) of the other articles could have been speedily deleted under other criteria, and 7 (28%) should have been prodded or even kept!

G1 explicitly excludes poor writing, partisan screeds, obscene remarks, vandalism, fictional material, material not in English, poorly translated material, implausible theories, or hoaxes! Finding mistakes in G1 is the low hanging fruit. Most CSDers mess up here, and if they mess up here, one doesn't have to look further!

NOTE: If a criterion other than G1/G3 can be used, it is usually preferred to do so. G1/G3 can be seen as "bitey" as they imply maliciousness or incompetence in the author. If you can use another category, do so.

G2

 * Test pages. This excludes the sandbox and the users' own user space.

Most cases that should be deleted G2 are the ones where the person made it fairly obvious that it was a test page (EG they make a statement to the effect that they were just practicing.) Other cases that might apply would be if they have a bunch of different coding items that are not really related to one another... EG a picture of George Washington, a graph showing the annual median income in Venezuela, a table of Super Bowl winners, etc. If it has content and is intended to be an article in full or part, it isn't G2. If it has content, but isn't intended as part of an article, G2 might be appropriate. Personally, I've seen a lot of less-than-encyclopedic quality but totally workable content, that looked a lot more like a first draft than a test page incorrectly deleted under G2.

It's usually a good idea for G2 candidates to be userfied, and the editor who created them to be notified of the page's new location. Since editors may not be aware that they have a userspace, this will not only clean up article space, but allow the user to continue their experiments with minimal disruption.

G2 is almost a subcategory of G3, but assumes good faith.

G3

 * Pure vandalism. This includes blatant and obvious misinformation, redirects created by cleanup from page-move vandalism, and blatant hoaxes.

G3 occasionally leads to the deletion of valid articles. An article is vandalized. A CSDer comes along, sees the vandalism. Tags the article for Speedy Deletion. An Admin comes along and sees the vandalized article and the tag, and deletes it. Unfortunately, the article started out as a valid article that was vandalized. The vandal thus succeeded in getting an admin to do his/her dirty work.

Moral of the story, always check the article's history. This should be a general guideline for anybody doing CSD work! Many articles have been deleted that shouldn't have. Also, if the editor is an established editor, give them space. The article might not be in a current state worthy of Wikipedia, but if they are established, the odds are that they aren't finished! Plus, by speedily deleting the work of an established editor, you only perpetuate the stereotypes.

NOTE: If a criterion other than G1/G3 can be used, it is usually preferred to do so. G1/G3 can be seen as bitey as they imply maliciousness or incompetence in the author. If you can use another category, do so.

G4

 * Recreation of deleted material. A sufficiently identical and unimproved copy, having any title, of a page deleted via a deletion discussion. This excludes articles that are not substantially identical to the deleted version, articles that address the reasons for which the material was deleted, and Content moved to user space for explicit improvement. Material moved or copied to circumvent Wikipedia's deletion policy is not excluded). This also excludes content undeleted via deletion review, deleted via proposed deletion, or to speedy deletions (although in that case, the previous speedy criterion, or other speedy criteria, may apply).

G4 explicitly is not for articles previously speedily deleted or PRODDED, but articles deleted through the discussion. And only then if the article is essentially the same as the one sent through AFD. If the article is a substantial rewrite or has addressed the reasons for deletion, it is not a G4 candidate.

G10

 * Pages that disparage or threaten their subject or some other entity, and serve no other purpose. These "attack pages" may include slander, legal threats, and biographical material about a living person that is entirely negative in tone and unsourced. These pages should be speedily deleted when there is no neutral version in the page history to revert to. Both the page title and page content may be taken into account in assessing an attack. Articles about living people deleted under this criterion should not be restored or recreated by any editor until the biographical article standards are met.

G10 gets overused as a result of BLP. People will see an article that says something unsourced about a living person, the something is negative, and instantly delete the article per a BLP violation. While articles that are entirely negative about individuals and unsourced should be deleted, CSDers should first check to see if they can add a source! But even if they don't do that, simply because an article includes a BLP violation does not mean that this gives CSDers free license to delete any article with negative information. First, the CSDer is expected to check to see if there is a version of the article without a BLP violation. Second, the CSDer should delete the contentious material first. For example, two deleted articles about a Prime Minister and President of African Countries: These two individuals were alleged to have been involved in atrocities committed in their countries under their rule. While these allegations would require sourcing, the rest of the articles discussed the two individuals as heads of state. Removing a sentence or two would have removed the BLP concern leaving decent start class articles. Instead, they were deleted. This is an abuse of G10.

Oh yeah, one other thing, BLP has a key requirement, the person discussed has to be living. An unsourced criticism about Hitler is not a BLP violation.

Redirects are often wrongly deleted under this criteria. Many famous people have negative nicknames, such as Butcher of the Balkans, Slick Willie or Mad dog of the Middle East. These redirects are not attack pages if these nicknames are well used or cited in an article. Redirects should only be deleted as attack pages for negative nicknames that are rarely used or completely made-up.

NOTE: If an article fits G10, it is better to delete it G10 than A7. Some admins will work attack pages and copy vio pages before they work any other category. If an article is a BLP issue, blank the page while you wait for an admin to delete it. The admin can read the history to verify that it was an attack page. No need to let wikiclones copy an attack page.

G11

 * Unambiguous advertising or promotion. Pages that are exclusively promotional, and would need to be fundamentally rewritten to become encyclopedic. Note that simply having a company or product as its subject does not qualify an article for this criterion.

G11 is by far and away one of the poorest written CSD criteria. What does it mean? What is an unambiguous advertising or promotion? If it comes from the company's website, yes that is obvious, but that should be deleted as a copyright violation. If it uses first person pronouns, that is another obvious sign that the page is advertising. But is the page that makes the claim, "The ABC Company is the number one global provider of X in the North West" pure advertising? According to some the answer is yes. G11 is a very poorly written criterion. This criterion has been debated several times at WT:CSD because critics of it say it fails both criteria 1 and 2 for CSD rationales. Criterion one says that the criterion has to be objective; as written, this is not the case. It is purely subjective. Criterion two says that the criterion has to be uncontestable. Again, this is not the case because nobody really knows what it means. G11 is an area where one should tread cautiously.

G12

 * Unambiguous copyright infringement. This applies to text pages that contain copyrighted material with no credible assertion of public domain, fair use, or a compatible free license, where there is no non-infringing content on the page worth saving. Only if the history is unsalvageably corrupted should it be deleted in its entirety; earlier versions without infringement should be retained. For equivocal cases that do not meet speedy deletion criteria (such as where there is a dubious assertion of permission, where free-content edits overlie the infringement, or where there is only partial infringement or close paraphrasing), the article or the appropriate section should be blanked with {{subst:Copyvio}}, and the page should be listed at Copyright problems. Please consult Copyright violations for other instructions. Public-domain and other free content, such as a Wikipedia mirror, do not fall under this criterion, nor is mere lack of attribution of such works a reason for speedy deletion. For images and media, see the equivalent criterion in the "Files" section here, which has more specific instructions.

While this criterion is rarely misused, please confim, prior to deletion, that the other web site isn't a mirror of Wikipedia.

A1

 * No context. Articles lacking sufficient context to identify the subject of the article. Example: "He is a funny man with a red car. He makes people laugh." This applies only to very short articles. Context is different from content, treated in A7, below.

A1 explicitly declares that the article has to be very short. This generally means 2 sentences or less. When an article is a paragraph or two it no longer qualifies. If an article includes references or external links, and those links are valid, it does not qualify. If the article has context, even as a hoax or vandalism, it does not qualify. If you can garner enough information to look for sources, regardless of whether or not sources exist, it does not qualify. In other words, "John Smith was the President of the United Auto Workers" is not an A1 candidate—it gives plenty of context, not enough to save it from AFD, but it does give it context. In a 2008 review of A1 deletions, only 35% of them actually met the definition of A1!

A7

 * No indication of importance (individuals, animals, organisations, web content, events). An article about a real person, individual animal(s), an organization (e.g. band, club, company, etc., except schools), web content or event that does not indicate why its subject is important or significant. This is distinct from verifiability and reliability of sources, and is a lower standard than notability. This criterion applies only to articles about web content and to articles about people, organizations, and individual animals themselves, not to articles about their books, albums, software, or other creative works. The criterion does not apply to any article that makes any credible claim of significance or importance even if the claim is not supported by a reliable source. The criterion does apply if the claim of significance or importance given is not credible. If the claim's credibility is unclear, you can improve the article yourself, propose deletion, or list the article at articles for deletion.

A7 only refers to real people, individual animals, organizations, web content, or events. Too often this is applied to other articles. An article doesn't have to prove that the subject is important or significant nor does it have to provide reliable sources; if it makes a credible claim that it might be important or significant, then it is not speedily deletable. It's important to distinguish "importance or significance" from "notability": A7 does not require that an article indicates that the subject meets a notability guideline, merely suggest that the article could be improved to a state where it does. "Credible" is added because some kid writing an autobiography of himself declaring himself to be the best lover the world has ever seen may be a claim to importance or significance, but it clearly isn't credible. The same kid, however, might be able to make a credible claim saying that he received an award from the president for saving another kid from drowning in a pool. That award may not be enough to save the article, as there may be no evidence of receiving the award from reliable sources, but it is enough to prevent it from being speedily deleted.

Other non-criteria
You should also be familiar with non-criteria for CSD.

Adminshopping
If an admin declines a speedy request, this does not mean that you should simply retag it and hope a different admin comes along. If an admin, or anybody unaffiliated with the article for that matter, removes a CSD tag, send it to AFD or PROD; don't simply replace the tag hoping somebody else will delete it. This is Adminshopping and should be discouraged. Admins should also not delete speedy articles that were declined, especially if it was declined by another admin. This only encourages sloppy CSD work and Adminshopping. If you disagree with another person or admin, discuss the rationale with that person. The only exceptions to this rule are clear COPYVIO and BLP Violations. However, if you find an article may meet another criterion, different from the one that was declined, it can be retagged with that new rationale. This is largely because some admins only handle certain criteria and leave others alone; the admin who declined the first request may simply not feel like he is able to judge the other criterion.

Getting it right matters
People at RfA are often criticized for using the wrong tag, even though an article clearly deserves to be deleted. Others have asked, why does it matter? The reason is simple. Since CSD has such a potential for abuse/error, it is imperative that people who engage in CSD get it right. In order to support a CSDer, one has to be sure that they know the difference between a G10 and a G1 and that they are going to adhere to the policy and applicable guidelines. It doesn't matter if the article could be deleted under a different criterion; admins need to get it right. As an admin candidate who wants to work in an unregulated area that has such potential for error, you really need to prove you can be trusted with the power of deletion. You have to show that you are near perfect!

Options
When in doubt, err on the side of caution. Send it to AfD/PROD. Even if it is obviously something needing to be deleted, if it doesn't cleanly fit one of the accepted criteria, send it to AfD or tag it with a Prod. At AfD, the verdict may be obvious, and it may be speedily closed, but at least you gave the article's creator a chance to defend the article and obtained a broader perspective than just two editors.

The difference between an article that falls under the speedy deletion criteria and an article that falls under other deletion criteria may seem like a technical distinction, which in the interests of expediency, perhaps we should sometimes dispense with. Indeed, if an article looks non-notable, it may be tempting to go ahead and stretch the speedy deletion criteria a bit to try to avoid the trouble of an AfD. However:
 * 1) The prod and draft processes gives editors time to look for sources to shore up the article's notability. Unless you've googled around, you probably don't know whether such sources exist, so why not tell the article creators about your concerns and give them an opportunity to address them, without an imminent speedy deletion looming? Once editors are in the position of trying to get a speedy deletion overturned, or creating a new version of the article from scratch, collaboration (and potential group discussion) has already been interrupted by the removal of the article from view by non-admins.
 * 2) The AfD process lets editors argue and counter-argue their cases, producing a record that users can refer back to when they're considering recreating the article. They can see whether they have better arguments for notability than the arguments that didn't prevail in the AfD. In marginal cases, the likelihood of article recreation is probably higher, so it's all the more reason for creating that record.
 * 3) Having multiple editors weigh in at AfD creates a greater likelihood that a variety of deletion arguments and counter-arguments will be made that will stimulate further research and writing, resulting in the best article that can be produced on that topic, being produced in that 7-day period. Basically you're helping steelman the article creators' arguments that an article on the topic should exist.

One of the challenges that CSD/NPP'ers have at RfA is that they are seen as "only destroying the work of others" or they "don't understand what it's like to have their work scrutinized/deleted." Many CSD/NPP'ers will find it difficult if not impossible to pass an RfA without substantive article building. The problem is that people who enjoy CSD/NPP work typically are not the types of people who likes writing articles.

But if you really want to impress people and help out the encyclopedia, salvage the article. Take an article that you were going to nominate for CSD, and help it meet Wikistandards. This is the best way to show that you understand the plight of the newbie editor. You don't have to take the article to FA quality, but taking an article that you were thinking about deleting and turning it into a start class article will differentiate you from 99% of your fellow CSDers. Plus, if you do it with a new article, you might be able to get some Did You Know recognition.