Wikipedia talk:Edit warring/Archives/2014/May

Does Change = Revert?
recently blocked me for violating 3RR, but has refused to provide the four revert diffs. Through discussion here and here, it seems that Bbb23's interpretation of 3RR is that any change to the existing text of an article is a revert and can be counted as such by an admin while justifying a block for violating 3RR:


 * "You changed the text in both paragraphs of the lead. Those changes constitute a revert. Some editor or editors wrote that language. You undid their "actions" by changing it."

And


 * "Remember, a revert is changing information on a page, and that include edits that you might not think of as disruptive".

Bbb23 explained his interpretation of WP:3RR:


 * "If an earlier editor or editors wrote something in the article, a change to that material constitutes a revert per the definition in the policy. That doesn't mean, however, that administrators don't have discretion not to count what technically constitutes a revert. Obviously, context matters, and, perhaps less obviously but not suprisingly, administrators may disagree on how to exercise their discretion." --Bbb23 (talk) 00:40, 10 March 2014 (UTC)

Is this accurate, because it comes as a serious shock to me; after 40,000 edits and 4 1/2 years on Wikipedia I'd never heard this before. Is copyediting virtually synonymous with reverting? If I used 5 edits to trim inaccurate and unsourced material from an article is that really a 3RR violation? GabeMc (talk&#124;contribs)  17:25, 10 March 2014 (UTC)

Comment It seems your issue is with application of 3RR by an admin, not with 3RR itself. The steps at WP:ADMINABUSE are more appropriate for pursuing your concerns.—Bagumba (talk) 17:52, 10 March 2014 (UTC)
 * It seems to me that any change that cannot be connected to a specific edit and a specific editor is not a revert. I.e., if you can't say that I reverted XXX editor, then you shouldn't call it a revert. Bbb23 said that this non-contentious copyedit is a revert, but then who did I revert? and, is the edit really a revert? Per your above comment: "Any admin worth a damn knows how to tell the difference between edit warring and fixing up an article", are any of these four edits reverts? the edit at 17:31, the one at 18:41, the one at 18:49 or the one at 19:09. Because these are the four that Bbb23 used to justify my 3RR violation.  GabeMc  (talk&#124;contribs)  17:10, 10 March 2014 (UTC)
 * I don't think so. This is more about clarifying the guideline then it is about Bbb23. If he is correct, then this should be made more clear. If he is wrong, then this should be made more clear, but either way this is an issue with the guideline as its currently being interpreted. might have something to add.  GabeMc  (talk&#124;contribs)  17:58, 10 March 2014 (UTC)


 * Similar problems with Bbb23 resulted in this AN thread and this WT:EW thread. vzaak 18:33, 10 March 2014 (UTC)


 * Your question: If I used 5 edits to trim inaccurate and unsourced material from an article is that really a 3RR violation? is unclear. WP:3RR says that A series of consecutive saved revert edits by one user with no intervening edits by another user counts as one revert.  You can use eleventy-seven edits to change everything in the article and it's only one revert if there are no intervening edits.  If there are enough unrelated intervening edits to get your editing up to 3 reverts, probably at least some of the editors involved should take a break already.  If the intervening edits are related it's probably edit-warring.  You didn't put the specifics of your case in front of us here, but really, it's hard to hit 4 reverts (properly understood) without some kind of bad stuff being done on the page by someone.&mdash; alf laylah wa laylah (talk) 22:42, 10 March 2014 (UTC)
 * I made four groups of consecutive edits: . Does this clarify the situation ? GabeMc  (talk&#124;contribs)  22:51, 10 March 2014 (UTC)

I did not analyze all of those links, but the policy has a big problem......a revert needs to be defined better, specifically to include the condition of as undoing the recent work of another editor, and in a contentious situation. And even iif "recent" needs to be left vague, it's a lot less vaguer than what we have now. Without that, it is subject to the whims of people misapplying or even misusing the policy. And I've seen it abused, especially to try to deprecate people with bogus claims of violating 1RR. Claiming that an initial edit was a "revert" of something that an editor did 2 years before.  North8000  (talk) 01:55, 11 March 2014 (UTC)


 * Well 2 years wouldn't inhibit some admins. As Kww says, he doesn't care how far back a previous version goes, and doesn't even care whether the revert was intentional: "I don't care whether the editor intentionally undid a previous edit or how long ago the previous version was, if I can identify a previous version of the article that used the wording that the edit created, that edit is a revert". The tradition on Wikipedia is to allow admins individually to make up their own ideas about what is blockable. Admins use terms like "discretion" to give a cloak of respectability to this policy. It seems a deliberate strategy to demoralise content builders (who are not admins) by creating fraught and uncertain environments for them to try and work in. The answer is not to try and create guidelines that look at every possible wrinkle, but to provide centralised control over the discipling of content builders, and to provide a bill of rights for content builders. That means that many current admins would lose their privilege to block content builders. And that's a consequence that the incumbent admins as the controlling power block will not countenance. --Epipelagic (talk) 03:50, 11 March 2014 (UTC)
 * The principal (but unprincipled) admin strategy is to refuse acknowledgement so it is unnecessary to engage with rational debate on system failings. In this way there is a pretence at dismissing system critics as trolls or vandals. This is a natural response for admins, since (a) the system critiques carry considerable weight and cannot be countered with rational argument (b) admins already have (self-assigned) draconic powers on Wikipedia, and as is usual might prevails. Content builders (who are not admins) have two choices: accept the miserable and unjust deal or leave. --Epipelagic (talk) 06:31, 11 March 2014 (UTC)


 * GabeMC highlights why 3RR is the worst policy on Wikipedia, bar none. There is nothing more useless at Wikipedia than the 3RR policy, because the thing we should be preventing is unhelpful bickering over content that happens in the article space.  I see no evidence of that kind of bickering here.  Admins should be investigating whether or not users are using the undo/rollback function (or the equivalent) as a means to win conflicts with other editors, not counting some arbitrary number of something.  If an editor is not engaged in an open conflict in the article space, what is the point of a block?  As an administrator myself, I have never invoked 3RR as a reason to block someone, because of the shortcomings of the very concept that is behind 3RR.  -- Jayron  32  11:04, 11 March 2014 (UTC)

Admins need to follow policy. We can fix the problem by fixing the policy. I'll try it BRD by adding "recent" to the definition.  North8000  (talk) 12:08, 11 March 2014 (UTC)
 * Did it. We should probably also add "of a contentious nature"   but it will take more work to get that added right.  North8000  (talk) 12:12, 11 March 2014 (UTC)


 * The particular tweak you just made to policy might counter Kww's creative endeavour to make life more difficult for content builders (though perhaps he will claim "recent" means within ten years). But how is it going the counter Bbb23's even more splendid offering, "a revert is changing information on a page"? However you turn and baste the policy, enterprising admins like these will use the fat to cook their own version. Nothing short of a disciplinary panel moderated by a bill of rights is likely to introduce some order and fairness into the system. --Epipelagic (talk) 16:31, 11 March 2014 (UTC)


 * No, we shouldn't add "of a contentious nature"; edit warring is edit warring, whether something is contentious or not. And as this discussion (from the WP:BLP talk page last month) shows, Wikipedia editors can even disagree on what is contentious. The word recent that you added should be fine since the 3RR policy's time frame is within a 24-hour period and "[a]ny appearance of gaming the system by reverting a fourth time just outside the 24-hour slot is likely to be treated as an edit-warring violation." So for that same reason, I don't see why your "recent" edit was needed; if certain people could not correctly comprehend the text regarding the 3RR policy's time frame before your addition, that is the fault of those certain editors, not the 3RR policy. Flyer22 (talk) 13:39, 11 March 2014 (UTC)
 * I believe that you are conflating two different time-related things. The 24 hours is the time frame for the 3RR counting.    My "recent" was a qualifier on the criteria of what makes an edit a "revert" for the purposes of prohibited actions and disciplinary actions.  Without that qualifier, even an edit that "partially undoes" a 5 year old edit is considered a "revert".   And don't forget that this definition is commonly applied to 1RR and 0RR. Without this fix, you can (and do) have ludicrous situations.   E.G. that any editing violates 0RR, and any revert is illegal if your did another gnome edit on the article that day, and that doing a lot of gnome work on any article in a day violates 3RR.   North8000  (talk) 16:24, 11 March 2014 (UTC)
 * Nope, no conflating. "Revert" in this context has to do with the 3RR policy, and "recent" is not what makes a revert; and arguing that it does make a revert is opposite of what you stated on the matter before. In that discussion, you argued "any modification of existing material is a 'revert'." I disagreed. Like I stated, "Going by that definition means that any extensive changes we make to an article, such as extensive copyediting, are a bunch of reverts." Either way, and as you know, your edit was reverted by a different editor, partly because that editor agrees with me about "recent" already being clarified by the 24-hour time frame. Flyer22 (talk) 16:37, 11 March 2014 (UTC)
 * , I'm not sure that's clear; after all, Bbb23 used my first edit as a revert that counted towards 3RR, but what recent edit was I reverting? GabeMc  (talk&#124;contribs)  16:53, 11 March 2014 (UTC)
 * GabeMc, like I already stated to North8000 above, "if certain people could not correctly comprehend the text regarding the 3RR policy's time frame before your addition, that is the fault of those certain editors, not the 3RR policy." I don't see how any old edit, clearly outside of the 24-hour time frame that 3RR makes explicitly clear, can be taken as part of a 3RR violation. Flyer22 (talk) 16:57, 11 March 2014 (UTC)
 * Well, Bbb23 used it against me and that there isn't anything specific to point to when asserting that he was wrong means we need clarification. According to Bbb23, an edit made 10 years ago cannot be changed unless you first revert. The clarification is needed, IMO. GabeMc  (talk&#124;contribs)
 * The 24-hour time frame is specific and shows that he is wrong. No WP:Administrator has used WP:3RR in that way, to my knowledge; you indicated similarly in your initial post in this section above. Get WP:Consensus to add "recent" if you want. But it is ridiculous to me to bend over backwards to "clarify" something simply because one WP:Administrator or even a few have interpreted policy differently than the vast majority of people at this site interpret it. Editors interpret our policies and guidelines differently all the time, often to suit their own needs; that does not mean that we should alter our policies and guidelines each and every time, or in any case, to try and address the few who see things differently. Perhaps Bagumba, who reverted North8000 on the "recent" addition, has something to state on this matter. Flyer22 (talk) 17:15, 11 March 2014 (UTC)
 * The 24-hour time frame refers to the 3 reverts, not the text being reverted. GabeMc  (talk&#124;contribs)  17:19, 11 March 2014 (UTC)
 * The 24-hour time frame is about what counts as a WP:3RR violation; policy is clear that any revert outside of that time frame is not a part of a WP:3RR violation, but is still edit warring. Splitting hairs about that policy in any way makes no sense to me. Flyer22 (talk) 17:24, 11 March 2014 (UTC)
 * Flyer, the concept of a revert is not just about 3RR. This entire policy is about reverting.  You cannot, by definition, have an edit war, unless someone is reverting.  It doesn't matter if those reversions happen in the space of 24 minutes or 24 days:  reverting is required for edit wars, and edit wars (the subject of this policy) do not have a 24-hour time limit.  WhatamIdoing (talk) 17:03, 12 March 2014 (UTC)
 * WhatamIdoing, of course the concept of a revert is not just about 3RR. I didn't state otherwise above. I stated that the 24-hour time frame above makes explicitly clear what is and is not a WP:3RR violation; I stand by that. Everything else on this matter is hair splitting and needless drama. Flyer22 (talk) 04:59, 13 March 2014 (UTC)


 * How about some language that puts the onus on the blocking admin to explicitly state which editor was reverted? After all, if an edit is a revert, then you should be able to name who was reverted. If a blocking admin cannot state who was reverted, then the edit was a change not a revert. GabeMc  (talk&#124;contribs)  16:33, 11 March 2014 (UTC)
 * Great idea, but that should be a part of a broader wiki-right. If you are going to get smacked, they must be specific on telling you what you did wrong.    I also think we still need the "recent".  North8000  (talk) 16:37, 11 March 2014 (UTC)
 * Well, I think its infinitely more likely to achieve consensus on some additional guideline language then it is to secure that as a user right. All we need to do is add that language or language like it to 3RR. Then any admin who fails to provide the required information would be breaking 3RR. GabeMc  (talk&#124;contribs)  16:51, 11 March 2014 (UTC)


 * Or how about some language that differentiates between changes and reverts? Is it accurate to state that a revert restores text to a previous state, but a change alters prose without substantively reversing the text's meaning?


 * Original text: "John Smith is a dull person."


 * Revert: "John Smith is a brilliant person."


 * Change: "John Smith has been described variously as intelligent and dull."


 * Original text: "The guidelines have advice on when to revert."


 * Revert: "The guidelines are helpful."


 * Change: "The guidelines include advice on when its appropriate to revert."

If we can achieve some clarity regarding the difference between reverts and changes, this will be a non-issue. Also, we might want to stress that one of the hallmarks of a revert is when text is reverted to a previous state.


 * Original text: "Everyone can edit Wikipedia."


 * Change: "Wikipedia is a encyclopedia that everyone can edit."


 * Revert: "Everyone can edit Wikipedia."

We need to differentiate between minor copyediting and reversions, which should be defined more clearly in the 3RR policy as only edits that restore or nearly restore language to a previous state. GabeMc (talk&#124;contribs)  16:47, 11 March 2014 (UTC)

Editor 1 writes: Country X is stupid. Two years later Editor 2 changes it to: Country X is not stupid. 3 minutes later Editor 3 reverts to: Country X is stupid.
 * I think it'd be a mistake to change the policy. I understand that GabeMc's case was borderline, but it's obvious from the page history that there was edit-warring going on.  Possibly it would have been more fair to have blocked both editors or neither.  However, hard cases make bad law, and the distinction made above between "change" and "revert" would I think only lead to more gaming of the system than we already have.  People routinely edit-war over content much more trivial than the examples given above.  Here's what's really going to happen if we put an expire-by limit on the first revert (imagine we're subject to a 1RR for clarity):
 * If we adopt the proposal here, Editor 2 is at 0 reverts while Editor 3 is at 1. This privileges the new version over the status quo ante.  However, BRD privileges the status quo ante, what then?  Also, if the codification of what counts as a revert are refined, the gaming will be refined too.  Anyway, the 1RR in ARBPIA articles has done wonders for that area, although there are still a lot of problems.  Most editors there understand the rules quite clearly, and putting some expiration date on that first revert would mess things up badly by giving the advantage to the first editor to make a stupid change in something old.  If people just don't edit-war, these kinds of issues wouldn't come up without clear administrator abuse, which can then be dealt with through community processes.&mdash; alf laylah wa laylah (talk) 17:18, 11 March 2014 (UTC)
 * I think WP:CONSENSUS remediates the above concerns. It doesn't matter how many reverts you have left; if consensus is against your edits then they will be undone. Your arguments above sound like content creation is no more than a 3RR numbers game, but in practice it isn't. An edit made to text written 10 years ago is not a revert, since editors are encouraged to copyedit and discouraged to revert. Also, I don't think that BRD privileges the status quo ante. It encourages bold edits, but hedges its bets against edits that are inconsistent with CONSENSUS. GabeMc  (talk&#124;contribs)  17:26, 11 March 2014 (UTC)


 * WP:BRD specifically says: Discuss the edit, and the reasons for the edit, on the article's talk page. Do not continue to revert, which is the beginning of edit-warring. Leave the article in the condition it was in before the Bold edit was made (often called the status quo ante). When the discussion has achieved mutual understanding, attempt a new edit that will be acceptable to all participants in the discussion. Your proposal makes this impossible in 1RR situations, where it's needed the most.  I'm not talking about a numbers game at all, actually.  Really, you're upset about your block, and I can sympathize completely, but I don't think we need to rewrite a rule that's generally working quite well.  The danger of unintended consequences is too high for my taste.&mdash; alf laylah wa laylah (talk) 17:40, 11 March 2014 (UTC)
 * I'm actually not upset about my block, which I don't really care about. What I care about is the fact that my first edit to any page on any day is a revert, and indeed every edit I make to existing text is a revert according to Bbb23. Have you ever copyedited an article to the FA level without making more than three changes to it in any one day? Copyediting Wikipedia is encouraged, but reverting is discouraged; therefore they are not synonymous as the guideline currently suggests. GabeMc  (talk&#124;contribs)  18:07, 11 March 2014 (UTC)
 * No, I haven't. Have you ever been blocked for copyediting when you were really, truly copyediting?  Has anyone?  As I understand it in the case that led to this discussion, you brought a report to ANEW and ended up blocked.  The point is that *you* thought there was edit-warring happening at that page too.  Normal collaborative copy-editing doesn't end up at ANEW in the ordinary course of things.  If it does, and for some reason some administrator takes such a complaint seriously, then there's a problem with the administrator, not with the rule.  It's too risky to change established rules to solve problems that we don't even really know exist. alf laylah wa laylah (talk)18:23, 11 March 2014 (UTC)
 * Well, IMO, I was improving and copyediting the project page and Static was edit-warring to stop me. I still see nothing contentious in any of the edits that I made. GabeMc  (talk&#124;contribs)  18:36, 11 March 2014 (UTC)
 * BRD also explicitly says that not everyone either can or should follow it. We cannot and should not require its use.  WhatamIdoing (talk) 17:08, 12 March 2014 (UTC)

Language proposal
I suggest that North8000 was taking the correct approach with this edit, which added the word recently in an attempt to clarify that an editor's actions should only be characterized as reverting if the changes are made to recently added text.


 * Existing language of WP:3RR: "An edit or a series of consecutive edits that undoes other editors' actions—whether in whole or in part—counts as a revert."
 * Proposed language: "An edit or a series of consecutive edits that undoes other editors' recent actions—whether in whole or in part—counts as a revert."

Examples
 * User:Example added a sentence in 2011. You remove it in 2014.   Not counted as a "revert" as far as the edit-warring policy is concerned, because you didn't reverse a "recent" edit.
 * User:Example added a sentence on Monday, 03 March. You removed it on Friday, 07 March.  ✅ Counted as a "revert" under this policy, because you reversed a "recent" edit.

Support

 * 1) Support - Without the qualifier recent, every edit made by every editor that changes any text is a revert. The idea that a copyedit of material from an unknown date and an unknown source is a revert is ridiculous, and it starts every editor off with one strike against 3RR. We can define recent as 24 hours, or longer, if we want. Maybe recent is anything that can be directly attributed to a specific editor.  GabeMc  (talk&#124;contribs)  17:10, 11 March 2014 (UTC)
 * 2) Support The current wording is terrible; where everybody is technically breaking the rule.  Rules should be written so the you can actually follow them as written, not be forced to routinely violate them.  And don't forget that this same definition is used for 1RR and 0RR.   North8000  (talk) 17:28, 11 March 2014 (UTC)
 * 3) Support, and note to Laylah below: this has come up because the *lack* of "recent" was used to game the system. If you let a rule lean too far towards being able to catch bad guys, you end up with a rule that can be used to catch anyone and which is even more prone to gaming than having balance.  Letting admins claim that every change is a revert is ridiculous.  Ken Arromdee (talk) 19:45, 11 March 2014 (UTC)
 * 4) As long as we understand that "recent" involves in the last few weeks, rather than the last few hours, then yes.  You are not "reverting" or "edit warring" if you remove something that was added last year (especially if you are removing something unsourced, in which case, what a few admins would call your "revert" is a policy-protected WP:CHALLENGE).  WhatamIdoing (talk) 17:10, 12 March 2014 (UTC)
 * 5) Support. Recent is desirable to avoid absurdity, yet not too inflexible. Rothorpe (talk) 20:15, 12 March 2014 (UTC)
 * 6) Support. To address the objections below: 1) the system can be gamed on both sides here. See the "more problems with the definition" section earlier on this page.  If you really want to minimize end runs around the system, you reduce the importance of the "system" in deciding what is an infraction and elevate the importance of common sense.  2) there has not been just one misunderstanding here.  You would not have so many pissed off editors here calling for change if just one of us has been wronged.  3) the whole idea behind 3RR is to slow down edit wars; 3RR does not, in and of itself, end those edit wars apart from slowing them down.  4) if editors are following BOLD, revert, discuss cycle, the edit count is of secondary importance.  I read Robert McClenon's remarks as saying "well, violating 2RR is bad too."  Of course, but whether it should be 2RR or 3RR is a separate issue.--Brian Dell (talk) 23:13, 17 March 2014 (UTC)

Oppose

 * 1) Hard cases make bad law.  I think the 3RR works fine, and, especially, changing the definition of revert in this way would lead to system-gaming and would also play havoc in areas subject to a 1RR, where that restriction has largely been for the good.  More details in my recent comment above.&mdash; alf laylah wa laylah (talk) 17:20, 11 March 2014 (UTC)
 * 2) We ought not to change the language because of one misunderstanding. A revert involves undoing another editor's work. That's clear enough. It needn't be recent; there are long-term, slow edit wars that involve reverting back to an old disputed version. SlimVirgin (talk) 17:24, 11 March 2014 (UTC)
 * How about because of half a dozen "misunderstandings" involving more than one admin? I see specific complaints that name two admins, including one who insisted at ANI that "any change" is "technically a revert".  If "any change" is truly "a revert", then everyone on WP:MOSTEDITS should have been blocked long ago, including you and me.  And if that's not the true definition, but we have a couple of admins claiming that it is, then maybe we need to correct their "misunderstanding"—of this policy and of what the community considers to be a revert.   WhatamIdoing (talk) 17:14, 12 March 2014 (UTC)
 * 1) There are slow edit wars where reverts wouldn't necessarily be limited to "recent" edits. I'll restate my comment from earlier, any disagreement from a specific incident with an admin should be pursued per WP:ADMINABUSE. 3RR already states that the timespan is 24 hour; let's not add ambiguity to "revert" that might lead to WP:LAWYERING over what time-span applies to reverts for other cases (e.g. WP:0RR).—Bagumba (talk) 18:07, 11 March 2014 (UTC)
 * 3RR says that the timespan *between the reverts* is 24 hours, not that the timespan *between the revert and what was reverted* is 24 hours. This is about the latter. Ken Arromdee (talk) 07:48, 13 March 2014 (UTC)
 * 1) The existing rule is reasonable, because either 3RR or 1RR only kicks in after the editor makes the second or fourth revert. If an editor changes ancient text, which under the current language is technically a revert, and it is then restored twice and she reverts it both times, she has reached 3RR.  It isn't unreasonable to require that she desist from further edits for that day and discuss on the talk page instead.  If 1RR is in effect, and she changes ancient text, and her change is restored, it isn't unreasonable in an area subject to sanctions to expect her to discuss and everyone to cool down.  Robert McClenon (talk) 23:40, 13 March 2014 (UTC)
 * 2) This change would only open the door to wikilawyering on what "recent" means, and the more specific you make it, the more unfair it becomes. The purpose of the policy isn't to pigeonhole editors or admin, but to provide structure for the application of common sense.  Any policy can be nit picked, we rely on the good judgement of each other to apply it.  I would note that 4 reverts are not required before an admin blocks, nor are admin required to block after a 4th revert.  It is just a clear and obvious line that is acceptable under most circumstances, and nothing more.  Dennis Brown &#124; 2¢ &#124;  WER  00:28, 19 March 2014 (UTC)
 * 3) Oppose First I think this is a very good discussion and I applaud everyone for their thoughtful contributions. That said, I am not comfortable with the addition of the proposed term. In general when crafting rules or guidelines, especially rules to which punitive sanctions might be applied, it's usually best to minimize the use of subjective terminology. "Recently" is a tad too vague for my comfort level. -Ad Orientem (talk) 01:08, 26 March 2014 (UTC)
 * 4) Oppose. This adds subjective language that doesn't appreciably improve policy.  Although I can see the intent, I don't think that we need to be any more explicit that normal copy editing is permissible.  Real-time, contemporaneous edit wars are more problematic and common, but I don't think we should open the door any wider to slow motion edit wars, righting great wrongs (or minor wrongs, in this case), or wikilawyering. NinjaRobotPirate (talk) 05:04, 30 March 2014 (UTC)

Discussion

 * , doesn't WP:CONSENSUS remediate the concern regarding those slow edit wars of which you speak? Why would we need 3RR, or why would it be a relevant guideline, to say that the old version is no longer the accepted community consensus? If someone went to the Beatles page and changed the "t"s to "T"s we wouldn't trace back to the edit-wars from two years ago, we would simply state that the current consensus is in favour of using lower-cased "t"s. GabeMc  (talk&#124;contribs)  17:32, 11 March 2014 (UTC)


 * It would depend on the situation. The current language has worked well for years, and grey areas can be sorted out by the admin the violation is reported to; most admins err on the side of caution. If you introduce the idea of "recent" actions, it adds a layer of confusion. Bear in mind that people can be blocked for edit warring without four reverts anyway. SlimVirgin (talk) 17:41, 11 March 2014 (UTC)
 * Right, but calling the first copyedit a reversion is nonsense when the identity of the reverted editor is unknown. After all, if we don't know who created the original text, then why is it a revert? Also, adding recent would clarify that as an editor begins working on an article that has sat neglected for years they need not count each and every change as a revert. Without the qualifier recent, any and all content editors could be blocked for improving articles, since rarely does a copyeditor self-limit to 3 changes per day per article (I've made as many as 75 in a day). At the rate of three changes per day, it would have taken me more than 3 years to copyedit Paul McCartney. GabeMc  (talk&#124;contribs)  17:54, 11 March 2014 (UTC)


 * If a tree falls in the woods: Another complication is that if we count changes made to material that was added 1–5 years ago, then how will a blocking admin know for certain that the editor in question isn't self-reverting? Per WP:NOT3RR: "The following actions are not counted as reverts for the purposes of 3RR: Reverting your own actions ('self-reverting')." The longer the span of time since the material was added the greater the chance that someone could be blocked for editing material that they originally added. Why not adopt the "if a tree falls in the woods" approach, whereby an edit cannot be said to be a revert unless the editor who was reverted can be named? GabeMc  (talk&#124;contribs)  18:03, 11 March 2014 (UTC)
 * , you keep stating: "3RR already states that the timespan is 24 hour", but the 24 hour period refers to the reverts, not the edited material. If any change to existing material is a revert, then editors cannot make more than three non-consecutive edits to any page ever, right? GabeMc  (talk&#124;contribs)  18:13, 11 March 2014 (UTC)


 * When blocking for 3RR the admin needs four examples of another editor's work being undone. But admins won't (as a rule) wikilawyer about this or try to find purely technical reverts. The point is always that there was disruption. Sometimes there will be a block for disruption because of 4 reverts within 24 hours; at other times it will be because of disruptive editing without having crossed that line. The admin who blocked you must have felt there was something disruptive about what you were doing. If you disagree there are other places to discuss that (e.g. AN/I), but the wording of the 3RR section of this policy wasn't the issue in your case (so far as I can see). SlimVirgin (talk) 18:19, 11 March 2014 (UTC)
 * , actually the blocking admin used my first edit to the page as counting towards the 3RR violation, so "the wording of the 3RR section of this policy" is an issue in that it implies that any change to existing text is a revert even if the editor who is theoretically being reverted is unknown. GabeMc  (talk&#124;contribs)  18:24, 11 March 2014 (UTC)


 * I didn't see the report so I don't know what was counted as a revert. I glanced at your edits and I can't see a 3RR violation. But I also glanced at the talk page and apparently you reported someone for edit warring for having reverted you (I'm writing from memory). So an admin saw this and felt you were unilaterally changing a WikiProject page over objections, and that it was disruptive editing. Perhaps he interpreted the other editors' objections to their work being undone as a revert, so where you saw a copy-edit, they saw the removal of a consensus version. I'm only guessing, though. I wouldn't have interpreted your first edit as a revert, unless you restored previous wording. But the wording of this policy wouldn't have made a difference because the issue is that someone saw your edits as disruptive, rather than as merely bold. SlimVirgin (talk) 18:40, 11 March 2014 (UTC)
 * Well, if I was changing the agreed upon consensus with my first edit to the page then the agreed upon consensus there is that the project language trumps all Wikipedia policies and guidelines, which I thought was inappropriate if not impossible. The intention of my edits was to make it clear that the project advice is not to be seen an alternative to policy, guidelines, or the MoS. I would say that if one cannot point to a discussion where said consensus was clearly established then per WP:SILENCE the assumed consensus no longer exists after dissent is voiced, which my first edit accomplished. GabeMc  (talk&#124;contribs)  18:50, 11 March 2014 (UTC)

(edit conflict) There are a couple of things that I think that the above discussions are forgetting. One is that this definition of a revert is also used for 1RR and 0RR. The second is that rules are mis-used (to conduct warfare, e.g. to deprecate people who they view as an "opponent") as often as they are used for their intended purpose. And in warfare (particularly on 1RR articles) I have seen somebody (who didn't like ONE revert that an editor made) look back many years to find the edit that a different gnome edit that they made modified, thus making the gnome edit a "revert" so that they could report someone who made ONE revert as having made two reverts. And technically, they did violate the rule as written, and a responsible editor had to go through all of that stress, so no, the system isn't working. A system that defines normal every day behavior as actionable mis-behavior is a BAD system.....that's how police states operate. Rules should be written so that people can follow them, as written  North8000  (talk) 18:20, 11 March 2014 (UTC)


 * It's 1RR cases that specifically worry me about the proposed change. In warfare any rule whatsoever will be gamed.  Do you not think that 1RR with the current definition works pretty well in ARBPIA areas?  I think adding another opportunity to grasp at technicalities will only increase the level of argument and won't stop determined warriors.  But we already have means to stop warriors.  This won't help us there.  Also, the current rule and the proposed rule are about equally hard to understand, so the police state comparison is kind of over the top in my opinion.&mdash; alf laylah wa laylah (talk) 18:27, 11 March 2014 (UTC)
 * Maybe with my lack of clarity it was over the top. What I meant is that particular aspect (write the rules so that everybody violates them and then everyone has to depend on the mercy of the police) mimics the police state approach.  North8000  (talk) 18:47, 11 March 2014 (UTC)
 * Why would a one-word addition to the 3RR guideline mess-up 1RR? Don't they already have separate definitions? GabeMc  (talk&#124;contribs)  18:32, 11 March 2014 (UTC)
 * They don't have separate definitions, and I gave an explicit example above of how it would mess it up. It would make BRD impossible, because the first revert wouldn't count as a revert so if there were only two editors involved it would make it impossible to revert to the status quo ante for discussion.&mdash; alf laylah wa laylah (talk) 18:36, 11 March 2014 (UTC)
 * That doesn't make any sense to me. Are you saying that if an article has a 1RR restriction then editors are only allowed to make one change to it per day? The issue here is that change and revert are not the same things because the community encourages the former and discourages the later. How are articles with 1RR improved if only one edit per day per editor is allowed? GabeMc  (talk&#124;contribs)  18:45, 11 March 2014 (UTC)


 * It works pretty well. I managed to do over 100 edits to Rachel Corrie, all copyediting, during a massive week long edit-war sparked by an Israeli court verdict in her case, and get barnstars from both "sides" in the process, by just stopping for the day every time someone made another edit.  I could do fifteen or twenty small edits in a row.  The rule works really well for damping down the agony on those pages, and it's explicitly understood that one's first edit to the article counts as a revert unless its a straight-up addition of new material.  You should go edit some articles covered by ARBPIA for a while to get a feeling for what might happen if your proposal were adopted.&mdash; alf laylah wa laylah (talk) 18:50, 11 March 2014 (UTC)


 * On 1RR you can make as many changes as you like, making sure you don't restore previous wording or do anything that could reasonably be interpreted as undoing someone else's actions. Note the word "reasonably." If you remove an image, that doesn't count as reverting to a version from 10 years ago that had no image. But if the image was an issue at some point, and if there's reason to believe that the editor who removed it knew it was an issue, then one removal might indeed count as a revert. That's where an admin has to exercise judgment.


 * Perhaps you (Gabe) want a policy that won't ever rely on an admin's judgment, but bot-like blocks for reverting would be worse than the current situation. We do want people to consider the context, and admins usually give the benefit of the doubt. SlimVirgin (talk) 18:57, 11 March 2014 (UTC)
 * Per North8000: "What I meant is that particular aspect (write the rules so that everybody violates them and then everyone has to depend on the mercy of the police) mimics the police state approach." I have to agree. With an open-ended standard that every single change made to any article for any reason is a revert so long as the blocking admin sees it that way gives them total carte-blanche. So yes, I agree that bot-blocks would be a bad thing, but to say that content editors are always at the mercy of admin discretion is beyond creepy, IMO. If Bbb23 is correct, then I could have been blocked multiple times for every article that I've ever improved to FA, because I often change an article prose more than three time in one day. Have you ever achieved FA status at an article that you never made more than three non-consecutive edits per day? GabeMc  (talk&#124;contribs)  19:04, 11 March 2014 (UTC)
 * Straw man argument going on here, Gabe. :) No one interprets every edit as a revert. Your first edit in this case was (perhaps) interpreted that way because you changed the consensus version of a WikiProject page, and didn't stop when objections were made. SlimVirgin (talk) 19:14, 11 March 2014 (UTC)
 * Re-read Bbb23's comments at the top of this thread and then tell me that he doesn't count every change as a revert. GabeMc  (talk&#124;contribs)  19:16, 11 March 2014 (UTC)


 * I don't agree with his interpretation as written above, but there may be a context I'm unaware of. I would count your first edit to that page as a revert if (a) it restored earlier language, or (b) there had been an issue with the lead (or page) going back and forth, people had found a consensus version, and you changed it unilaterally, even if not reverting to earlier words (or the removal of consensus words). If neither of those things pertain, I would not count your first edit as a revert. SlimVirgin (talk) 19:28, 11 March 2014 (UTC)
 * My first edit did not re-introduce any previous language nor was it part of a back and forth, because it was a BOLD edit before any disagreement had taken place. As far as I can tell, I only reverted once, maybe twice, but not three or four times unless you can count the first and last edits, which both introduced new language that had not yet been reverted. Bbb23 counted the first edit because I was changing what had been written, which I think violates WP:BOLD. Editing Wikipedia is encouraged, so how can the first edit be a revert which is generally discouraged? GabeMc  (talk&#124;contribs)  19:36, 11 March 2014 (UTC)


 * According to Bbb23, this edit and this edit are reverts that count against 3RR, but the thing is, I'm the one who added the text to the article in the first place, so the potential for a 3RR violation is increased by copyediting your own prose, especially if you added it so long ago that finding a diff would be extremely difficult, hence the need to specify that 3RR only applies to recent edits. Also, if Bbb23 is correct, then I can make only two more changes to this article today without violating 3RR. GabeMc  (talk&#124;contribs)  18:30, 12 March 2014 (UTC)

Language proposal (b)
Per WP:EW: "A revert means undoing the actions of another editor." In this statement, revert is linked to Help:Reverting, which states: "Reverting means undoing or otherwise negating the effects of one or more edits, which results in the page being restored to a previous version. Partial reversion involves restoring one part of the page to a previous version, but leaving other contributions intact." Per Revert only when necessary: "Reverting is reversing a prior edit. See Reverting. You should generally avoid reverting." However, per WP:BOLD: "Wikipedia not only allows you to add, revise, and edit articles: it wants you to do it." So the notion that any change = a revert contradicts Wikipedia guidelines. Reverts are not equal to changes. Reverts are specifically edits that restore or partially restore previously removed material, whereas a change includes the introduction of new material and/or the copyediting of existing prose for grammar and clarity; edits made for this reason should not be characterized as reverts that count toward a 3RR violation.


 * Existing language of WP:3RR: "An edit or a series of consecutive edits that undoes other editors' actions—whether in whole or in part—counts as a revert."
 * Proposed language: ''"An edit or a series of consecutive edits that undoes other editors' actions—whether in whole or in part—counts as a revert. An edit that does not remove, restore, or partially restore previously included material should not be characterized as a revert."

Support (b)

 * 1) Although the exact Wikipedia definition of revert is debatable, the key element is that it "restores or partially restores" previous language that has since been removed. Introducing new language to a page or copyediting that page for grammar and/or clarity is not a revert, unless we are seriously twisting the definition of that word.  GabeMc  (talk&#124;contribs)  19:43, 11 March 2014 (UTC)
 * 2) I support what GabeMc is trying to do here.  Once Gabe has made a diff to the policy, complainants can raise the diff here and offer their specific complaints instead of objecting at this point that the specifics are unclear.  re "hard cases make bad law" then don't make the law.  Apparently I'm ineligible for template editing rights because of the one dubious block I have received in my editing career.  Stop having the whole Wiki world turn on a technicality and there wouldn't be a need to put some flex in here.--Brian Dell (talk) 23:24, 17 March 2014 (UTC)

Oppose (b)

 * 1) I think this is an example of hard cases make bad law. SlimVirgin (talk) 19:49, 11 March 2014 (UTC)
 * 2) This doesn't make sense to me. The only way I can interpret it sensically makes it logically equivalent to the existing definition, which I assume can't be your intention.&mdash; alf laylah wa laylah (talk) 20:45, 11 March 2014 (UTC)
 * 3) No evidence that this specific WP:ADMINABUSE dispute is a general problem that needs repetitive language inserted.—Bagumba (talk) 22:16, 11 March 2014 (UTC)
 * 4) Unnecessary and not helpful. Dennis Brown &#124; 2¢ &#124;  WER  00:30, 19 March 2014 (UTC)

Discussion (b)
This one isn't phrased well. Deleting material that someone else added is also a revert. Ken Arromdee (talk) 19:46, 11 March 2014 (UTC)
 * Well, if you added "John is nice", and I removed it I would be restoring the material to a previous version. Also, the rest of the guideline covers this, so that aspect would be unaffected by this proposal. GabeMc (talk&#124;contribs)  19:55, 11 March 2014 (UTC)
 * , does this edit resolve your concern with the proposal? GabeMc  (talk&#124;contribs)  20:07, 11 March 2014 (UTC)
 * This is already covered. It would be an edit that "remove[s]...previously included material".  WhatamIdoing (talk) 20:49, 12 March 2014 (UTC)


 * , I know that Reverting is an essay, but its language strongly supports what I'm saying here: "Technically, any edit can be said to reverse some of a previous edit; however, this is not the way the community interprets reversion, because it is not consistent with either the principle of collaborative editing or with the editing policy." Is there anyway that this distinction can be made here? GabeMc  (talk&#124;contribs)  20:01, 11 March 2014 (UTC)


 * You want to tighten things because you had a bad experience, but you risk causing further wikilawyering in future. I think instead you should challenge that admin's interpretation so that he doesn't make it again (if we've understood him correctly), but otherwise move on from your block. Not every change is a revert. A revert undoes someone's work (their addition or their removal of material) in a meaningful sense, not in the empty technical sense of "every time an article is created it reverts someone's decision not to create that article." SlimVirgin (talk) 20:17, 11 March 2014 (UTC)
 * Although given the realities of special:newpages that might be a pragmatic interpretation...&mdash; alf laylah wa laylah (talk) 23:18, 11 March 2014 (UTC)
 * Well, first of all the block really doesn't bother me; I am not a candidate for a popularity contest nor will I be anytime soon, so please don't assume that my only motivation here is revenge for the block. My motivation is that Bbb23's interpretation cannot be challenged by anyone due to the guideline's vague language, thus he will inevitably block others for similarly poor reasoning. A guideline that can be used to block content editors should not be so vague, IMO; a block is a serious thing that should be based on community consensus, not every admin's personal interpretation, which tend to be self-serving. If an edit does not "remove, restore, or partially restore" previously included material than how can it be said to be a revert? What's wrong about wanting to clarify this here so that next time a bad block is given the victim can point to something in the guideline that explicitly defends their position? GabeMc  (talk&#124;contribs)  20:31, 11 March 2014 (UTC)


 * , actually yes, my intention is to bring the 3RR guideline in-line with the community's working definition of revert since the guideline as currently written is obviously grey enough that its left room for interpretation to the point that Bbb23 counts any changes as reverts. GabeMc  (talk&#124;contribs)  20:55, 11 March 2014 (UTC)
 * I'm not sure why this is even contentious. WP:EW links to Help:Reverting, which links to Reverting, which states: "Technically, any edit can be said to reverse some of a previous edit; however, this is not the way the community interprets reversion, because it is not consistent with either the principle of collaborative editing or with the editing policy." So, why can't we make that clear here at WP:3RR? GabeMc  (talk&#124;contribs)  21:26, 11 March 2014 (UTC)
 * I edited that paragraph just the other day. Previously, it held forth the nonsensical view that every single edit could, if you squinted hard enough, be considered a "revert" because you are "reversing" the previous editor's alleged "choice" not to include any material that didn't make it into the first saved version of a page, or, I suppose even the first version, because when you create a page, you are "reversing" someone else's choice not to do so.  Here we were, all thinking that SlimVirgin was an excellent content contributor, but in that "reality", she's just someone who has somehow gotten away with making 135,000 "reverts".  That definition is nonsense, not supported by the community, and should be explicitly rejected.  WhatamIdoing (talk) 20:49, 12 March 2014 (UTC)
 * Bagumba, I think that a prolific content editor with a clean block log getting blocked—in part—because of the guideline's vague language is prima facie evidence that some admins will Wikilawyer this loop-hole to their advantage. If a loop-hole is evident then it should be closed; redundancies are better than contradictions because they reinforce the consensus opinions instead of confounding them. GabeMc  (talk&#124;contribs)  22:22, 11 March 2014 (UTC)
 * Respect your opinion, but the explanation at AN3 hinted that the blocking admin believed more than 3RR was involved. At this point, I think your position is clear.  I'd suggest waiting for others to chime in to establish consensus whether anything needs to be done to this policy.—Bagumba (talk) 22:52, 11 March 2014 (UTC)
 * That sounds like that falls under "everyone is guilty, so whether you're sent to jail depends on police discretion", and he's just explaining his use of discretion. The official explanation for the block was that an admin can block someone for 4 reverts whenever he wants.  It's not as if his decision that something more than 3RR was involved can be challenged--he claims that he can block a user for 4 reverts whether there's something more or not. Ken Arromdee (talk) 07:44, 13 March 2014 (UTC)

Another suggestion
Regardless of who is right or wrong about whether GabeMc violated 3RR in this specific case, the situation illustrates that good-faith editors can disagree about what constitutes a revert, or three reverts.

I will repeat here a suggestion I have made several times before, which is that when an editor (particularly one without an edit-warring history) appears to have overstepped 3RR inadvertently, especially in a borderline situation, administrators should be encouraged to notify or warn the editor rather than block.

In cases of necessity, perhaps administrators could be empowered to order a user off a specific page for 24 hours, in lieu of blocking for that time, which disables the editor from working on other pages unrelated to the dispute. Newyorkbrad (talk) 01:19, 12 March 2014 (UTC)
 * That's why we shouldn't be counting anything. There should be no blind-bookkeeping-and-automatically-blocking-when-some-arbitrary-limit-is-reached.  Automatic penalties without any deeper analysis ALSO makes bad law.  Instead an admin needs to assess a situation, decide if an editor is using the ability to undo the actions of others a means to win a conflict through sheer force-of-will, and only block if that is what is going on.  That's edit warring, and it is bad.  If an editor is not trying to win a conflict by putting their preferred content in over-and-over, it isn't a problem.  If they are, it IS a problem, regardless of the numbers and timeframes involved.  Admins should be looking into situations, not counting edits and times and blindly hitting the block button when some random limit is reached (or refusing to block if the limit isn't).  -- Jayron  32  01:44, 12 March 2014 (UTC)
 * It would be difficult for admins to be that stringent if they are not that familiar with the domain of a given article. That's probably where 3RR came from.  WP:AVOIDEDITWAR provides a number of alternatives to continued reverting before 3RR becomes an issue.—Bagumba (talk) 01:57, 12 March 2014 (UTC)
 * That's one of many reasons why admins who do not themselves have considerable experience developing non-trivial content should never be permitted to block long term productive content builders. --Epipelagic (talk) 13:54, 12 March 2014 (UTC)
 * Could you please elaborate on your feelings here? I don't see how an admin's contributions to the project necessarily have any bearing on whether they're qualified to judge whether an editor is edit-warring, and in my experience even the most productive of editors are capable of edit-warring, perhaps precisely because they've become too invested in the articles they've developed. Thanks. DonIago (talk) 14:28, 12 March 2014 (UTC)
 * Well it seems to be a general pattern – that the worst blocking admins rarely seem to have much experience with building non-trivial content themselves. Vandals and productive content builders seem to blur in their minds, and they start treating them alike. An admin who has worked hard building content is more likely to have nuance and respect for the process, and be less likely to make knee jerk blocks when a more skilful approach that doesn't require blocking can be just as effective. Anyway, individual admins should never be blocking productive content builders just at their own whim. Actions as draconian as that should be supervised by a disciplinary panel. --Epipelagic (talk) 14:49, 12 March 2014 (UTC)
 * Hm. Fair enough, I guess, though this sounds a bit more like theory than established fact to me. I don't really see how whether someone's effective at building content indicates that they're less prone to edit-warring though; I could see an editor getting defensive about their content and consequently exercising inappropriate levels of ownership. I'd think things like 3RR are clear-cut enough that both editors should be able to steer clear of it and admins should be able to determine when it's been violated. As for a block being "draconian"...personally I view blocks up to, say, a week or so as being little more than a "slap on the wrist". I mean, really, it's a whole seven days...surely it won't kill someone to be unable to edit articles for such a brief period of time.
 * All that said, I would be amenable to a hard rule that an editor must have been expressly warned about their edit-warring before they can be blocked for such, and I'd support blocking an editor from working on a specific article as opposed to imposing a projectwide block. DonIago (talk) 15:20, 12 March 2014 (UTC)


 * Some commentators in the preceding thread advocate a loose policy on the grounds that hard cases make bad law. Others want a tighter policy, presumably because bad law makes hard cases. But either way, refining policy documents is not going to lead to better outcomes. The core dysfunction results from the way we use our over one thousand life-appointed administrators. Each admin is allowed individually to interpret policy according to their idiosyncratic whims, resulting in the absurd administrative anarchy we see today. It is though we have a Wikipedia police force but no judicial system. Each Wikipedia policeman acts independently as apprehender, judge, jury and enforcer. --Epipelagic (talk) 05:02, 12 March 2014 (UTC)

I think that Newyorkbrad made a good suggestion, and that Epipelagic made some good and important points. But IMO neither is a good reason for not fixing a terribly written definition.  North8000  (talk) 11:30, 12 March 2014 (UTC)


 * Yes, of course policy should be improved where it can. But the problem is not so much with the policy documents, which are generally quite good. The problem is with the implementation of policy, with the manner in which policy is mishandled by the admin corps. There are also problems with the lack of some sort of constitution which could provide background perspective and function like a rudder. There is, for example, nothing equivalent to a mission statement for admins, so admins don't really know what they are here for. Some admins seem to think they are here to block as many users as possible. Nor is there some sort of balancing bill of rights for content builders. Some admins treat content builders as though they are no different than vandals.
 * Newyorkbrad's comments are excellent, and I move that something to that effect be added to WP:EW. --Epipelagic (talk) 14:25, 12 March 2014 (UTC)
 * I also agree with Brad. An admin should be required to provide a warning before handing out a 3RR block. I didn't know I went over, but if Bbb23 had warned me that he was counting the first edit I would have stopped and not been blocked. GabeMc  (talk&#124;contribs)  16:16, 12 March 2014 (UTC)


 * It's common for admins to ask someone not to edit a particular page for 24 hours if they've been reverting a lot. It's also standard practice to warn people before filing a 3RR report. And people have to provide a diff to the version reverted to when filing a report. So all these things already happen. This discussion was triggered by a case of one admin not doing any of them, but there's no need to change a policy because of one admin interpreting it differently. SlimVirgin (talk) 16:52, 12 March 2014 (UTC)


 * And there are cases when a warning isn't appropriate because the edit-warring is out of control or because one or both of the edit-warriors make a report at ANEW, in which case one must assume they know the 3RR (this comment is not in reference to the specific case that instigated this discussion). I'm just worried that these proposals are going to end up adding a lot of qualifications to the rule, which in turn will create a lot of unintended loopholes and misunderstandings.  I think the rule generally works quite well in practice and is clear enough.  I'm opposed to changing it without overwhelming evidence that there's a serious problem with it.  I just don't see that there is.  Whether or not  is right about administrator abuse, abusive administrators will abuse any rule.  It's not possible to make rules so clear that they can't be abused, and the more specific they're made in futile attempts to do this, the more unintended opportunities for abuse we're likely to see.&mdash; alf laylah wa laylah (talk) 17:04, 12 March 2014 (UTC)
 * Those are exceptions. If an admin decides to take the time to block someone then they should first provide a warning. Why is this controversial? RE: "It's not possible to make rules so clear that they can't be abused", is a wild generalization. All laws of prohibition do just that. Is a speed limit gameable? GabeMc  (talk&#124;contribs)  17:16, 12 March 2014 (UTC)


 * I'll reply at your talk, which feel free to ignore if it's boring, since it's so off-topic.&mdash; alf laylah wa laylah (talk) 17:54, 12 March 2014 (UTC)


 * That's not a reason to stop trying to improve the policy documentation. What you say is a reason to take the interpretation of policy away from loose cannon admins, and set up instead a proper disciplinary body that can do the job sensibly. --Epipelagic (talk) 17:23, 12 March 2014 (UTC)
 * , can you please explain how requiring a warning prior to blocking changes the substance of the guideline? Admins would not need to adjust their behavior beyond providing a fair warning before blocking for all the same reasons that they used to block for. GabeMc  (talk&#124;contribs)  17:36, 12 March 2014 (UTC)
 * I didn't say that it did change the substance of the guideline.&mdash; alf laylah wa laylah (talk) 17:55, 12 March 2014 (UTC)
 * If adding a warning requirement does not change the substance of the guideline then why are you resistant to the suggestion? GabeMc  (talk&#124;contribs)  18:32, 12 March 2014 (UTC)
 * If it doesn't change the substance, then why put it in? The more words there are in a rule the more unintended consequences one can expect.&mdash; alf laylah wa laylah (talk) 21:08, 12 March 2014 (UTC)
 * The intent of the proposal is not to change the substance of why admins block for 3RR, its attempting to changing the protocol for when they block. Per WP:EW, "[blocks] are intended to prevent, deter and encourage change in disruptive behavior, not to punish it." So why would a warning be bad thing? How many blocks could be prevented by requiring a warning? GabeMc  (talk&#124;contribs)  23:08, 12 March 2014 (UTC)
 * Asked and answered, but I'll rephrase. If it's already clear that the edit-warrior knows the rules, why should it be necessary every time to issue a warning?  Plus, changing when they block evidently would require a change in substance, which you already said you weren't trying to do.  Look, I'm not opposed to your goal here, I'm opposed to the means by which you're trying to implement it.  I think any changes to 3RR would potentially have dire unintended consequences.  If you're trying to avoid bad blocks I think it'd be better to do it by trying to hold bad blockers accountable for their actions rather than by adjusting a rule that's central to the functioning of large swathes of Wikipedia and is working well, in my opinion, for the most part, even in some of the most troubled topic areas that we have.&mdash; alf laylah wa laylah (talk) 23:17, 12 March 2014 (UTC)
 * Look at it this way, the only reason why Bbb23 considered my series of edits that day to number 4 and not 1 is that Static edited during the same time frame, but how would someone even know that their series of edits was approaching a 3RR violation unless they checked the edit history after each edit to see if anyone edited in the meantime, which few if any people do. If requiring a warning is such a bad idea then I think you would be able to provide some concrete examples of what might happen. As it is your concern is quite vague. What's the potential downside to a required warning? GabeMc  (talk&#124;contribs)  23:22, 12 March 2014 (UTC)
 * I'm still not sure if your proposal is about your case or not, but look, you knew well enough that there was some controversy happening on the page to report the other editor to ANEW. If you already know that, why is it hard to check the history regularly and just stop one revert earlier?  Editors on 1RR pages do check the page history all the time.  I keep it open in another tab and refresh it before I hit save.  Obviously that's not necessary on regular articles, but seriously, the fact that you went to ANEW with diffs means that you could have started checking the page history a little bit.  It's totally possible that Bbb23's block was based on miscounting, I really don't know, but it's not so plausible that you or anyone else can not know when it's time to start checking the page history and counting edits.&mdash; alf laylah wa laylah (talk) 23:31, 12 March 2014 (UTC)
 * Its not really about me, its about the incident bringing to light a grey area that I think should be illuminated. To clarify what happened, I saw that Static had unilaterally undone my work, which I still say is uncontentious, so I reverted it in the strictest sense one time. Then I added some new language to a different paragraph and I realized that he again undid my work when we edit conflicted, so yeah, I knew there was edit warring going on because he was reverting me and edit conflicting, but I honestly thought that the bulk of my edits were not reverts, and I still do, but that's not at all the point, which is if Bbb23 had warned me I would have stopped and not been blocked, which I think is a superior outcome 9 times out of 10, especially when dealing with veteran content creators and not anons or newbies. GabeMc  (talk&#124;contribs)  23:42, 12 March 2014 (UTC)

From User talk:WhatamIdoing: Let's look at that dispute:


 * G's bold change
 * S removed only one of the new sentences ("#1")
 * G's restored sentence #1 and then made other consecutive edits
 * S reverted a change to a completely unrelated sentence ("#2") (G accepts the reversion of sentence #2, and never touches it again.)
 * G made more bold changes, all unrelated to either sentences #1 or #2
 * S removed sentence #1 again
 * G made more bold changes, again, all unrelated to either sentences #1 or #2
 * D reverted to the previous day's version

GabeMc (talk&#124;contribs)  23:42, 12 March 2014 (UTC)

Language proposal (c)
Per 's above suggestion, lets introduce language that requires an admin to warn an editor prior to blocking them for 3RR.


 * Existing language in Edit warring:
 * "Administrators decide whether to issue a warning or block"


 * Proposed language rough draft:
 * Version I: "Before an admin may block an editor for a 3RR violation they must first warn the editor at their talk page."
 * Version II: "Before an admin may block an editor for a 3RR violation, the editor must have been warned at their talk page prior to the most recent relevant revert."

I won't start any !voting until we think we have a general agreement on the proposed language. GabeMc (talk&#124;contribs)  16:16, 12 March 2014 (UTC)


 * I'd vote for that. Rothorpe (talk) 16:25, 12 March 2014 (UTC)
 * Can we rephrase this to some variant of "the editor must have been warned"? If I, a non-admin, warn an editor that they're edit-warring (say, with the uw-ew template), and they continue to edit-war, then an admin reviews the situation, I feel they should be able to act without needing to issue a warning of their own. DonIago (talk) 16:31, 12 March 2014 (UTC)
 * Good point. What do you think of version II? GabeMc  (talk&#124;contribs)  16:40, 12 March 2014 (UTC)
 * Better. Thank you. I think I'd Support that, but I'd like to see how this conversation evolves before committing. DonIago (talk) 17:59, 12 March 2014 (UTC)


 * Oppose. Usually the person filing the report warns, sometimes an admin does, but there are circumstances in which no warning is needed (e.g. for a serial reverter). Adding this to the policy would tie people's hands unnecessarily. Gabe, editors should ideally not be counting reverts. Your proposed changes would all encourage counting and wikilawyering. SlimVirgin (talk) 16:56, 12 March 2014 (UTC)
 * You're assuming that a blocking admin must file a 3RR report to block someone for violating 3RR, but as far as I can tell they do not. What's the harm in requiring an admin to give a warning before issuing a block? GabeMc  (talk&#124;contribs)  17:00, 12 March 2014 (UTC)
 * Alf.laylah.wa.laylah explains it well above. SlimVirgin (talk) 17:12, 12 March 2014 (UTC)

That's a premature oppose SlimVirgin, we haven't agreed yet on the wording. I would like to see the wording more in line with Brad's suggestion, "when an editor (particularly one without an edit-warring history) appears to have overstepped 3RR inadvertently, especially in a borderline situation, administrators should be encouraged to notify or warn the editor rather than block." Could we also include Brad's other (excellent) suggestion, "In cases of necessity, perhaps administrators could be empowered to order a user off a specific page for 24 hours, in lieu of blocking for that time, which disables the editor from working on other pages unrelated to the dispute." SlimVirgin says above that some admins already do this. But clearly others don't, so let's spell it out as a more skilful and less damaging means than crude blocks. --Epipelagic (talk) 17:10, 12 March 2014 (UTC)
 * I think some of Brad's language is far too specific and subjective. What's borderline and how far does encouraged go? I think if 3RR is a bright line for admins then content editors need a bright line as well. If you cannot be bothered to provide a warning before blocking an editor regardless of the situation then you shouldn't be entrusted with the block button, IMO. GabeMc  (talk&#124;contribs)  17:32, 12 March 2014 (UTC)


 * These things happen already, Epipelagic, but it should be left to the admin to take all the factors into account. SlimVirgin (talk) 17:16, 12 March 2014 (UTC)
 * That's the problem, Slim. Some admins issue bad blocks due to their bad judgment, which might be inevitable, but its not inevitable that the victims of a bad blocks have nothing to point to in their defense. This seems like a blank check that non-admins do not get. GabeMc  (talk&#124;contribs)  17:19, 12 March 2014 (UTC)


 * Per Edit warring: "Administrators decide whether to issue a warning or block; these are intended to prevent, deter and encourage change in disruptive behavior, not to punish it." If prevention is the purpose of giving blocks then I fail to see how requiring a warning would undermine the guideline. GabeMc  (talk&#124;contribs)  17:19, 12 March 2014 (UTC)

Goodness. I get the impression from what you are saying SlimVirgin, that everything is fine just as it is. It's all very well to say, "it should be left to the admin to take all the factors into account". That perhaps works with the better admins, but we have a lot of admins who are not up to that standard. That means that things are not fine as they are. You say, "These things happen already". But they only happen sometimes. At other times appropriate things are not happening. --Epipelagic (talk) 17:36, 12 March 2014 (UTC)


 * , if blocks are meant to be corrective and not punitive then how does requiring fair warning undermine their intent? GabeMc  (talk&#124;contribs)  17:47, 12 March 2014 (UTC)

Oppose Edit_warring doesn't require a warning to report at AN3. More importantly, I can easily see the side effect of WP:GAMING the system. I know by saying this I'm guilty of WP:WABBITSEASON as well, but I'm failing seeing why WP:APPEAL and WP:ADMINABUSE can't satifactorily address the occasional dispute.—Bagumba (talk) 19:08, 12 March 2014 (UTC)


 * So far we have two over-eager opposes – both from incumbent admins. Both opposes are premature since we haven't agreed what the proposal is yet. I don't suppose it will make any difference what proposal we offer, since admins will tend to oppose it anyway, unless it increases their already over considerable powers. has opposed every single proposal so far, offering little more than unsubstantiated claims that the status quo is fine as it is. --Epipelagic (talk) 20:52, 12 March 2014 (UTC)
 * Likewise, I suppose it won't make any difference, but I'm subject to the same rules as an editor, and I make tens of thousands more edits than blocks. Whatever makes you sleep better.—Bagumba (talk) 19:12, 13 March 2014 (UTC)
 * Please don't take that comment personally. It's just that that particular edit of yours happened to conform to a wider pattern in the way admins knock back reform attempts. --Epipelagic (talk) 20:02, 13 March 2014 (UTC)
 * WP:ADMINABUSE does not apply to the most common block, which is a 24-hour one. The same with WP:APPEAL, which is pretty irrelevant when you consider that the blocked editor is not guaranteed as audience a) within the block period, and b) with an admin other than the one who issued the block. GabeMc  (talk&#124;contribs)  19:17, 12 March 2014 (UTC)
 * Bagumba, at the top of AN3 it says, "You must notify any user you report." Following that requirement has the practical effect of warning the user that someone is unhappy with their behavior.  WhatamIdoing (talk) 20:38, 12 March 2014 (UTC)
 * Understand your point about the time window. There is always going to be a tradeoff between the additional delay of stopping a disruptive editor where all the warnings in the world aren't going to matter, and the inconvenience to the editors who will stop and made nice when they were made 99% certain—not 90%—that a block was imminent. Your specific incident seems to be the extreme case, where no warning was given to you except for the WP:AN3 disclaimer at the top that read "When reporting a user here, your own behavior will also be scrutinized" before you proceeded to report the other editor.—Bagumba (talk) 19:12, 13 March 2014 (UTC)


 * I'm a little concerned about an unqualified "must". A warn-first rule implies not only that you need to post the warning, but that you need to wait a reasonable number of minutes before blocking, so that the person has a chance to notice the banner and read it.  You don't block people for a "post-warning violation" on an edit that was saved one or two seconds after you posted a warning.  But if you've got high-speed, high-volume reverting going on, you might actually want to block first and warn later.  This probably won't come up more than once or twice a year, but I'd be slightly more comfortable with saying "must normally".  On a related point, we should have a general principle about "warnings" that they expire after some reasonable, humane length of time.  I know I've been "warned" (notified, really) that discretionary sanctions are in place on some articles, but I have no idea which ones any longer.  WhatamIdoing (talk) 20:38, 12 March 2014 (UTC)
 * Why should warnings be based on each admins individual discretion? That's the problem here, they already have too much leeway. If a warning slows the blocking admin enough that the edit-warring stops before a block is needed then our mission is accomplished. Blocks are not supposed to be punitive, they are intended as preventive and corrective. How many blocks would never have been issued if we required a prior warning? GabeMc  (talk&#124;contribs)  21:07, 12 March 2014 (UTC)

Whew, this is dispersing into a lot of complex worthwhile threads, probably enough so to keep anything from happening. :-)  :-).   I'd like to still try to bring the baby step idea at the beginning of this to fruition which is to just add the word "recent". Sincerely,  North8000  (talk) 21:03, 12 March 2014 (UTC)
 * I think adding the word "recent" would be a bad idea because it gives first mover advantage to edit-warriors instead of supporters of the status quo ante as is required for BRD to function effectively. We could safely add the word "recent" if we changed to an NRR rule with even N instead of odd, but then what would we do with 1RR pages?  0RR is too few and 2RR is too many.  I really recommend that editors who want to change anything about the current wording edit for a few weeks on some high traffic ARBPIA pages to see what the stakes are.&mdash; alf laylah wa laylah (talk) 21:14, 12 March 2014 (UTC)
 * As far as I can tell these proposals aim to clarify only the guideline for 3RR. I'm not sure why you think this would automatically apply to all edit warring situations. 1RR and 0RR should have their own language that preserves the intent of those guidelines. GabeMc  (talk&#124;contribs)  21:36, 12 March 2014 (UTC)
 * Because the definition of "revert" in WP:3RR is shared by all of them. If you want to rewrite the language describing 1RR just so you can rewrite this language you're going even deeper down the rabbit hole.  I'm not convinced there's enough of a problem to justify risking the problems that might arise in trying to fix it.&mdash; alf laylah wa laylah (talk) 22:32, 12 March 2014 (UTC)
 * WP:3RR links to a different section of the guideline than do WP:1RR and WP:0RR, so why would the language in one section dictate the implementation of another? Anyway, why would a required warning be a bad idea? I've still not seen that argument made to any satisfaction. What are the biggest downsides to having to warn people before you block them? The number one reason why people stop editing Wikipedia is negative encounters with other editors, and I am quite sure that "other editors' often means admins issuing blocks. Should we protect Wikipedia as a whole, or just the admins? GabeMc  (talk&#124;contribs)  23:04, 12 March 2014 (UTC)
 * From WP:1RR: A "one-revert rule" is often analogous to the three-revert rule as described above, with the words "more than three reverts" replaced by "more than one revert".&mdash; alf laylah wa laylah (talk) 23:12, 12 March 2014 (UTC)
 * Right, but what does that have to do with warnings? Under what circumstances is giving a warning a bad idea? GabeMc  (talk&#124;contribs)  23:14, 12 March 2014 (UTC)
 * Under circumstances where it's obvious to everyone that the edit-warrior knows the rules because of previous actions, because of them having made complaints themselves and getting boomeranged, because they're editing on pages with something like this on the talk page: Template:Arab-Israeli Arbitration Enforcement, because they're editing a page with editnotices like this: Template:Editnotices/Page/Jerusalem, because it would add another set of opportunities for wikilawyering. Maybe there are others; those are just off the top of my head.&mdash; alf laylah wa laylah (talk) 23:23, 12 March 2014 (UTC)
 * I mean that how do they know they are approaching 3RR? If I go and edit that page right now I won't know if my edits count as one or four unless a) I check the page history to see if all my edits were consecutive, and b) if someone warned me that they thought I was approaching a 3RR violation. If admins are going to count changes as reverts then an editor won't always know where they stand. I didn't, in fact I was discussing the changes at talk when I was blocked, which was after the supposed edit warring occurred and ended. I honestly thought I made maybe two reverts, then I stopped. They only indication that I had violated 3RR came when Bbb23 blocked me, which is too late, IMO, and it will drive editors away. GabeMc  (talk&#124;contribs)  23:28, 12 March 2014 (UTC)
 * I can see situations in which warning an editor is pointless (obvious vandal, etc.) but I'm having difficulty imagining a scenario in which it's actually counterproductive, especially when one can issue said warning in under two minutes. DonIago (talk) 04:36, 13 March 2014 (UTC)
 * Warning isn't counterproductive, the requirement to warn is counterproductive. Surely the requirement to warn doesn't mean warn and then block immediately.  It must mean warn and then wait some unspecified amount of time to see if the warnee will self-revert and then block if they don't or something like that.  So how's that going to work in practice?  Are you going to specify how long the warner must wait to see if the warnee self-reverts, what?  What if a warning is issued, a time period is waited, no self-revert, the administrator blocks, the edit-warrior says they were going to self-revert but they were just out walking the dog?  I just don't see how this is either necessary or feasible.  But I've said enough... more than enough... about it by now.&mdash; alf laylah wa laylah (talk) 04:42, 13 March 2014 (UTC)
 * What's the guideline for vandalism-blocking? I imagine the principle's the same, in that editors usually/always must have at least one warning for vandalism before they can be blocked. Have there been discussions regarding an appropriate time frame in such instances? DonIago (talk) 04:52, 13 March 2014 (UTC)
 * Good idea, DonIago. We can adopt the same standard that's used for vandalism warnings.
 * Alf, we aren't asking the person to self-revert. We're asking the person to stop.  The purpose of the warning is to see whether we can achieve the same practical effect as a block (no future edits on that page today by that person) without actually having to issue a block.  You can't self-revert when you're blocked, either.
 * Also, I don't see ay reason why the warning needs to be delivered by the admin. If you post a note that says, "Hey, when I keep removing most of your edits, I'm trying to stop you, I'm not trying to build on what you're doing.  Please take it to the talk page instead of edit warring" should be sufficient. WhatamIdoing (talk) 15:39, 13 March 2014 (UTC)


 * 1) Oppose because under most circumstances, an admin should be warning before blocking (socking and radical vandalism excepted). Placing it into policy here would be saying it is ok to NOT do this in other types of blocks.  Dennis Brown &#124; 2¢ &#124;  WER  00:32, 19 March 2014 (UTC)
 * And if the admin doesn't warn? Note here that the system is currently gamed by editors who, when they see the opportunity to demand a trumped up block of another editor who spends more time discussing than anything else, are very quick to deny the accused an opportunity to self-revert and ONLY THEN starting the complaint process.  These complainants won't get their satisfaction if the accused self-reverts because what these complainants really want is the satisfaction of getting their targets blocked, not having an article read a certain way.  Admins CAN see through this sometimes but they also decline to and having a 3RR definition that is highly technical and doesn't mention "discuss" 'even once invites admins to not bother considering whether the complainant has clean hands or not.  Admins are additionally hardly ever second guessed.  I extend an invitation to review in cases like this and that's simply declined: "...your recent block means that you are not eligible..."  It appears we are not going to get this policy improved until a sufficiently large number of editors have been wrongly blocked by it since there is a great deal of denial going on around here about the extent to which the system is being exploited.--Brian Dell (talk) 18:57, 22 March 2014 (UTC)

Even vs odd
The even- versus odd-numbered thing is mathematically wrong. The number of cycles doesn't affect the first-mover advantage. It works like this: The exchange opens with a score of 0–0. No one has made any reverts.
 * 1) Round 1:
 * 2) * You make a bold change.
 * 3) * In this round, I can safely revert you if the rule is 1RR (odd) or more.
 * 4) * I revert you.
 * 5) * Your score: 0 reverts.  My score:  1 revert.
 * 6) Round 2:
 * 7) * You revert my reversion.
 * 8) * In this round, I can safely evert you if the rule is 2RR (even) or more. I revert you.
 * 9) *Your score: 1 revert.  My score:  2 reverts.
 * 10) Round 3:
 * 11) * You revert my reversion.
 * 12) * In this round, I can safely revert you if the rule is 3RR (odd) or more.  I revert you.
 * 13) *Your score: 2 reverts.  My score:  3 reverts.
 * 14) Round 4:
 * 15) * You revert my reversion.
 * 16) * In this round, I can safely revert you if the rule is 4RR (even) or more.
 * 17) * I revert you.
 * 18) * Your score: 3 reverts.  My score:  4 reverts.

See? Every time, your score is lower than mine. I will be constrained by the rule one "move" before you will. It doesn't matter whether the number of rounds is limited to one or ten. When only two people are involved in the "game", the first mover (aka WP:BEBOLD) wins in every instance.

When more than one person is involved in reverting, then the advantage is reversed: the first mover always loses. This is true even when one of the players is passive, e.g., if your first move is to revert what a third editor added yesterday, rather than to make a bold change of your own. WhatamIdoing (talk) 16:03, 13 March 2014 (UTC)


 * No. Let's stick to 1RR for clarity.  Here's how it plays under the current rules:
 * The article currently says [something you disagree with that's been in for a while]
 * You take it out. Your score: 1 revert.  My score: 0 reverts.
 * I put it back in per BRD. Your score: 1 revert.  My score: 1 revert.
 * Now we have to stop and talk. That's good.


 * If we change to a situation where the first change is not a revert it goes:
 * The article currently says [something you disagree with that's been in for a while]
 * You take it out. Your score: 0 reverts.  My score: 0 reverts.
 * I put it back. Your score: 0 reverts.  My score: 1 revert.
 * You take it out. Now we're both at 1 and the article is not at status quo ante so BRD fails.


 * It's the same for any nRR for odd n under the proposed change.&mdash; alf laylah wa laylah (talk) 16:17, 13 March 2014 (UTC)
 * You're missing the point. The result is identical for any even n as well. See how 2RR works in your example:
 * The article currently says [something you disagree with that's been in for a while].
 * You take it out. Your score: 1 revert.  My score: 0 reverts.
 * I put it back in per BRD. Your score: 1 revert.  My score: 1 revert.
 * You take it out again. Your score: 2 revert.  My score: 1 reverts.
 * I put it back again. Your score: 2 revert.  My score: 2 revert.
 * Now we have to stop and talk. That's good."
 * That's also identical to what happens with 1RR, 3RR, and eleventy-two RR. The number of "rounds" is irrelevant, because each of us get 'n' rounds.  There would only be an even-odd difference if 'n' means your changes plus my changes (total reverts made on the article today, rather than reverts made by me today).  WhatamIdoing (talk) 00:11, 19 March 2014 (UTC)


 * Oh, plus. Asking an editor to self-revert before reporting is standard (although informal) etiquette on 1RR pages.  It's one reason the rule works so well.  The editors give one another the chance to self-revert and erase the violation before reporting at AE and it often calms everyone down.  I was just assuming that that would be the point of a warning before blocking.  Perhaps not.  So it goes.&mdash; alf laylah wa laylah (talk) 16:21, 13 March 2014 (UTC)
 * I'm not sure that anyone wants to "change to a situation where the first change is [never] a revert", its that the first edit isn't always a revert, particularly when the material cannot be attributed to a specific edit or editor. If I add some new material to an article and then three days later you remove it, that's a revert. If I add material to an article and several weeks or months pass before someone copyedits my addition, then they are not reverting me; they are editing. I think that we should develop a chart similar to the one at Reverting that shows the various situations. E.g., if you hit the actual revert button then its always a revert, but if you copyedit material without undoing the substance its not a revert.


 * Take for example a hypothetical situation whereby I am editing an article under a 1RR restriction. My first edit of the day there is to disambiguate a link, an uncontentious change that according to Bbb23 and others might be a revert, but that's not the spirit of the guideline or the principle of collaborative editing. Am I really at my limit at this article when I haven't changed the substance of any material? An article with a 1RR restriction isn't one where no more than one uncontentious change can be made, is it? The purpose of these guidelines is to prevent edit-warring, not to limit the amount of improvement one editor can make to an article in one day.


 * I'll say it again, via reductio ad absurdum, all changes cannot equal reverts because copyediting is encouraged whereas reverting is discouraged. If all changes are indeed reverts, and some changes are uncontentious improvements, then by Bbb23's logic some uncontentious improvements are reverts, which are discouraged. Its simply not logical to say that all changes are also reverts, nor is it in the spirit of how the "community interprets reversion, because it is not consistent with either the principle of collaborative editing or with the editing policy." GabeMc  (talk&#124;contribs)  18:14, 13 March 2014 (UTC)


 * Comment User:Bbb23 by my count is already mentioned over 20 times on this page. I'm not here to defend them, but it seems like an overreaction to create more rules for an action by one individual that may or may not be wrong.  There are better venues than here to demand an apology or to seek corrective action of that individual.  Moreover, it's an oversimplification that this is a pure 3RR issue when the block explanation mentioned other reasons.   I'm all for improving Wikipedia, just don't think this page is the place for this specific issue with a specific admin.  Good luck, however consensus decides to proceed.—Bagumba (talk) 19:57, 13 March 2014 (UTC)
 * RE: "There are better venues than here to demand an apology or to seek corrective action of that individual." Such as? Arbcom won't hear cases for short blocks, so that exempts 75% of all blocks from scrutiny. Bbb23 decided when the conversation was over and to my knowledge there is nothing that can be done once he stops the discussion. Anyway, this is really less about my silly block then it is about a very basic component of any high-quality corrective system, which all include warning requirements. Its absurd to me that we allow admins to block and potentially drive-away editors without asking them to warn first. IME, there are at least three groups of Wikipedians when it comes to an issue like this: the first group, which I call Chicken Littles, resist change almost obsessively as though the entire project were a house of cards that could be brought down by one change to the guidelines. The second group, the Power Mongers, resist any change that could in anyway limit or reduce their ability to act untethered. The third group, the Unreasonable ones, want to experiment with guideline and policy changes in an effort to fine-tune our project, but when the first two groups bond together to prevent change there is little to nothing that can be done. "The reasonable man adapts himself to the world; the unreasonable one persists in trying to adapt the world to himself. Therefore all progress depends on the unreasonable man." - George Bernard Shaw, 1903 GabeMc  (talk&#124;contribs)  20:08, 13 March 2014 (UTC)
 * Ha! Yeah. George Bernhard Shaw's Maxims for Revolutionists is the perfect book from which to take quotes out of context.  What do you make of "If you beat children for pleasure, avow your object frankly, and play the game according to the rules, as a foxhunter does; and you will do comparatively little harm. No foxhunter is such a cad as to pretend that he hunts the fox to teach it not to steal chickens, or that he suffers more acutely than the fox at the death. Remember that even in childbeating there is the sportsman's way and the cad's way."  The points Shaw's making in that book are not as clear as you seem to think they might be.  There's a certain amount of irony present.  And enough with the self-serving division of Wikipedians into three parts.  We are not Gaul, after all...&mdash; alf laylah wa laylah (talk) 20:23, 13 March 2014 (UTC)
 * Right, and this talk page isn't Parliament, but FTR I said "IME, there are at least three groups of Wikipedians when it comes to an issue like this", not that everybody on Wikipedia falls into one of those three groups. Are you asserting that admins do not pathologically oppose any and all reforms that check their powers? GabeMc  (talk&#124;contribs)  20:32, 13 March 2014 (UTC)
 * Actually, that's kind of natural, as admins have an inherent COI factor in such discussions. (but not necessarily a wp:coi) North8000  (talk) 21:19, 13 March 2014 (UTC)
 * I don't know how many types of Wikipedians there are in total, but there's definitely one type who gets busted for violating a core site policy and then spends his next 1,000 edits trying to alter that policy in order to retroactively justify his violation. We're not going to alter a core site policy simply because Gabe is mad that he got blocked. And really, that's pretty much all that's going on here&mdash;I haven't seen much attempt to even pretend otherwise. This is Rule #8 in action. MastCell Talk 21:31, 13 March 2014 (UTC)
 * Which rule is it when an arrogant and overbearing admin tries to make a discussion personal in a feeble attempt to undermine a reform attempt that might check his power? I'm really not mad about my block; its totally irrelevant to me, but it made me realize that, outside of an essay, there is no clear-cut definition of what a revert is. Why isn't the term revert clearly defined in a policy or guideline or both? It seems like a pretty central topic to the project and since its interpretation leads to many blocks I think that its important. GabeMc  (talk&#124;contribs)  21:39, 13 March 2014 (UTC)
 * Why isn't the term revert clearly defined in a policy or guideline or both?
 * Lack of definition = more flexibility. &mdash; goethean 23:34, 13 March 2014 (UTC)
 * Why does anyone need so much flexibility in a 3RR guideline when all they have to do is cite WP:EW instead, which does not require a 3RR, 2RR, or a 1RR violation? It seems like we have several layers of redundant flexibility that only aim to cushion admins from the great embarrassment of making a mistake, but why is the gander eating better food than the goose? GabeMc  (talk&#124;contribs)  23:59, 13 March 2014 (UTC)
 * Because if you cite EW, they can defend themselves by getting other people to agree that their actions are not edit warring. If you cite 3RR, they have no defense.  3RR is a zero tolerance policy (another phrase for "bright line rule")--if it's three reverts, someone is automatically guilty and you don't need to justify the punishment by anything more than counting the reverts.  That's what makes 3RR+loose definition of revert useful to admins who want to be able to block anyone any time. Ken Arromdee (talk) 15:16, 14 March 2014 (UTC)
 * Furthermore, should we make the mistake of "clearly defining" the contents of WP:EW (beyond "do not use your edits to fight"), then we end up in a situation where editors will find loopholes, and wikiLawyering will become more fashionable that it already is.  We want the PowerMongers/Admins to have flexibility, because the alternative (according to ChickenLittles) is that the Unreasonables will have the more room to wiggle.  There is a very good reason the general rule is left almost completely up to admin discretion, and the 3RR is just a line in the sand, not a daily entitlement of wiki-battle-ammo.  p.s.  I'm not familiar with GabeMc, nor with the page in question, but I've thought long and hard about WP:EW, and making it 'firm' would be  a bad idea.  HTH. 74.192.84.101 (talk) 07:10, 17 March 2014 (UTC)
 * If you push a rule so far in one direction that you can prevent one sort of abuse, you often enable another sort of abuse in the process. That's what you're doing here.  By setting it up so that everyone is a rule violator, you avoid the problem of users finding loopholes in the rules, but you exacerbate the problem of admins blocking without good reason. Ken Arromdee (talk) 21:47, 18 March 2014 (UTC)

Please stop this futile wrangling
Please stop this futile wrangling on policy. Policy will never capture every possible nuance. Even if it did, it would be too complex for a workable document. It's not the policy that needs fixing. It's the manner in which the policy is interpreted and implemented. It's the absence of a framework that allows a judicious interpretation of policy. This is where Wikipedia governance fails. You cannot bridge this absence by tinkering with policy.

At present both the interpretation and the implementation of policy is left to individual admins. We have a police force but no judiciary. Individual admins are allowed to decide who are the villains, then act as one man juries who assess the guilt, then act as judges who interpret the policy (law) and sentence the offenders, and finally act as enforcers who carry out the sentence. What avenues of appeal are available are equally deficient. This anarchic mess of loose-cannon admins cannot change until the absurd right that every single admin has to block content builders is withdrawn and replaced by a properly constituted disciplinary board.

Then to make things worse, we have policy but no perspective on the policy. Nothing functions effectively on Wikipedia as a constitution. There is, for example, nothing like a mission statement that sets out what admins and content builders are here for. Many admins seem to think the purpose of being an admin is to block as many content builders as they can. Likewise, many seem to think the purpose of content builders being here is to provide admins with fodder. Further, there is nothing like a bill of rights for users. Without moderating documents like these, the Wikipedia governance is directionless and lacks mechanisms for achieving balance. Content builders are at the mercy of what is little more than a vigilante system. --Epipelagic (talk) 22:19, 13 March 2014 (UTC)
 * I agree, so lets make an effort to correct this dysfunction. GabeMc  (talk&#124;contribs)  22:25, 13 March 2014 (UTC)
 * Well that brings up the third reason we are in such a quandary. The mess we are in cannot be remedied while the admin corp control their own terms of reference. A major function of pages like this one is just to let over-trusting content builders vent until they run out of steam. --Epipelagic (talk) 00:14, 14 March 2014 (UTC)
 * Maybe, but for every active admin there are 100 active non admins that have a stake in what we discuss here. Admins aren't the problem; the problem is with the apathetic non admins. GabeMc  (talk&#124;contribs)  00:26, 14 March 2014 (UTC)
 * It is a complete myth that these matters are decided by "the Wikipedia community". It comes down to who actually monitors the relevant pages. The vast majority of content builders came here to write an encyclopedia. They have no interest in and little awareness of administrative matters, and certainly don't want to get involved in internal political bickering. If they are treated badly, they usually give up and leave rather than attempt to challenge the status quo, like you are doing. The particular page we are using here is a Wikipedia cul-de-sac. Little more than 300 users monitor this page, and most of those are probably admins or are inactive. Policy moves are usually debated on drama boards watchlisted by the usual denizens of drama boards. These include admins and their retinues, the admin wannabees, and the habitual dramaboard peanut gallery. This is the community that largely decide policy. It is a community that seems generally happy with the status quo, and seems to regard all the drama and the anguish it generates as a plus. it is not a crowd that is interested in setting up a fair system for content builders. The admin wannabees are usually keen to impress admins, and often spearhead attacks against proposals that are not congenial to admin privilege. The admins themselves have huge reserves of over one thousand members to draw from. If a proposal admins want defeated looks like gaining acceptance, there are hundreds of admins available in the wings. These proposals rarely have more than 200 participant all told, so admins easily defeat anything that does not enhance or at least retain their powers. It is easy to document the truth of this with the countless failed proposals of the past. This is why I characterised the admin system above as a vigilante system. --Epipelagic (talk) 01:09, 14 March 2014 (UTC)
 * Admins have to follow policies. So let's write the policies better.  Starting with adding one word  "recent" per the first proposal. :-).   North8000  (talk) 00:56, 14 March 2014 (UTC)
 * Yes, it's always good to improve policy. But doing that alone will not fix the problems. Creative admins will drive their tanks though your policy no matter how cleverly you try to weave it. --Epipelagic (talk) 01:13, 14 March 2014 (UTC)
 * I'm with you in spirit on that, given that Wikipedia policies are inevitably fuzzy to some degree, but as a minimum, they matter a lot.  North8000  (talk) 01:25, 14 March 2014 (UTC)
 * Let's not fool ourselves into the belief that improvements to policies solves "problems". Editors solve problems. But that tank you refer to can be driven by any editor and need not be an admin.--Mark Miller (talk) 01:29, 14 March 2014 (UTC)


 * We can only imagine what it would be like if Wikipedia was dedicated to an exemplary admin system, something we could have pride in. The price Wikipedia pays for indulging incumbent admins by continuing with the system we actually have must be very high. It's a shame, but unpleasant and dysfunctional structures crop up everywhere. C'est ça la vi. --Epipelagic (talk) 04:35, 14 March 2014 (UTC)

--"If nothing we do matters, then all that matters is what we do." DonIago (talk) 04:52, 14 March 2014 (UTC)


 * This sense of powerlessness is the price content builders pay because the admin system has been positioned in a place where it is beyond remedy. It is, of course, wrong that that was allowed to happened. Admins never attempt a rational defence of their system, as you see right here (though they attack the messengers who critique the system mercilessly when they think they can get away with it). --Epipelagic (talk) 05:31, 14 March 2014 (UTC)

BRD and 3RR vis-a-vis the status quo
If BRD gives preference to the status quo ante, 3RR does not, because if I add a contentious statement to an article that does not in any way undue the actions of another, the first person to revert me is one behind me in their 3RR quota, and I can revert them once and we would still be tied. This seems like a contradiction. Any thoughts? GabeMc (talk&#124;contribs)  19:04, 16 March 2014 (UTC)
 * WP:AVOIDEDITWAR. WP:GAMING is only going to get so far if it is against WP:CONSENSUS.—Bagumba (talk) 19:16, 16 March 2014 (UTC)
 * Well, that's true, but sometimes only two people are involved in an edit war; what then? GabeMc  (talk&#124;contribs)  19:18, 16 March 2014 (UTC)
 * There's alternatives detailed at AVOIDEDITWAR. The key to keep in mind is there is no deadline.—Bagumba (talk) 19:30, 16 March 2014 (UTC)
 * Right, but if I inserted the text-string: "George Washington was a Mexican immigrant", and for whatever weird reason only you and I watched the page it would remain indefinitely in defiance of the status quo ante. GabeMc  (talk&#124;contribs)  19:36, 16 March 2014 (UTC)
 * Tag it and let consensus run its course.—Bagumba (talk) 19:58, 16 March 2014 (UTC)


 * Then you report the other editor to a noticeboard before anyone hits three reverts and you'll have editors swarming all over it. It works fine, believe me. If the added material is egregious enough it's covered under WP:NOT3RR and you can keep taking it out until the other editor reverts four times and then they'll be blocked. If it's not that egregious, as in your chosen example, which may well be true for all we know under certain interpretations, then what's the big deal if it stays in the article for a little while? In fact, a few falsehoods may do some good, and even print encyclopedias have them, often intentionally included (e.g. Fictitious entry). The nonsense is never going to stay in there indefinitely because you can start an RfC, ask for a 3o, slather it over every wikiproject talk page you can think of, and even bring it to AN/I if you're in the mood for sleazy kicks, and it'll be swarming with editors soonest.&mdash; alf laylah wa laylah (talk) 20:05, 16 March 2014 (UTC)
 * And as a note, if the edit is not about George Washington, but about some WP:BLP, there already is an explicit exception to the edit-warring rule. The system is messy, but it functions.  I agree that the use of the word 'revert' is basically some Wikipedia-specific floating context-dependent-definition, a kind of wikiRevert rather than the usual meaning... but this intentional fuzziness isn't really an unusual state of affairs at WP.  Look at the policy on wikiNotability, for instance; colloquial lowercase notability is *not* what is documented at WP:N.  74.192.84.101 (talk) 07:20, 17 March 2014 (UTC)
 * Good points all, but my original point still stands, in that if BRD gives preference to the status quo ante, 3RR does not, because if someone added a contentious statement to an article others would have to gain consensus to remove it, since the addition of new material does not undo the work of others per se and its not a revert in any sense. In this way the intent of BRD contradicts the intent of 3RR; those wanting to remove a bold edit are at a disadvantage in terms of 3RR, because the first removal counts toward 3RR whereas the bold editor starts with clean slate. GabeMc  (talk&#124;contribs)  17:18, 17 March 2014 (UTC)
 * They don't need consensus to remove it if it's contentious and unsourced. All they need is one more editor that wants it out than wants it in.  If it's sourced, then it doesn't matter whether it stays in or out while the the discussion proceeds.  The discussion is the important thing in that case.&mdash; alf laylah wa laylah (talk) 17:37, 17 March 2014 (UTC)
 * Not everything that can be sourced is also factual. I could add a sourced statement that is not at all true and the status quo ante is not upheld because as you say, "[we] need one more editor that wants it out" or else it stays indefinitely. If editors need to build a consensus to remove all additions, then that's not really a preference for the status quo, right? GabeMc  (talk&#124;contribs)  17:46, 17 March 2014 (UTC)
 * Can you give an example of this, because this is getting a little abstract for me. Nothing stays in indefinitely, by the way.&mdash; alf laylah wa laylah (talk) 18:19, 17 March 2014 (UTC)
 * Its only a hypothetical; I know that contentious additions have little chance of remaining indefinitely, although that is an equally general assumption. As for an example of a sourced statement that is untrue, from the Guardian: "Jimi Hendrix's 'Purple Haze' was the consequence of Stanley's Monterey Purple acid". This is patently false and impossible, since "Purple Haze" was written and recorded several months before the Monterey festival. Nonetheless, you can find several so-called reliable sources for this statement that defies the laws of physics. GabeMc  (talk&#124;contribs)  18:28, 17 March 2014 (UTC)
 * Obviously I know that sourced statements can be untrue. But your example is unfortunate, since it's actually arguably possibly true, rather than being a violation of the laws of physics.  Do you know for sure that Stanley didn't share purple acid with his friends in SF before the festival?  Is there a source that says that "Monterey" purple acid was the only purple acid Stanley produced?  Is it not possible, even plausible, that he might have sold or at least shared purple acid before the festival and it just became known as "Monterey" after because of the impact it had on the festival?  The statement's not as self-evidently false as you seem to think, and to declare the Guardian to be wrong because of some logical argument you cooked up is original research.  In a case like this you might quite reasonably be asked to supply sources declaring that the error was an error. Anyway, if this is the level this conversation is at, it's really hard for me to see that there's anything that needs to be fixed here.  We're just talking about the normal operation of Wikipedia at this point.&mdash; alf laylah wa laylah (talk) 18:49, 17 March 2014 (UTC)
 * You seem to be confused. According to the Guardian, Stanley's acid inspired Hendrix to write "Purple Haze" in December 1966, six months before the Monterey festival took place in June 1967 at which Stanley shared copius amounts of purple acid. How can a song be inspired by events that had not yet taken place? Further, according to the NY Times: "He did not, contrary to popular lore, release a product called Purple Haze; in interviews, he sounded quite miffed that anything emerging from his laboratory could be thought to cause haziness rather than the crystalline clarity for which he personally vouched." So no, its not my OR, Alf, its a known fact that this myth exists and that numerous reliable sources have repeated this logical absurdity. According to John Peel, people started calling the purple acid "Purple Haze" after the song, and not the other way around. GabeMc  (talk&#124;contribs)  18:58, 17 March 2014 (UTC)
 * What am I confused about? Did I say the Guardian was wrong? Did I?  You said that the statement was "patently false and impossible" and that believing it "defies the laws of physics."  That's wrong even if the Guardian is wrong as well.  I don't know and I don't care anything about the factual issues about the song, the color of the acid, what Owsley Stanley said about the reputation of his lab, who cares about any of it?  We're talking about sourced statements that are so absolutely untrue that... what?  It's going to destroy the encyclopedia if they sit in an article for a few hours while people discuss them?  You see what you did when I argued with your "patently false and impossible" sourced statement?  You produced contrary sources.  You didn't just reassert that the statement is "patently false and impossible".  This is how the process works.  Pfft.&mdash; alf laylah wa laylah (talk) 23:36, 17 March 2014 (UTC)
 * Alf, you asked for an example so I gave you one. The Guardian claim defies the laws of physics—specifically causality—because it ascribes a cause to an effect that pre-dates the alleged cause by 6 months. Its kinda like when G.W. Bush said that his favorite childhood book was The Very Hungry Caterpillar, which wasn't written until he was in college. GabeMc  (talk&#124;contribs)  18:20, 18 March 2014 (UTC)
 * Anyway, the point of this sub-thread is that 3RR does not favor the status quo ante as you have repeatedly asserted. If I add an absurd statement that is technically "reliably sourced" then you have to build a consensus to remove it, not the other way around as BRD suggests. Also, since BRD is an essay and 3RR is a guideline, what bearing does BRD have on this discussion? The guideline trumps the essay and the guideline does not favor the status quo ante. GabeMc  (talk&#124;contribs)  18:34, 18 March 2014 (UTC)


 * 1) Round 1:
 * 2) * I add an absurd statement that is technically reliable sourced without changing, removing, or re-adding any pre-existing text, i.e. a non-revert.
 * 3) * You revert me.
 * 4) * Your score: 1 revert.  My score:  0 reverts.
 * 5) Round 2:
 * 6) * I revert your reversion.
 * 7) *Your score: 1 revert.  My score:  1 revert.
 * 8) Round 3:
 * 9) * You revert my reversion of your revert.
 * 10) *Your score: 2 reverts.  My score:  1 revert.
 * 11) Round 4:
 * 12) * I revert your reversion.
 * 13) * Your score: 2 reverts.  My score:  2 reverts.
 * 14) Round 5:
 * 15) * You revert my reversion.
 * 16) * Your score: 3 reverts.  My score:  2 reverts.
 * 17) Round 6:
 * 18) * I revert your reversion, which restores the absurd statement from round 1 and leaves you without another revert to restore the status quo.
 * 19) * Your score: 3 reverts.  My score:  3 reverts.

So, the reverting editor will get to the 3RR limit before the bold editor every single time, assuming that the dispute involves only two parties, which it often does. In this way, 3RR is geared in favor of the bold editor, not the reverting editor as BRD suggests, hence the inherent contradiction; BRD and 3RR favor opposite parties to a two editor dispute. The same is true if four editors are involved if two are on each side; in fact the balance only shifts to the status quo ante when there are more editors in favor of the consensus version than support the new addition. GabeMc (talk&#124;contribs)  18:51, 18 March 2014 (UTC)


 * BRD doesn't actually give a preference to the status quo ante (and neither does WP:Consensus). BRD merely prefers not messing with the mainspace while you sort out disputes.  BRD expects that the end result—the WP:NODEADLINE result—will be changes.  WhatamIdoing (talk) 00:17, 19 March 2014 (UTC)
 * Except that when invoking BRD, most editors are implying that even one more revert will constitute edit-warring, which does favor the status quo in that any bold edit may be reverted pending consensus to restore, whereas 3RR implies that the reverting editors should gain consensus to remove the disputed new addition. GabeMc  (talk&#124;contribs)  00:24, 19 March 2014 (UTC)
 * I agree that we have a remarkable number of people citing that essay who have either never read what they are citing or who have serious reading comprehension problems. But that problem will never be solved, or at least not without many people politely correcting every misuse of the term that they see.  WhatamIdoing (talk) 23:56, 16 April 2014 (UTC)


 * I have never as an admin used 3RR as a reason that I can remember, . It's purely mechanical, and being purely mechanical is incapable of providing the flexibility for dealing with the range of situations. The discussions above for how to exactly distinguish the edits that constitute the revert are what a bot would need, It is true  that the original definition was much too vague for a bot; but it was also much too inflexibility for a human.  I did not come here to act like a incompetently-programmed robot. I think it was based upon the type of thought necessary to play or program games, where the rules need to be as exact as possible  to give structure to the game and prevent a free-for-all, even if they are lacking in fairness. We're not here to play games, but to edit, and the necessary distinguishing feature is what is constructive, not what fits within a pre-formmed pattern of moves.  The only reason an admin would care about an exact definition is to avoid the onus of thinking and taking responsibility. The only reason an editor would care is to arrange and time their actions to try to come as close to the boundary as possible while trying to trap their opponent into an error.  I like games, but games by their nature are a way of detaching oneself from consequences and from the intent to actually accomplish something. They are for recreation--and to provide a structure for patterned training. We're not here to practice strategy, or to se how clever we can be, but to write an encyclopedia.  The first step in eliminating edit warring is to stop making rules about how much edit warring will be permitted.  DGG ( talk ) 04:43, 26 March 2014 (UTC)
 * Hear, hear.--Brian Dell (talk) 18:38, 26 March 2014 (UTC)

Protection
I request protection of the page. Here is why: Currently, the page is unprotected. A person who wanted to do something that would violate 3RR could just edit its section. He could then do the planned edits and nobody would have a reason to warn him, because 3RR would allow it. Thanks, Lightning BOLT! (talk) 11:59, 22 May 2014 (UTC)
 * Page-protection requests should be submitted at WP:RFPP. DonIago (talk) 12:56, 22 May 2014 (UTC)
 * If the edit lacks consensus, then the changes made by it would be inapplicable. isaacl (talk) 13:09, 22 May 2014 (UTC)