Wikipedia talk:Edit warring/Archives/2013/March

Define R

It is not clear to me where 3RR applies. The terminology of the "R" action of 3RR seems to be the so-called "broad definition" involving {revert, restore, override} and {contribute, edit} and {war}.

But terminology aside, let us define the "R" action of "3RR" in terms of a critical process that locates 3RR as following WP:BRD such that it is never the case either where disagreement should set in (because of close-minded skepticism) or where contention surrounds the article's content (due to a controversial change).

For simplicity, assume only two highly respectable editors. The unwanted procession between a bold editor and an open-minded skeptic goes like this:
(0) nothing
(1) a Bold contribution
(A) a mere Revert
(2) a Bold Revert
(B) a Bold Revert
(3) a volatile Revert
(C) a volatile Revert
BRD sets Discussion at 2 or B, after 1 and A, then 3RR sets violations at 3 or C.

Here is the critical thinking: Boldness is encouraged in contributing or Reverting, because Boldness does not necessarily imply disagreement, rather it forcefully inspires one to think harder about a contribution or Reversion at 2 and B. Thus Discussion can delay until boldness-inspired thoughts are done, in other words until the open-mindedness that opportuned is found critically to be close-mindedness. The delay of Discussion sets in until 3 or C. Woo hoo!

The 3RR rule sets in to force Discussion after BRD and after another round of Boldness. If Discussions don't start, volatility starts (in the above logic). So we have the next logical step: if discussions don't start 3RR rule sets in where R is defined in the help page as

any edit (or administrative action) that reverses the actions of other editors, in whole or in part, whether involving the same or different material. It can involve as little as one word. A series of consecutive saved revert edits by one user with no intervening edits by another user counts as one revert.

Now despite the terminology stating action R is {administrator action, editor action, involvement, revert edit(s)} where is that 3rd R in the simplified, enumerated process? Is it at 3, or C? Why? Are the disagreeable and contentious states avoided? — CpiralCpiral 10:22, 18 February 2013 (UTC)

I have absolutely no idea what you are attempting to communicate. Jeepday (talk) 11:20, 18 February 2013 (UTC)

I have two ideas about how to take that comment. If a general warning, since there is no specific interaction attempted, that I'm incomprehensible (I get poohed for that uncomfortably often), and you're just not yet willing to spending a lot of time studying it to prove anything, that others are not going to either, and that I run a risk that my (possible) efforts will never have need or want to be understood. Thanks. If a complement, there is always laughing at myself for compacting so much into so little space. — CpiralCpiral 01:21, 19 February 2013 (UTC)

To answer your question: By the letter of the rule, (3) is the first violation of 3RR. But don't rely on it. You are starting from a wrong premise: Neither of these two editors is "highly respectable", as they are both engaged in an edit war. (As this page explains: "[3RR] is a convenient limit for occasions when an edit war is happening fairly quickly, but it is not a definition of what "edit warring" means [...]"). Preoccupation with the number of edits is a sign of Wikilawyering. We've all seen too many cases in which a somewhat more experienced editor wages an edit war with exactly as many reverts as the 3RR rule "allows", gloating when a less experienced editor is drawn into a 3rd revert. Such behavior is anything but "highly respectable". — Sebastian 06:38, 19 February 2013 (UTC)

Thank you for answering the question positively, such that it allows me to further assert that two 2nd reverts occur at 2 and B, and two 3rd reverts at 3 and C. I made content remarks about terminology, and you answered with content remarks about the character of the audience. I will return an opinion about the audience of 3RR while continuing to work-in more assertions about points 0, 1, 3, and C.

<snip>— CpiralCpiral 22:38, 19 February 2013 (UTC)

Actually, (4) and (D) are the first violation of 3RR I suppose (C) is still a trap, though. — Arthur Rubin (talk) 23:34, 19 February 2013 (UTC)

Ooops, you're right! I guess I have to learn how to count to 4! — Sebastian 04:24, 20 February 2013 (UTC)

After much consideration of Arthur Rubin's 4 and D, I can well say that indeed the third R is a violation that does occur at 4. But I offer you three consolations Sebastian: 1) the count can best be termed as three, as you correctly intuited originally, and not four. 2) I think I understand your earlier remarks about wikilawyering traps, and 3) I bend on the war-like aspect of 3RR. — CpiralCpiral 13:02, 20 February 2013 (UTC)

The trap occurs at 4 because the violation is the trap. — CpiralCpiral 13:02, 20 February 2013 (UTC)

Cpiral, your reply is TLDR, so I'm taking the liberty to only reply to your last sentence. Your definition of "respectable" obviously differs from mine. If you lived in a place where noise of 115dB is permitted between 8AM and 10PM, and your neighbor, with whom you have a disagreement, would deliberately max out that bright-line rule and make noise during exactly that time and at that volume, then I at least would not respect that. Would you? — Sebastian 04:24, 20 February 2013 (UTC)

I'd disrespect that, yes. — CpiralCpiral 13:02, 20 February 2013 (UTC)

I removed the too-long-didn't-read part that elicits no comments. — CpiralCpiral 19:27, 1 March 2013 (UTC)

3RR and Clean-up work

I have over the last 20 hours or so been editing the Hugo Chávez page. Mainly I have tried to get the article into past tense (reflecting his death). But I also did a little bit of clean-up when I went though part of the article, in the form of removing stuff that I found to be either outdated or superfluous. The article is overtly long and has now been tagged as such (by someone else than me). This, however, made me start thinking about whether I am technically in a 3RR violation, as a result of having removed stuff in more than 3 edits within the 24 hours timeframe. Is there any policy on how to apply 3RR when you are cleaning up an article? There hasn't been any edit war going on, as I am aware of. With regards, Iselilja (talk) 20:00, 6 March 2013 (UTC)

You probably won't get in any trouble for routine updates that (we assume) are not controversial. One of your editorial changes was reverted by someone else, so you should not restore that one without checking for consensus. EdJohnston (talk) 20:15, 6 March 2013 (UTC)

Editwarring vs "machine gun editing" vs "cluster-bombing"

Its easy to spot 3RR when two editors go back and forth. But what about a single series by a single editor? For example,

Machine-Gun editing When "Editor X" is new to an article and arrives with a truckload of individual edits that through some combination of outright reverting, deleting, modifying, and/or substitution makes (for a random number example) twenty sequential changes to existing text, does that violate 3RR when

• A The existing text was static for six months?

• B The existing text is fairly dynamic with several active editors?

Cluster-bomb editing

• C Do the answers to A or B change when "Editor X" packs all the changes into a single edit?

Lines of distinction

• What if its only 5 changes? 10? 100? Does number matter, or is any number >3 all that matters? Do we count changes to old text the same as changes to recent textt?

Thanks for comments If you care, there is a matter that inspires this question and I left a note for the other editor, but I am not saying what article it is right here because I seek generic education about 3RR and not some dysfunctional backdoor way to do battle over this particular article. NewsAndEventsGuy (talk) 14:36, 16 March 2013 (UTC)

In my opinion, it is established that any sequence of edits by X without intervening edits is considered at most 1 revert. A single edit may be considered edit warring under some circumstances, but it's not a 3RR violation.

I proposed (above) a further restriction that a series of edits by X, even with intervening edits, should be considered at most one revert if none of the intervening edits was reverted, but it didn't get consensus. (This was, believe it or not, to protect a now-indefinitely-banned editor who I was in conflict with.) — Arthur Rubin (talk) 06:58, 18 March 2013 (UTC)

The existing rule is easy to understand: any number of consecutive edits by X count as at most one revert for 3RR purposes. It is hard to see a good reason to change this. The more complex the rules are, the more likely it that someone who is blocked for 3RR will complain that they were not treated fairly. EdJohnston (talk) 15:34, 18 March 2013 (UTC)

Facepalm Somehow I overlooked the sentence about consecutive edits, thanks. NewsAndEventsGuy (talk) 15:57, 18 March 2013 (UTC)

1RR

The present version of the policy includes:

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".

I believe that to be incorrect; 1RR restrictions usually have a stricter definition of revert (e.g., removing text is not automatically considered a revert) or require that the reversions be of the same section or of the same text. Policy pages using the word "often" should be descriptive, rather than perscriptive. 0RR is extremely rare, except what I would call 0mRR; you may not revert reversion of your own edits.

Of course, I haven't checked all the restrictions applied under ArbComm rulings or by community consensus at WP:AN, but I don't recall editing an article which had a "standard" 1RR in some time. — Arthur Rubin (talk) 07:12, 18 March 2013 (UTC)

I've never, so far, encountered another variant then the one described here, except for the one in force at Tea Party Movement (which is the cause of Arthur's request). Indeed, usually, 1RR is not spelled out in detail, but understood to refer to WP:1RR, which has moved around, but at least since July 2010 has pointed to Wikipedia:Edit_warring#Other_revert_rules, which had, essentially,the current form. I think it's very much preferably to keep this consistency - variants only lead to confusion. --Stephan Schulz (talk) 07:43, 18 March 2013 (UTC)

Well, I don't recall any instance where a simple deletion which was not an actual (partial) reversal of a recent edit has been determined to be a "revert" for the purpose of 1RR, although I agree the "same text" restriction appears unusual. As we all know, a simple deletion is considered a revert for the purpose of 3RR, regardless of whether it's a revert of a specific edit. — Arthur Rubin (talk) 15:20, 18 March 2013 (UTC)

If there is anything special about revert-counting for general 1RR purposes I don't know where that was discussed. On some particular articles or subject areas a variation may be specifically imposed. There is a modified 1RR in ARBPIA and WP:TROUBLES that was approved by Arbcom which allows reverting IP edits under some conditions. There is a different modification of 1RR in WP:ARBAA2 that was enacted at WP:AE and only applies to the Nagorno-Karabakh article. EdJohnston (talk) 15:45, 18 March 2013 (UTC)

Withdraw. If no one else recalls other definitions of "revert" for 1RR, I'm probably wrong. — Arthur Rubin (talk) 00:20, 19 March 2013 (UTC)

Reversion of old text not being counted as a revert

Like others in the past, I'd like to see this clarified. See my recent comments here[1] where what I see as a pov reversion wasn't counted as a revert as it was old text - I've linked to a discussion last year about this. Dougweller (talk) 10:28, 22 March 2013 (UTC)

I'm interested in this too. Doug, I tried to follow your link but could not find what you were pointing at. NewsAndEventsGuy (talk) 10:46, 22 March 2013 (UTC)

My reading of the edit warring guidelines is that we have restrictions and sanctions in place in order to allow edit wars to be prevented from happening in the first place, or from continuing once they've started. Our policy needs to reflect that. 3RR includes reverts of edits made recently by editors because it is often a recent changes to an article that sparks an edit war. It is difficult with reverts to older content, because it is less obvious whether or not these are contributing to the continuation of the edit war. I would be inclined to generally not count reversion of old text as a revert because, if it is old text, it is probably not part of a current edit war. If the old text is added again, the addition would be a revert, as would subsequent removals. The only exception would be to allow admins discretion in dealing with tendentious editors who remove slightly different parts of text with each revert to avoid their edits being labelled as reverts. Sanctioning someone for removing old content (with a reason, of course) is not preventative because an edit war only begins when current editors are editing back and forth (the editor who reinstates the removed material begins the edit war because they're reversing the recent action of another editor). Boldly removing long-standing content which you don't think should be there needs to be encouraged, and we should encourage editors who disagree with such removals to discuss the issue. ItsZippy ^{(talk • contributions)} 10:58, 22 March 2013 (UTC)

I'm inclined to count old ones because (A) the continuous-edit rule is easy to grasp so a string of changes to old text is still just one, (B) by doing count/donotcount we inject confusion rather than clarity, and (C) by not counting them we encourage slow-motion edit wars. NewsAndEventsGuy (talk) 11:09, 22 March 2013 (UTC)

A better link is [2] (the editor's been blocked now by ItsZippy for continuing to edit war). In this specific case I see an editor pushing a pov, and that pov included removing some 3 year old text (you could argue the removal was correct, but that doesn't count for 3RR). If I saw a report like this where the removal of old text was something I thought was related to the editor's current reverts, I'd probably block. Dougweller (talk) 13:05, 22 March 2013 (UTC)

(edit conflict) I think something like this is hard to write policy for: the nature of edit wars means that anything we agree should be in policy will be circumvented by clever edit warriors. I think admin discretion is necessary here: if the very old edit is unrelated to the conflict, I'd be inclined to let it slide; if it's being used to game the system and avoid 3RR on a technicality, I'd want to take action. But we do that anyway: I've blocked users for edit warring who have not technically violated 3RR; I get the feeling that any policy we write regarding old edits will be subject to just as much discretionary interpretation that any policy that is not overly-restrictive will be so vague that it doesn't give any useful guidance. ItsZippy ^{(talk • contributions)} 13:06, 22 March 2013 (UTC)

Works for me. NewsAndEventsGuy (talk) 15:40, 22 March 2013 (UTC)

If the edit which reverted old text can be seen as part of an ongoing POV campaign, I think it might be counted as a revert. If it's truly a new idea for article improvement, maybe not. EdJohnston (talk) 16:48, 22 March 2013 (UTC)

I agree. Can we figure out some form of wording that makes it clear that this is up to Admin's discretion? I still think it would be useful given the fact that I'm not the only editor who has asked about this. Dougweller (talk) 17:25, 22 March 2013 (UTC)

"The following is not open for debate - this is a matter of discretion" Pardon the levity. Although silly I agree it would be useful NewsAndEventsGuy (talk) 17:43, 22 March 2013 (UTC)

At present WP:EW contains no rule on whether to count removal of long-standing text as part of any given series of reverts. Any rule that actually does give guidance is probably not helpful. So I suggest leaving the policy alone for now. I.e. no change. A somewhat-related topic came up on this page in September 2012. We decided at that time that it was usually possible for admins to distinguish normal 'improvement' edits from reverts but this was not susceptible to a rule. This question does come up periodically: people ask if a routine series of improvements will accidentally break 3RR. In cases like that (where the intent is obvious) it is no problem. EdJohnston (talk) 19:07, 22 March 2013 (UTC)

I agree with Ed. This issue is going to vary on a case-by-case basis. Depending on the report, it may not be clear to the reviewing admin that there is an old edit that is being reverted. Recently, I handled a report in which the reporter made it easier for me by highlighting it because, otherwise, it would have looked like an addition to me. Even then, one has to analyze the whole picture and decide what's appropriate. To explain all this in the policy would be difficult, if not impossible.--Bbb23 (talk) 19:18, 22 March 2013 (UTC)

No argument, but what about text saying "We rely a lot on admin discretion." Then everyone knows and this very thread won't need to be replowed by the next person asking the same things? NewsAndEventsGuy (talk) 10:01, 23 March 2013 (UTC)

(ec) It was said in the last discussion I saw that 3RR is supposed to be a "bright line", such that when someone steps over it, they know and admins know and everyone knows they've stepped over it. Any deliberate fudging of the rule, particularly on something so fundamental as what counts as a revert (given that the rule is all about counting reverts), would seem to negate the very reason for its existence. Either ditch the rule and trust admins to know edit warring when they see it, or (if we must have such a rule) make it clear what it really means, by not leaving questions like this unanswered. Victor Yus (talk) 10:06, 23 March 2013 (UTC)

It's a good argument that if 3RR is a "bright line" then the age of the text doesn't matter. Right now I'd argue that the line is not so bright if age matters. And that edit warriors reverts even if of old text are for a reason and intentional, so why ignore them? Dougweller (talk) 11:01, 23 March 2013 (UTC)

Anyone who patrols ANEW knows that a violation of 3RR doesn't always result in a block. I would hate to lose the discretion I have in handling reports, with respect to 3RR, with respect to what counts as a revert, with respect to "old text" (doesn't come up as often), and with respect to other variables too numerous to list. And just because I close a report differently from another admin doesn't mean that either of us is "wrong" or that it's a bad thing. Sanction variation in other contexts is common at Wikipedia. We are not fungible, and that's good.--Bbb23 (talk) 14:57, 23 March 2013 (UTC)

I basically agree with Bbb23 and EdJohnston that a policy-mandated interpretation on this issue is not desirable and that admin discretion is best. Heimstern Läufer (talk) 15:10, 23 March 2013 (UTC)

Which is current practice and I concur. I still think that we could add something even if only "the age of the material reverted may be taken into account if not the reversion is not relevant to the other reverts." Dougweller (talk) 15:13, 23 March 2013 (UTC)

I'm glad to hear that breaches of this "rule" don't always result in a block, and I would hope that admins would continue to use their discretion over this and many other factors. However we ought to make an effort to avoid making people think they haven't broken the rule, when (at least according to the judgement of some admins) it may turn out that they have. Hence: if it's likely that people are going to think that changes to old material don't count as reverts for the purposes of this rule (and I think intuitively that's what I would have thought if I didn't know better), while in fact such changes are sometimes counted as reverts, then we ought to point this out explicitly. Something like: "Note that your actions may count as reverts even if they do not undo a recent action by another editor; however, the age ... [continue as in Dougw.'s suggestion above]." Victor Yus (talk) 15:33, 23 March 2013 (UTC)

It's also occurred to me while reading Victor's latest post that the archives of 3RRN will now contain a statement that a revert didn't count because it was too old, and that might be seen by some(including the editor I brought there for 3RR) as policy/practice, something I'm sure wasn't intended by ItsZippy. Dougweller (talk) 16:44, 23 March 2013 (UTC)