Wikipedia talk:Edit warring/Archives/2013/September

A flaw in definining a "revert"?
Can I clarify something here, it is something that has always bugged me about how 3RR operates. A "revert" for the purposes of determining whether 3RR has been violated is defined as a series of consecutive saved revert edits by one user with no intervening edits by another user counts as one revert.

OK. The problem I have with this, is that I can effectively remove the same content in three different ways, and one of them would see me violate 3RR. Let me provide a simple example to illustrate my point: an editor adds four unsourced claims to List of vegans in four sequential edits which I decide to remove.


 * 1) I remove all four names in a single edit. That counts as one revert.
 * 2) I revert each individual edit, resulting in four consecutive edits. By the definition provided above that counts as one revert.
 * 3) I revert each individual edit, but another editor also editing the article intermingles their edits with mine, meaning that my edits are no longer consecutive. By the definition of the rule I have made four reverts and violated 3RR.

Basically I have carried out exactly the same action, but one method is punishable by a block. I am pretty sure this is not how the rule is intended to work, but that is how it reads in a literal sense. It just seems very arbitrary to me. I think 3RR should only apply to the same material i.e. it is a revert cycle you are trying to discourage, not necessarily a revert series. Betty Logan (talk) 06:56, 3 September 2013 (UTC)
 * Its designed to capture the situation where you remove name 1, then another re-adds name one, then you remove name 2, and another editor re-adds name 2, you remove name 3, and another editor re-adds name 3... which is just as bad as edit warring over one point in the article. I think this is just a place we need to count on Admins to exercise good sense in applying the rule. Monty  845  14:47, 3 September 2013 (UTC)
 * Right, so the rule is really designed to actually tackle repeats of a previous action? That makes much more sense to me. So what we are really counting is "redos" i.e. you redo an edit that either you yourself or someone else has done previously? Betty Logan (talk) 17:43, 3 September 2013 (UTC)
 * That is how it is supposed to work. If you want to remove four names, then you should do it in one edit.  Some editors will make one change, find it reversed, and continue making changes against consensus.  TFD (talk) 05:23, 4 September 2013 (UTC)
 * Ideally, I can see how removing all four names in a single edit would keep it simple, and my example was slightly pedantic just to illustrate my point; but some articles get a lot of activity where sometimes an editor can find themselves undoing unconnected edits throughout a day i.e. removing unsourced claims, spam links, reversing erroneous changes to sourced content etc. Mainly I was just interested in how 3RR was interpreted in that particular context. Betty Logan (talk) 05:38, 4 September 2013 (UTC)
 * Again, I would hope that an admin would count them as essentially consecutive if the intervening edits are entirely unrelated. We have the RFA process because admins are required to excercise good judgement and discretion, and this is one of those cases. To try and explain every possible scenario in the policy just isn't practical. Monty  845  14:28, 4 September 2013 (UTC)
 * I like the way Betty Logan interprets 3RR and I agree with their views, but don't forget the main issue (at least as I see it) is "within a 24-hour period". Yes, there is bound to be special circumstances to an unique long term issue but, most 3RR's should stay within the suggested time interval.  Mlpearc  ( powwow ) 18:58, 5 September 2013 (UTC).


 * I think the most obvious answer to all this is to use common sense. One of the five pillars of Wikipedia is that we don't have firm rules, and trying to create firm rules and water-tight definitions simply doesn't work, because the more complex definitions became as we try to legislate for every situation, so the more places there are for some situation or other to fall foul of some problem in the definitions. Over the years, as policies and guidelines have expanded with attempts to plug every loophole and define every doubtful case, more and more room has been created for unproductive wikilawyering about the exact correct interpretation, wasting time that could have been more productively spent. In a situation where editor A reverts several different things within a couple of minutes, while in between those reverts editor B makes some unrelated edit, anyone with a grain of common sense will treat editor A's edits in the same way as if editor B had not happened along, while if editor A's later reverts are related to editor B's intervening edits then that is a different matter. We don't need a rule to tell us that.


 * I said that I thought that was "the most obvious answer", but there is, in my opinion, an even better answer. What counts as one "revert" absolutely doesn't matter at all unless for some reason you think it matters exactly how many reverts an editor has made. The only reason for thinking that is the so-called "three revert rule". However, disruptive edit warring is disruptive edit warring, and blockable, whether it breaks the "three revert rule" or not, while non-disruptive editing is not disruptive, and not blockable, whether it technically breaks the "three revert rule" or not, so in neither case does it matter whether the "three revert rule" has been broken. This is just one of many examples of ways that the "three revert rule" is totally unhelpful. I wouldn't like to guess how much editor- and admin- time has been wasted over the years by pointless arguments that start because an editor blocked for edit warring protests angrily "but I didn't break the 3-revert rule", is told "you weren't blocked for that, you were blocked for edit warring", and it carries on from there. In cases where the editor was blocked for breaking the three revert rule, it is even worse. In those cases, I have known absurdly long time-wasting arguments about whether a particular sequence of edits did or did not technically constitute a breach of the three revert rule, when it has been abundantly clear that there had been edit warring, whether or not the three revert rule has been broken. I have even known cases where an editor has in perfectly good faith deliberately waited till 24 hours has just passed to make a fourth revert, genuinely thinking that doing so is acceptable, and has then been totally taken aback to find that a block comes anyway. The "three revert rule" was introduced in good faith, but it it serves little if any useful purpose, and causes endless unnecessary problems. If we simply forgot all about it then we wouldn't have to spend time discussing such technicalities as the one which is the subject of this section. JamesBWatson (talk) 11:41, 10 September 2013 (UTC)

not constructive, unconstructive
I like "not constructive" better because it's more emphatic. NE Ent 01:43, 12 September 2013 (UTC)
 * I don't agree. Firstly, I don't see that it is "more emphatic": on the contrary, merely saying that it isn't constructive seems to me weaker than stating that it is positively unconstructive, just as saying that something isn't cold is weaker than saying that it is hot. Secondly, the single word "unconstructive" seems less cumbersome to me than the 2-word expression, but of course that is only a matter of personal judgement. Also, the word "unconstructive" has a long history of use in Wikipedia in numerous places, and is clearly accepted by a large consensus. JamesBWatson (talk) 10:33, 13 September 2013 (UTC)
 * unconstructive defines it as "not constructive." In the spirit of saying it is hot, how about "destructive"? NE Ent 11:49, 21 September 2013 (UTC)

Some exempt?
It has become my general impression that certain highly appreciated editors are exempt from having to comply with the 3RR rule. Anyone else have the same impression? If not, I will look forward to having to apologize. --SergeWoodzing (talk) 01:22, 21 September 2013 (UTC)
 * I wouldn't say that anyone is exempt. (Except maybe Jimbo, but I bet there are admins who would love an excuse to block him for however short the block stand) However, the 3rr rule does not mandate a block for violating it, instead, its a line beyond which no one is going to second guess an admin's decision to issue a block. Remember, the goal of the Edit Warring policy is not to block anyone, the goal is to stop edit warring, and blocks are sometimes s necessary means to that end. Personally, I don't like to block anyone for edit warring unless they have a recent, unambiguous, warning against edit warring, even if such isn't required by policy. Generally, more experience editors understand that failing to heed a 3rr warning is one of the best ways to get yourself blocked, and stop once reminded, they stop. That may mean they "get away with it" but it also means policy served its purpose. Monty  845  03:44, 21 September 2013 (UTC)

Hypothetical question

 * A reverts X’s edit
 * B reverts A’s revert.
 * A reverts X’s edit
 * B reverts A’s revert.
 * A reverts X’s edit

Is A in violation of 3rr if the original edit happened a while ago? Thanks in advance, XOttawahitech (talk) 23:08, 27 September 2013 (UTC)
 * This tech question seems to miss the point, which is "Is this an EW" instead of the technical Q "Did this break 3RR".  3RR is just one tool for determining the existence of an EW.  In my view, if they're not talking, seems like A EW'd at line 3 and B at line 4, even though neither broke 3RR at those points.NewsAndEventsGuy (talk) 23:41, 27 September 2013 (UTC)
 * thanks for taking the time to address my concern. Just wondering if "talking" includes short comments made through edit summaries, and if so what would be the next step to resolving the issue without breaking 3rr rules. XOttawahitech (talk) 14:28, 28 September 2013 (UTC)
 * Talking about 3RR, any next revert by A within 24 hours of the first revert will be a violation of 3RR, but as NewsAndEventsGuy said both of them are already edit warring. But my views differ slightly as I think even if they are talking besides reverting, they are still edit warring. And edit summaries should not be used for discussions. Best solution would be (for both A and B) to start a discussion at talk page. If that doesn't help, opt for any of the Dispute Resolution Methods. -- S M S   Talk 14:53, 28 September 2013 (UTC)
 * Thanks for taking the time to join this discussion. I am confused by your comment above where you say that the next revert by A would be considered 3rr, but what about the next revert by B? XOttawahitech (talk) 19:31, 28 September 2013 (UTC)
 * Your example needs another set of reverts to generate the hypo I think your actually interested in. A 3rr violation occurs on the 4th revert. Even if we count A's initial edit as a revert, A is only at 3. If you add the next set of reverts, then it becomes dispositive whether A's first edit counts as a regular edit, or a revert. My position is that at some point, the material from X that is being contested stops being an edit to revert and starts just being part of the article the removal of which is just a bold edit, and not a revert for counting purposes. Exactly how long that takes is a case by case determination, at least months, probably years, and honestly, if your in a situation where your worrying about the exact number of reverts that counts for a 3rr violation, your in an edit war either way. Monty  845  23:09, 28 September 2013 (UTC)
 * @Ottawahitech Among the five reverts you mentioned, A has reverted 3 times and B reverted 2 times, so B is still 2 reverts away from 3RR. Now that Monty845 has raised the time aspect while considering A's first edit as revert or not, I would say that A's first edit is a revert if there is any indication that A was aware of X's edit even if X did that edit months ago. And that could be when A either used the "Undo" function or A was previously involved in an edit war with X involving the same content that A is reverting currently or any other evidence that suggests A having knowledge of X's edit. This is strictly my understanding of this policy and I can be wrong. On a side note, (in case if you are interested in finding out when A or B will get a block) 3RR is just an extension of the Edit warring policy, that clearly draws a line that shouldn't be crossed but administrators have a certain leeway for determining whether someone has violated the Edit warring policy even if they haven't crossed 3RR. And I have seen numerous times editors getting blocked for less than 4 reverts. So its better to stop short of even 3 reverts and move to the talk. -- S M S  Talk 09:51, 29 September 2013 (UTC)

I think the pithy summary by says it all, and it's so good it might be worth including at WP:3RR. He said,
 * "[I]f you are in a situation where you are worrying about the exact number of reverts that counts for a 3rr violation, your in an edit war either way"

NewsAndEventsGuy (talk) 11:31, 29 September 2013 (UTC)
 * I wont comment on the wisdom of adding it, but if you do, make sure to correct my bad grammar, your->you're. Monty  845  18:49, 29 September 2013 (UTC)
 * It may be pithy when used in a given case, but I wouldn't add it to the policy.--Bbb23 (talk) 18:53, 29 September 2013 (UTC)