Wikipedia talk:Edit warring/Archives/2019/November

Clarification on 3RR

Something came up elsewhere that has me confused. I have, for 12 years, thought it worked one way but it sounds like I may be wrong.

Please consider the following sequence, which takes places in a 24-hour period at a particular article, with editor A and editor B. The article is not subject to any special restrictions, and none of the 3RRNO exemptions apply.

1. A removes a block of long-standing content in good faith.
2. B reverts A, also in good faith.
3. A removes the content again.
4. B reverts A.
5. A removes the content again.
6. B reverts A.
7. A removes the content again.
8. B reverts A.
9. A removes the content again.
10. B reverts A.

After which step has editor A violated 3RR?

After which step has editor B violated 3RR?

For the sake of argument, let's say that nobody has used the talk page, but I'd be curious if that would affect people's answers. — Rhododendrites ^talk \\ 19:45, 2 November 2019 (UTC)

7.- MrX 🖋 19:47, 2 November 2019 (UTC)

@MrX: You are saying that editor A has violated 3RR after step 7 (removing the content for the fourth time)? And editor B, presumably, violates 3RR after step 8? — Rhododendrites ^talk \\ 19:50, 2 November 2019 (UTC)

@Rhododendrites: Yes, that is my understanding.- MrX 🖋 19:53, 2 November 2019 (UTC)

Ok. Me too. That doesn't seem like everyone's interpretation, however. I'll wait for some additional opinions here. — Rhododendrites ^talk \\ 20:36, 2 November 2019 (UTC)

I observe editors and Admins recently interpreting the first removal as not-a-revert. This is inconsistent with every definition and description of a revert on our policy and guidelines pages, but I think this interpretation is frequently cited.

So I see many editors saying A violates at 9 and B violates at 8.

I also have observed various interpretations (inconsistent with the language of our documentation) of what is a revert, e.g. full or partial, same or different content, etc. I think this confusion is fairly widespread, even among Admins. It's a pretty important fundamental question, and it should be resolved and documented even more clearly than is currently the case. SPECIFICO talk 02:51, 3 November 2019 (UTC)

I'm no less confused than anyone but I believe the reason some say A violates at 9 and B violates at 8 is pragmatic: it's too hard for an onlooker to determine whether A's first edit is the continuation of an edit war, or whether it is a good-faith attempt to improve the article (hey, that has to be true sooner or later!). Of course the picture is different if there are diffs from a week or a month ago that show A should have known that step 1 was a continuation of an edit war. Participants might understand the POV that step 1 is pushing, and how that is part of A's ongoing contribution to the article. However, it's too hard for an admin to justify a sanction unless step 1 can be shown to be not the first edit for that disagreement. Johnuniq (talk) 06:14, 3 November 2019 (UTC)

Revert aspects have been debated extensively at this talk page. One thing I've maintained is that an editor coming to an article and changing text that was there for days, weeks, or years shouldn't count as a revert with regard to an edit war if that editor was not in any dispute with another editor over that content. If they are coming in to aid another editor, then I can see how WP:Tag team applies. And a tag team effort can be an edit war. A slow edit war can be an edit war. But if we defined a revert as any change to an article, just about every edit we make to Wikipedia could be counted as reverts since Wikipedia articles are continuously being changed. But we don't consider every edit a revert. Flyer22 Reborn (talk) 07:57, 3 November 2019 (UTC)

Agreed, and I'm reminded that I have seen admins who would sanction A at step 7 saying that "obviously" step 1 was a revert of someone's text. Johnuniq (talk) 09:22, 3 November 2019 (UTC)

Here's the source of my confusion: a 3RR first violated at step 8 rather than step 7 means the burden is on the reverter to argue for the stable version, while the initial editor can force the contested change a fourth time without violating 3RR. That seems contrary to the spirit of how, well, much of Wikipedia works. I've long thought of 3RR as being compatible with BRD, but it sounds more an asterisk to BRD. Yes, yes, there are dispute resolution processes and neither party should be edit warring and all that, but why would our most toothy policy err on the side of the contested change?

Best I can think of is that because 3RR isn't written for the same material being added/removed multiple times, but rather for reverts on a page, building in a presumption that one of our other policies will intervene in the case of the same material ... but it's not clear.

Beyond that, the AN3 archives are full of blocks of [mostly new] users at step 7. Maybe the blocking admins would say it was for edit warring in general, and not a 3RR violation necessarily? — Rhododendrites ^talk \\ 14:58, 3 November 2019 (UTC)

Johnuniq, in my opinion, blocks after step 7 directly follow from the definition of the revert. According to 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. "

Obviously, if we interpret this clause literally, any removal (full or partial) and even a modification of 'any existing text should be considered as a revert. Meanwhile, a 3RR is supposed to be simple and clear, to avoid possible misinterpretations or gaming. Therefore, it would be correct to amend this sentence as follows:

"An edit or a series of consecutive edits that undoes recent actions of other editors—whether in whole or in part—counts as a revert. An action is considered recent if it was made within a 24 hr period."

That would eliminate this ambiguity, so good faith users will not be at risk to be blocked at step 7.

Anothert important reason for doing that change is the following. Wikipedia articles have a tendency to an uncontrolled growth, and removal of redundant content is gradually becoming more important than adding new texts. I think the policy should be more friendly to the users who are cleaning Wikipedia from redundant stuff.--Paul Siebert (talk) 15:59, 3 November 2019 (UTC)

• I asked a couple of admins about such situation long time ago, and the answer seemed to be the following. It depends. Edit #1 may or may not be a revert. If one can clearly show that edit #1 was a revert of certain edit by another contributor before ([diff]), then it will be regarded as revert. However, if this is just a removal of sourced text, this usually will not be regarded as a revert. The bottom line: not every removal of sourced text is a revert. Regardless, the general advice is not to be engaged in edit war. If one of the contributors reported such case as above to 3RRNB, then both contributors would likely be blocked. My very best wishes (talk) 18:20, 3 November 2019 (UTC)

I would say that it's more important for a rule to be clear and unambiguous than it is for the rule to be perfect. I think the whole question should be raised at a community-wide venue and that a universally understood rule be written there. Then any improvements on that wording should meet the simple test that they are likewise clear and unambiguous. We would then have uniform expectations and enforcement. Such an approach would be easier than trying to parse all the disagreements, exceptions, and unintended outcomes in the current language. SPECIFICO talk 21:31, 3 November 2019 (UTC)

• This will always be subjective, because defining what constitutes a "partial revert" is subjective. The only way to achieve some clarity here is to use "exact revert" everywhere instead of simply "revert". What it means? If a contributor X inserted new text "ABCDE", then the exact revert would be a removal by contributor Y very same text "ABCDE" or a larger text that includes "ABCDE". A removal of "ABCD" would NOT be an "exact revert" and would not count. Hence, there would be no "partial reverts" in the policy. That makes sense because the purpose of the 3RR rule is to prevent fruitless "circular" edit warring that does not improve content. Having a "partial revert" means the contributors are actually moving somewhere. My very best wishes (talk) 16:04, 4 November 2019 (UTC)

BRD represents the limit of acceptably bold editing

Proposed edit

Insert the underlined

Wikipedia encourages editors to be bold, but while a potentially controversial change may be made to find out whether it is opposed, another editor may revert it. This may be the beginning of a bold, revert, discuss (BRD) cycle, which represents the limit of acceptably bold editing. An edit war only arises if the situation develops into a series of back-and-forth reverts. Nevertheless, not every revert or controversial edit is regarded as edit warring:

Why?

As has been repeatedly discussed, for years, at WT:BRD, inexperienced editors misunderstand BRD as a recommended way to edit, when it is actually an advanced, high speed, editing method that comes very close to edit warring. Practicing BRD, plus a few points of over-reach, becomes edit warring. Someone has suggested that BRD is too prominently linked from policy pages, like this one, WP:EW, and yes, it is, and without sufficient warning that BRD is a dangerous editing style for the inexperienced. —SmokeyJoe (talk) 12:57, 27 November 2019 (UTC)

Oppose. Your suggestion flies in the face of long-established principles like 3RR. Your proposal suggests (though it is too cryptic to be sure) that someone who reinserts something that has been reverted is now edit-warring, when they haven't even broken 1RR let alone 3RR. In fact, if BRD is understood correctly, which is "if your edit is reverted, your next step is discussion", it is a good way to reduce warring. I've seen it work countless times. Zero^talk 04:29, 29 November 2019 (UTC)

You’re pointing that 2RR is not edit warring, but is across the line from BRD. OK, I see your point. —SmokeyJoe (talk) 13:15, 29 November 2019 (UTC)