Thoughts on the edit warring policy

I think the "ABA" structure is fine, but I agree with sending a notifier (especially when it's a third party who notices the issue, possibly months later) rather than immediately suspending someone. If a particular edit war becomes lengthy or contentious, it can be brought to the mods' attention. Or, if someone racks up many notifiers, perhaps signaling careless editing behavior, like not checking the edit history before making changes—that would be worth a chat with a mod at that point.

As an aside, I keep up with reading Edit Banned and I've noticed that recently, people brought in for edit warring are often not given the link to that page, even though they're supposed to explain the process they ought to go through instead of reverting an edit. Granted, a percentage of people given the link clearly don't read or understand it anyway, based on their answers, but not giving them the link at all seems unfair. I think the responses need to be consistent.

In my previous post, I said longer back-and-forth edit wars are more of a problem than ABA cycles, and I also previously commented that I think we could possibly be more lenient with how we handle ABA cycles. I didn't say that there was a problem with taking the ABA structure into account on its own; I just commented on the general discussion about possibly adjusting how we handle them.

Edited by GastonRabbit on Jan 1st 2024 at 1:13:47 PM

As a burned user, I would agree that ABA is too strict because it's easy to trigger such a case by accident. In my case, I was cross-wicking an entry from the work page to the trope page and got banned because someone removed that entry from the trope page and I did not see the comment in the history. That's already ABA without realizing it. ABABA is a more telling proof of misconduct especially since the punishment is quite drastic and humiliating.

I agree that we probably shouldn't jump to automatic suspension unless there's something to suggest that the user is deliberately going against consensus. That can be ABABA, combative edit reasons, or ignoring private messages asking them to discuss the edit.

Making that the criteria might work. Regarding combative edit reasons specifically, those can already earn someone a suspension separately from edit warring, but in the context of reinstating an edit someone else undid, it makes it clear that it wasn't an accident.

Edited by GastonRabbit on Jan 5th 2024 at 10:59:53 AM

Yeah, that's what I was thinking. An Edit War can still be ABA, but ABA shouldn't be an insta-suspension unless there's more evidence of intent. More often than not it turns out to be a complete accident, or just a misunderstanding of the rules, which can be hashed out in a discussion venue.

It's probably a good time to move from calling the policy too harsh to drafting/suggesting what needs to be changed about it, as it seems people agree in that regard.

I mean, my thought the entire time has been that the only thing that needs changing is the way the rule gets enforced. Don't alter what it means to be in an Edit War — just don't suspend off the bat for one. That's less about changing the policy, but me calling it harsh is me suggesting what needs to change here.

As a #1 cause of suspensions, it's kinda very important that it's properly defined and isn't treated like a suggestion instead of a rule. If we think the intent should be weeding out uncooperative tropers, an edit revert would be a good, but not a definite sign of, that's what the page should be saying.

Well, no. I think that what it means to be in an edit war shouldn't change, because it makes sense and it would be super confusing to change it on everyone. I just don't think that it warrants an immediate suspension. Most other violations don't, and an edit war can be resolved just by reaching out to the parties involved.

In my mind, the only necessary change is in how people react to an ordinary ABA situation.

Edited by WarJay77 on Jan 5th 2024 at 2:16:30 PM

what Warjay said. i had suggested a two-strikes system, basically an edit war is still ABA, but not automatically a suspension on first offense. But if someone informs or notifies them about the edit war, then the second strike can be either a second ABA edit war, or a new edit on this one.

This prevents people from doing a dozen ABA edit wars, and have that not count as any kind of rulebreaking as long as they dont go to ABABA

Edited by Tremmor19 on Jan 6th 2024 at 5:39:26 AM

Sounds good to me. Is there an existing system for mods that can incorporate the strike concept?

Every user has a mod-only File that acts as a notepad when suspension is considered. Also notifiers are technically our strike system.

Edited by Amonimus on Jan 6th 2024 at 2:49:02 PM

Yes, if we respond to the second A in the ABA cycle with a notifier instead of a suspension, the system does already keep track of how many of a specific type of notifier someone has received, so if we added a notifier for edit warring, it would say how many edit war notifiers someone received.

Edited by GastonRabbit on Jan 6th 2024 at 9:54:47 AM

I think that while such a "strikes" system would reduce the number of suspensions, it still would require moderators to intervene every time someone repeats an edit. Most of the other problems mentioned in the OP would still apply, even if in weakened form. It's a half-solution at best.

On the first page, I previously mentioned the idea of acting if the notifier is ignored (instead of the current system of acting right away) and the edit is reinstated another time without any discussion after the notifier is sent, in addition to the strike system that was discussed on this page. Would that be any better?

Edited by GastonRabbit on Jan 7th 2024 at 2:47:24 PM

From what I've personally summarized from the conversation so far

• Reconsider what counts as "edit-warring". Instead of the A-B-A editing cycle immediately counting, treat it more like a red flag for a potential, but preventable, prolonged back-and-forth.
• Loosen the suspension policy, from a single revert, to leaving up to moderator's digression if the revert is an indication of bad faith editing.
• Even for cases that would count, prioritize informing the offending troper of the edit-warring policy without suspending them first.
• Encourage people to invite people suspected of edit-warring to the discussion avenues first, and opening Ask The Tropers reports only if that doesn't work or if there's a sign of bad faith editing.
• Implement an "edit-warring" Notifier.
• Launching a trope is to be considered as "community contribution" as TLP drafts require voting, instead of counting as the sponsor's edit.
• Encourage tropers to settle disagreements privately, instead of asking for "permission to avoid edit-warring" in Ask The Tropers. Ditto from "preventing potential edit-warring", one doesn't need to "ask to ask" multiple editors.
• Clarify that third parties reverting an edit, while "technically edit-warring", is not what the policy is for, though is still worth attention.
• Assume a revert accidental if there's no evidence for otherwise, but still requiring asking the troper in private. One possible reason is even covered by Edit Stomp.
• Check and adjust Edit War page where necessary.

As for the potential notifier, wingedcatgirl's proposal:

• The edit you made on [page] appears to have reinstated a previous edit of yours without discussing its removal. Wikis like TV Tropes are collaborative projects, and since anyone can edit any page at any time, disagreements are most constructively handled by discussing the issue and coming to a consensus, rather than by brute force.

My proposal (which I've written before noticing there's one already)

• edit-warring: Greetings. The [page] edit you made appears to have reverted another edit to a (close enough) version previously made by you, without a justifiable reason (such as consensus or fixing violations). Per Edit War, it is crucial that disagreements with edits are handled by getting approval first, from involved parties or the community, and edit reasons don't count as a discussion. It's also a good practice to check the page's history tab to avoid accidental reverts.

Neglected to mention that in addition to discussion pages being one way (and the primary way) to hash things out after the notifier is sent, I was thinking another way to do it would be to add additional custom text after the notifier to use as a starting point for discussing the disputed edits over PM, in the event that only two people are involved (and using the discussion page if discussion is needed from more than two people, due to the use of PMs alone being impractical in those cases).

Edited by GastonRabbit on Jan 7th 2024 at 4:39:12 AM

@Completely Normal Guy That term "consensus" is it's own can of worms. If a small group declares something to be true, but it either isn't or gets changed in the work, then that could spawn an "edit war" right there. We do have a bit of a clique problem.

But yes, I'm all on board with encouraging people to talk things out privately. The problem comes when someone starts getting snippy or inventing new rules and that's when the mods should get involved.

You're not wrong but a lot of people don't really use the Discussion Pages these days, especially for relatively small edits. Things either go straight to P Ms or don't get talked about all.

But consensus is the primary way of settling things on the wiki and how it's always operated?

Edited by Amonimus on Jan 8th 2024 at 6:55:36 PM

Catching up, and yes, I always thought the ABA thing was a bit strict. Hell, I even ran afoul of it as a younger troper, because I just forgot I had already added the entry to begin with before it was deleted and I re-added it, though I didn't get caught at the time. Like others said, it's too easy to see as an accident. I've also never fully believed in the "a suspension is just a warning" principle because it doesn't really line up with the other principle of "if you get suspended multiple times, even for different things, we're not likely to reinstate you," especially when it comes to things that are newbie mistakes like edit warring or TLP issues.

I think a suspension after an ABABA, or even after an ABAB, would usually make sense, but I think other factors should be considered, like whether the participants use edit reasons and thus give an idea for their motive (i.e. a personal bias about the subject, or just general rudeness). I also fully agree with using notifiers first, even for an ABA, just to make things clear and seek consensus. I think ATT should only be used if you think you need other tropers' input on a situation, like if it's a disputed piece of information being warred over.

That's fair. I was just going based on past experience of seeing edit wars or disputes hashed out in ATT, so it was my immediate choice. Reserving it for contested edits seems reasonable.

Mmm. ABA is the easiest pattern to pick up on, and the soonest, but defining that as a edit war means that the rules are going to catch people out a lot more for not noticing the edit history, or just trying to discuss things the wrong place or way than actual unwillingness to discuss.

Suspension as a warning does make sense if you really sit down and consider just how many people are only in the appeals thread for unconcious edit warring. Rookie mistakes shouldn't be treated with the same seriousness as a guy sending threatening messages or wiping out pages he doesn't like.

The punishment not fitting the crime (and the impact these suspensions have on the Edit Banned workload) is why this discussion started, and why we're trying to figure out how to reform the system for handling edit wars.

On the first page (in a post I linked to on this page), I mentioned the idea of switching to using notifiers as a first resort for the second A in the ABA cycle (instead of the current system of it being a no-questions-asked reason for a suspension) and only suspending if that doesn't work.

Edit: Neglected to mention that the ABAB pattern (mentioned in mightymewtron's post) is already suspension-worthy under the current system (for both the second A and the second B), except that results in two people getting suspended, while ABA results in one suspension (for the second A). If we want to loosen the restrictions, ABABAB makes more sense for what's suspension-worthy for two participants if ABABA is the new suspension-worthy pattern for suspending a single participant.

Edited by GastonRabbit on Jan 9th 2024 at 11:31:36 AM