Wikipedia talk:Wheel war/Archive 3

From Wikipedia, the free encyclopedia
Jump to: navigation, search
Archive 2 Archive 3 Archive 4

Old discussion available at Wikipedia talk:Proposed wheel warring policy/Archive

The previous version of the project page (with multiple proposals)


The difficulty that we have to address is the definition of what wheel warring is. We all know it when we see it of course. I'm worried that by trying to nail down a definition we will cause problems. I have undone the actions of other admins on quite a few occasions but I have never wheel warred.Consider this- Admin A blocks a user. The user is on a shared network so the block causes collateral damage. Admin B unblocks the user after an email from a innocent user. B for some reason forgets to tell A. The next day A notices vandalism coming from that user and blocks him again. According to some of the definitions we are considering A is guilty of wheel warring! Theresa Knott | Taste the Korn 16:20, 18 February 2006 (UTC)

A is not guilty of "wheel warring", he's guilty of violating the 1RR on admin actions. If we set the bar at some vague notion of "wheel warring", then rules lawyers will have a field day and wheel warring won't be curbed. — Phil Welch 00:29, 19 February 2006 (UTC)
Admin A would not even be guilty of breaching 1WW, s/he would be at the limit and ought at that point to get messages telling him/her to be more careful about checking WHOIS before placing long blocks (or hopefully more polite messages explaining the situation). Overturning another admins decision once can sometimes be justified, I'm not sure it could ever be justified overturning it more than once. Physchim62 (talk) 06:56, 19 February 2006 (UTC)
Right. If admin B goes back and unblocks without talking it over with A, then B would be in violation. — Phil Welch 06:59, 19 February 2006 (UTC)
I think most of us are in agreement that B would have done something naughty there, yes. Maybe excusable if it were several days later, but the very least that should have occured, even before B's second revert, is that the potential dispute be brought to the attention of outside parties (WP:AN, ArbCom or Jimbo depending on the exact situation). Physchim62 (talk) 07:24, 19 February 2006 (UTC)
For this specific example, I suggest that editors check out Wikipedia:Blocking policy proposal, which has to do with IP blocks on shared networks.


Just a note for anyone paying attention should any of these ideas be implemented: please update the WP:1WW shortcut to point to wherever this is ultimately enacted at (if it is). —Locke Coletc 04:41, 23 February 2006 (UTC)


As of 04:39, 6 April 2006 (UTC), the last edit to 1WW is nearly 2 weeks old. There are now 5 related proposals under discussion. I suggest that we move forward; I feel that multiple proposals may make it difficult to build consensus behind any one.

At this time, the straw poll stands at:

proposal # author support oppose
1 Philwelch 4 14
2 Carnildo 3 3
3 Theresa 6 5
4 John Reid 20 7
5 Locke Cole 10 6

Clearly, none of us has got it quite right; but I think we are all tending toward a certain point. I'd like to ask consent to archive the existing page as it stands now; and attempt a refactor that includes the most popular and accepted points of all 5 proposals. I'd also like to incorporate useful portions of Wikipedia:Wheel war.

From that point on, while we may edit the proposal on its page, let's not fork it again. We can discuss it on talk. John Reid 04:39, 6 April 2006 (UTC)

Unfortunately, it appears there already is an existing "fork" at WP:0RR. It has not been put to a poll yet, but it also seems to enjoy some degree of consensus. I'm not sure how we ended up with two redundant proposal pages; this is the first time I've heard of WP:1WW. I'll discuss this futher below. —Ilmari Karonen (talk) 11:49, 12 April 2006 (UTC)

"Admin zero-revert rule" proposal

There's another similar proposal, started in March, at WP:0RR. It's somewhat more detailed than the proposals on this page, and somewhat stricter than Locke Cole's proposal. (My interpretation of John Reid's proposal would be compatible with the 0RR one, but it appears John himself may disagree.) It's been fairly stable for over a month now. It hasn't been put to a poll, but there seems to be a rough consensus on the talk page in favor of the latest versions.

Should we put it to a poll and see how it fares compared to the alternatives listed above? Or is there sufficient consensus to implement it based on discussion alone (since, as we all know, voting is evil)? —Ilmari Karonen (talk) 11:49, 12 April 2006 (UTC)

Polls are evil, let's avoid one if at all possible. I'll take a look at 0RR more closely tomorrow (I should be asleep now :P) and comment. But I can't imagine it "stricter" being a deal breaker. =) —Locke Coletc 13:12, 12 April 2006 (UTC)
I prefer it (0RR) to anything mentioned on this page. Johnleemk | Talk 15:44, 12 April 2006 (UTC)


Oh boy, another fork! Apparently some folks got started at Wikipedia:Admin zero-revert rule unaware that WP:1WW has been cooking since early February. I'm going to merge it into this discussion, since the substance is extremely similar. John Reid 18:06, 13 April 2006 (UTC)

My proposal came mainly out of a discussion on the ArbCom channel. Sorry I missed the other one, but nobody told me about it. It's probably a good thing that the two threads have merged, as I think we need something more than we have now. Kelly Martin (talk) 00:06, 14 April 2006 (UTC)

Discussion of Kelly Martin's proposal

This discussion originally took place at Wikipedia talk:Admin zero-revert rule.

So far this proposal sounds very good. Matthew Brown (Morven) (T:C) 07:02, 9 March 2006 (UTC)

Can we have it, please? --Tony Sidaway 07:04, 9 March 2006 (UTC)

It's pretty much what already do (admitedly in my case with the added view that whoever pulled my descission is responcible for any problems that arise as a result). The only posible issue is that I'd like a garentee that arbcom would look into every case of admins pulling blocks placed on themselves other than mistakes due to the autoblocker.Geni 12:34, 9 March 2006 (UTC)

Sounds fair enough. The "clear and immediate danger" exception is a fairly big loophole, though, and at least ought to be accompanied by a link to WP:PANIC. I'd almost be inclined to remove it and let any cases of actual clear and immediate danger be covered by IAR. But I'm willing to support this either way. (The "or else" part of the nutshell summary ought to be eliminated or reworded too. As it stands, it sounds almost comically menacing.) —Ilmari Karonen (talk) 12:36, 9 March 2006 (UTC)

Some questions

A reasonably good proposal in priciple, although I think some modifications (i.e. more exceptions) might be in order:

  1. Scenario: Admin A blocks vandal X, using a shared IP. Another user of the blocked IP realizes he is blocked and that admin A is off-line and e-mails admin B to request unblocking, something which is granted. Some hours later, the original vandal goes on a vandalism spree. Should admin A be allowed to reblock?
  2. Scenario: Admin A closes an AFD debate as a "delete", admin B disagrees and decides to simply undelete the article. Should Admin A be allowed to redelete?
  3. Scenario: Admin A protects an article because it is being targeted by vandals. An hour later, admin B decides the attacks are over and unprotects. Three hours after that, the article is subject to a new barrage of vandalism. Is admin A entitled to reprotect?

I would argue that in scenario 1 and 3, the undoing of the other administrator is not a wheel war, the situation has changed, and the admins in question may be the best of friends and endorse eachothers actions wholeheartedly here. In scenario 2, admin B's decision is "out of process", and the rules mandate that a disagreement be taken up with Admin A, the Admins' noticeboard or DRV. Also, an interesting counterpart to Scenario 2 is

  1. Scenario: Admin A closes an AFD debate as a "keep", admin B disagrees and decides to simply delete the article. Should Admin A be allowed to undelete? (This is slightly different from the first one because no administrerial action is technically made when closing a debate as a "keep".)

Sjakkalle (Check!) 12:36, 9 March 2006 (UTC)

In both of the deletion-process scenarios, no, admin A should not be allowed to redelete or undelete. Discussion is warranted, not reversing action. There is no criticality in such matters; it matters very little whether some page is deleted or not; it can sit "in the wrong state" for a few hours or days while consensus is reached.
In the blocking and protection situations, admin A needs to remember that he does not stand alone against the forces of chaos and enlist the aid of a third administrator (we have over 800 of them, surely admin A knows someone he can find) to repeat his or her action. I see no need for either of these modifications, especially since either of them would create loopholes for rulelawyers to surf.
Far too many wheel wars are conducted in part by admins who believe that they are Wikipedia's "last, best defense" against whatever ill is threatening Wikipedia at the moment. One of the purpose of this policy is to put a stop that attitude. We are supposed to be working together as a community, not running about as a bunch of lone wolves passing in the night. Kelly Martin (talk) 12:46, 9 March 2006 (UTC)
Why not apply the same "out" for vandalism for the 3RR to this 0RR? If by repating a reversed administrative action, an admin prevents obvious, currently-occurring vandalism, it should be exempt from this rule. There have been times when I've been unable to get another admin's attention, and not everyone has access to/desire to use IRC. If Admin A protects an article from a vandalbot, Admin B mistakenly unprotects, and the vandalbot returns to insert pictures of penises into an article, how can we possibly fault Admin A for taking quick, decisive action to prevent obvious and serious vandalism?
That said, I agree with the deletion/undeletion scenarios, but Sjakkalle presents an interesting inconsistency with the Admin A keeps/Admin B deletes/Admin A undeletes scenario. android79 14:47, 9 March 2006 (UTC)
I still think it's unnecessary to make an exception for vandalism. Admins should make an effort to network with one another anyway. With over 800 admins, I really think there's no excuse for not being able to find another admin to take quick action. And putting in an exception for vandalism invites attempts to surf the loophole thus created. No, I think this is exactly the "I am the last defense against the forces of chaos" attitude that is so poisonous in an administrator. Kelly Martin (talk) 18:12, 9 March 2006 (UTC)
I just don't want to have to do what amounts, essentially, to asking for approval to do something that prevents obvious damage to the encyclopedia. I've been entrusted by the community to use these tools. If another admin makes an obvious mistake, I should be allowed to correct it – with an explanation in a prominent place (the admin's talk page, ANI) of course. android79 18:36, 9 March 2006 (UTC)
This is a very, very bad idea if applied blindly, yet it seems from the above comment that is precisely the intention. When clearcut vandalism is taking place, you revert it on the spot, you don't "network". -- Curps 04:19, 10 March 2006 (UTC)

I agree that all above scenarios can be resolved by administrators on the ground working together. Scenarios 1 and 3 in particular suggest an administrator who has decided that he owns an article (albeit 3 is a little artificial because in practice administrators do not unprotect in the presence of very recent vandalism) Scenario 2 and its converse, a keep debate reclosed as a delete, is a classic case. There will be gaming in this area and it should be treated firmly. --Tony Sidaway 18:19, 9 March 2006 (UTC)

  • It seems to me that Scenerio #3 most likely and frequently occurs when the protected page in question is either a talk page, user talk page, one linked off the main page, or any other page in which WP:PPol specifically states that it is not appropriate to protect them unless in extreme circumstances, and only in a very short time period. other than have an exception, do you suggest any other solution to that problem? Zzyzx11 (Talk) 18:20, 9 March 2006 (UTC)
    • Here's a possible situation: Suppose I notice that the Today's FA is being bombarded with vandalism. But if I protect the page (which I may do temporarily to repair extreme and severe damage), I would violate the Admin zero-revert rule. So while I attempt to contact another admin, the vandals have more time to screw up the page even more. Zzyzx11 (Talk) 18:29, 9 March 2006 (UTC)
      • See Wikipedia:Don't panic (and possibly also Wikipedia:Adminitis). —Ilmari Karonen (talk) 19:00, 9 March 2006 (UTC)
        • Actually, I have never been in that actual situation. But I brought it up because I have been in the opposite position of unblocking an article linked from the main page only to be reverted moments later. Instead of actually getting into a wheel war, I just post a message on ANI saying "why are we protecting today's FA..." and have someone else (like our fearless ratified director of the Wikipedia:Today's featured article) get into the debate while I do something else... Zzyzx11 (Talk) 20:42, 10 March 2006 (UTC)

In general, I do agree with Sjakkalle's original claim above that there are situations where admin actions may be (seemingly) undone and redone, possibly even several times, without there actually being a wheel war going on. I'd hope that these would be covered by common sense, and in any case they generally can be avoided by asking for help, but I wouldn't be opposed to a note stating something like "This policy is not intended to apply to cases where the situation has clearly and significantly changed between the undoing and the redoing of the admin action in question, especially where the stated reason for undoing the action no longer applies. For example, if an article unprotected after a vandalism spree is over, and the vandalism later begins anew, the same admin who originally protected the article may protect it again.". —Ilmari Karonen (talk) 21:40, 10 March 2006 (UTC)


Should this apply to rollback? Haukur 12:41, 9 March 2006 (UTC)

IMO, yes. If you're going to revert war with another admin, at the very least have the decency to do it the manual way. —Ilmari Karonen (talk) 12:45, 9 March 2006 (UTC)
Agreed. It applies to the use of rollback. Kelly Martin (talk) 12:46, 9 March 2006 (UTC)

No, it should not apply to rollback. That's what we have the 3RR for. This muddies the waters between admin actions and regular editing actions. Any editor can manually accomplish a rollback, and there are many scripts, available to all, that emulate the rollback feature. android79 14:38, 9 March 2006 (UTC)

While I don't really have a strong opinion on whether this needs to be applied to rollback or not, I do believe that if you ever find yourself wanting to rollback a rollback, you should 1) think twice about what you're doing, and 2) revert manually with an edit summary that explains why you'd do such a thing. There just is no valid reason for two successive rollbacks, much less three. —Ilmari Karonen (talk) 15:23, 9 March 2006 (UTC)
"There just is no valid reason for two successive rollbacks, much less three." You could say the same thing about reverting in general. I agree that, under usual circumstances, no one should find the need to rollback a rollback, but that's largely an issue of etiquette, not abusing the extra buttons. android79 15:30, 9 March 2006 (UTC)
To elaborate, I can think of the following cases where one might want to rollback a rollback:
  1. You're revert warring with another admin.
    • Don't. Just don't. If you do anyway, at least do it manually and provide an edit summary.
  2. You're dealing with a vandal who is emulating rollbacks manually or with a script.
    • The 0RR rule doesn't really apply here, but I'd still recommending a manual revert with a summary like "reverting sneaky vandalism by X", possibly after blocking if the vandal is still at it.
  3. You're dealing with a vandal who is an admin.
    • If that's really the case, you're out of your depth. Ask for help on WP:ANI, the mailing list, or IRC. Then deal with as above.
  4. You want to correct an honest mistake by another admin, and can't be bothered to contact them first or even to do a manual revert.
    • Fine, be lazy and rollback, then. The 0RR doesn't actually forbid this, it just says the original admin shouldn't then rollback you. If you find yourself on the receiving end of such a double rollback and disagree, see case 1.
Did I miss anything? —Ilmari Karonen (talk) 15:38, 9 March 2006 (UTC)
I agree with applying this to rollback. If there are admins rolling back other admins, then I would hope the ensuing public discussion would emphasize that this is not acceptable. (An issue pops up if WP:RFR goes through, but we're not there yet.) I disagree with the distinction made above that "an issue of etiquette" is not abuse, as the entire body of policy deals with what is acceptable behavior on Wikipedia, which is a stricter way of saying 'enforceable norms or etiquette'. Overall support for this proposal, though I find it sad that we have to spell out that admins shouldn't treat others who have been entrusted with admin privileges as if they are trolls or blundering newcomers. - BanyanTree 16:57, 9 March 2006 (UTC)
It is a matter of etiquette. Let's say Alice performs a rollback, and Bob rolls back Alice's rollback
  1. manually with a blank or non-descriptive edit summary;
  2. automatically with a script that provides no reason in the edit summary;
  3. automatically with the button.
Under this proposed policy, Bob could only be desysopped for action 3, when actions 1 and 2 achieve the same effect and are just as rude. android79 17:11, 9 March 2006 (UTC)
I would consider 2 to be an attempt to game the system, which is covered under the proposal. Manual reversion (per 1) has two effects: it allows descriptive edit summaries and it slows down the process of reversion from the "zap"-mentality of rollback - whether admin or script-based. Even if an admin is in an unholy rage and doesn't bother explaining his/her reverts, it can be hoped that the soothingly pastel pink/violet box informing users that they are editing an old page would offer a chance for the user to realize that they have lost their cool. If not, at least it is clear that the admin is being stupid as a user, and not as an admin, and thus not wheel-warring. (I don't expect admins to be models of decorum and logic all the time, as I certainly am not, but when they lose it they shouldn't lose it with the extra buttons.) In any case, I trust that if there really is a borderline case then the 'crats will make a judgement call that deadminship is not appropriate or that it will go back to an new RFA where voters will realize that the removal of admin privileges was bogus. - BanyanTree 18:10, 9 March 2006 (UTC)
Yes, both actions 1 and 2 are "gaming the system" in some way. That's what makes inclusion of rollback a bad idea – it opens up the possibility of gaming the system. Each other action that's been proposed to fall under the 0RR is admin-only, and it's obvious if Bob reverses Alice's admin action – either he unblocked Carlos, or he didn't. When you can apply a rule to non-admin actions that's supposed to apply to admin actions, that's where we run into problems. android79 18:21, 9 March 2006 (UTC)

I've added rollback explicitly for now because rollbacks contain no proper edit summary (I would argue that they should, but that's a technical matter). If an administrator reverts a rollback, it would be a good idea if at least that revert contained an edit summary, if only so that those watching Recent changes will havea clue as to what's going on. --Tony Sidaway 18:03, 9 March 2006 (UTC)

I don't agree that rollback should be added. Rollback is an editing function; anyone can rollback, it's just that it's slightly more convenient for admins (1 click instead of 3). Edit-wars are already covered under existing policies, and we should avoid instruction creep. This policy should deal with purely administrative actions, i.e., actions that can only be carried out by those with admin abilities. Jayjg (talk) 23:16, 9 March 2006 (UTC)

Rollback should not be included. Rollback is only to be used against vandalism in any case, using it in any other context is an abuse of admin power anyway. Rich Farmbrough 23:05 19 April 2006 (UTC).

Question on enforcability

According to the proposal, the enforcement will be temporary or even permanent desysopping. Is it a problem that we have very few people capable of imposing such sanctions? Isn't it easier to treat wheel-warring the same way as we treat edit-warring, with short blocks of up to 24 hours? Sjakkalle (Check!) 13:35, 9 March 2006 (UTC)

Blocks are not an effective way of dealing with abuse of administrative power. Kelly Martin (talk) 13:37, 9 March 2006 (UTC)
Hmm... I just did a test myself. It seemed pretty effective. Being blocked knocks out the use of rollback, deletion and protection. Only the blocking seems in tact. Sjakkalle (Check!) 13:42, 9 March 2006 (UTC)
While blocking as a short-term action pending desysoping is a possibility, I think offenses of this nature deserve a more serious penalty than a mere 24 hour block. Adminship is supposed to be no big deal, so losing it should also be no big deal. Kelly Martin (talk) 14:20, 9 March 2006 (UTC)

This rule will not have teeth unless it carries a desysopping penalty. We should aim at making it a rule that is so universally followed that no administrator would even think of using his sysop power to perform a reversal. Regaining of sysop rights after an offense should always be subject to approval by RFA. --Tony Sidaway 18:08, 9 March 2006 (UTC)

The fundamental flaw of this sort of penalty is that it allows a few select individuals (the Stewards) to supplant the standing consensus of the community. The community made people admins by consensus, it should require a consensus to desysop someone. There are plenty of current admins who could not win an RfA right now. As long as only a select few (those who get desysopped) must face an RfA while the rest of us slide by without review, you have handed a very subjective power to individuals who were not elected with the clear knowledge that they would be exercising this power. NoSeptember talk 18:23, 9 March 2006 (UTC)
This rule is only intended to curb abuse of administrator power, not improve the overall quality of administrators. --Tony Sidaway 18:31, 9 March 2006 (UTC)
Determining abuse is a subjective thing. Should a Steward, who likely got elected because they know multiple languages, be making these decisions? Especially when the decision will likely be the difference between remaining an admin and being permanently desysopped (because you picked up enemies and can no longer pass RfA)? I would be OK with a temporary desysop with the understanding that the admin would be automatically resysopped after 48 hours or whatever. But we had people who were "temporarily desysopped" (read the statement made at the time) on Feb. 6, who are now permanently desysopped for all practical purposes. A temporary sanction is fine, but there is nothing to indicate that such sanctions will indeed be temporary. So the Stewards will become one person ArbComs. We didn't elect Stewards for this. Finally, we are creating an all or nothing penalty, screw up once and you are gone. We treat our vandals better than that. NoSeptember talk 18:46, 9 March 2006 (UTC)

may -> will

I suggest changing "Do not repeat any reversed administrative action within 24 hours. (Or a Steward may remove your administrative privileges.)" to "....(Or a Steward will remove your administrative privilages.)". Comments? --Rebroad 17:48, 9 March 2006 (UTC)

This assumes that Stewards are omniscient. :-) android79 17:51, 9 March 2006 (UTC)
...and never forgive nor use common sense. —Ilmari Karonen (talk) 18:10, 9 March 2006 (UTC)

No Thanks, the rule will backfire

This may be fine as a general guideline, but if it becomes a firm rule, the consequences will likely be bad. This policy seems to be an invitation to trolls and vandals to game the system to get their way and succeed in attacking good admins. If a vandal/troll who gets blocked (or has an article deleted etc.) is able to convince another admin to reverse the action, it is now the other admins (who may realize that the undoing admin was played), who are under the gun, primed to lose their adminship if they reestablish a proper action. So perhaps no one will want to risk immediate desysopping, and the vandal/troll gets to wreck havoc upon Wikipedia. We are trying too hard to make rules fit dynamic situations. There are too many instances that the rule being applied could be worse than not applying it. Instead, we should just insist that there be discussion, with the other admin involved or with the community on the noticeboard, and real humans can make consensus decisions based on each situation. NoSeptember talk 18:08, 9 March 2006 (UTC)

...and "we should just insist that there be discussion" differs from the proposed policy how? I suppose one might have a discussion, find the consensus uniformly against oneself, and then decide to screw it and revert anyway, but I doubt that's what you were thinking of.
Anyway, I think there's an underlying interpretation issue here. The way I read the proposal, if admin A does something and admin B undoes it, admin C is still free to redo it (but only once). If a vandal/troll is blocked and then "able to convince another admin to reverse the action", there ought to be plenty of other admins available to reinstate the block if it really was merited. Or one could just try to talk sense to the admin who did the unblocking. —Ilmari Karonen (talk) 18:20, 9 March 2006 (UTC)
"Hi, you unblocked X whom I blocked, but since then he has vandalized articles A, B and C and here are the diffs. Please reblock."
If the admin isn't available (or even if he is) bring it up on WP:AN/I and it should be easy enough to get agreement. --Tony Sidaway 18:28, 9 March 2006 (UTC)
If they are vandaliseing repeatedly someone else will probably block them before you can get to AN/I.Geni 20:10, 9 March 2006 (UTC)
Oppose I feel some Admin will make mistakes, or "power trip" for which another admin must

correct for. This policy gives individual admin too much power.

--Masssiveego 23:51, 9 March 2006 (UTC)

We need a real zero-revert rule

The issue I see with the 0RR as drafted is that it actually gives *more* power to crusaders, and doesn't actually stop very common kinds of wheel wars.

Here's what I mean: One admin protects a page, or blocks and editor, or whatever. Then a second admin comes along and undoes all of his actions. The second admin has *already* started the wheel-war, and, in fact, *won* the wheel war, because the first admin cannot revert the second admin.

Here's another common scenario; Admin A blocks an editor (or deletes a userbox). Admin B comes along and unblocks/undeletes. Admin C reblocks/redeletes. Admin D unblocks/undeletes again, etc. The wheel war is in full force, yet no-one has violated the 0RR as proposed.

Unlikely? Perhaps, but that is pretty much exactly what has happened recently in a number of recent blocking cases and various userbox wars.

This proposed policy is an attempt to stop admins from being disrespectful to other admins, and to stop wheel-warring; yet, as currently formulated, it does neither. A real 0RR would insist that an admin cannot undo the administrative actions of another admin within a 24 hour timeframe. Full stop (of course, inserting all the various caveats regarding resonable acts, emergencies, endangering, etc.). Jayjg (talk) 23:14, 9 March 2006 (UTC)

That was my first reaction when reading about zero-revert. Wheel wars can and have happened without an admin repeating an action at all. I think that good faith assumptions (with regard to the original admins action) should have some weight and that the first revert should be where the bright line is. Maybe there needs to be a couple admins involved when reverting....if you want to revert something you should find someone that agrees with you. Rx StrangeLove 00:03, 10 March 2006 (UTC)
I would have no problem supporting a rule akin to what Jayjg describes, although I'm concerned about writing the exception cases so they cannot be gamed. Kelly Martin (talk) 03:30, 10 March 2006 (UTC)
Jay's description is very good but I disagree with his conclusion. The currently proposed policy implies that if two admins strongly disagree about an admin action then that action will not go through because the second admin will, as Jay puts it, win the wheel-war. But we need not view it in those terms. If two admins disagree whether an admin action should be performed they should discuss it before it is applied (again) and maybe get more opinions. The proposed policy will lead to this.
And as Kelly points out it is very difficult to write a policy that disallows reversion of admin actions in normal cases and allows it in exceptional necessary cases.
Jay is correct that 0RR is a misleading name for this policy. We should come up with another name. Haukur 14:48, 10 March 2006 (UTC)
How about Wikipedia:No wheel wars? —Ilmari Karonen (talk) 16:12, 10 March 2006 (UTC)
Fine by me though people seem to have variant definitions for what constitutes a wheel war. Wikipedia:Don't repeat admin actions? Clear, though not very catchy. Haukur 16:46, 10 March 2006 (UTC)

24 hours

The problematic issue with this proposal is the 24 hour limit. The static limit is exactly what allows someone to game to system as discussed in the previous sections above. Somehow, the static limit needs to be removed. — Dzonatas 16:21, 12 March 2006 (UTC)

Unnecessary and gives too much power to 2nd admin

This much like the 3rr is flawed in that one of the 2 sides in the (wheel/edit) "war" is given extra power. In this case the 2nd administrator's (the reverser) is given absolute power over the other administrator's decision. Example, an admin unblocks a user that was blocked a few days ago as a suspected sockpuppet because there was no concrete evidence for the original block, now a 2nd admin re-blocks the user because he feels the circumstancial evidence was enough and then logs off for the night, then an admin finally runs checkuser(which is backed up so a pre-emptive block is common) and clears the user of the accusation, however he cannot unblock the user for another 20 hrs since the 2nd admin's block is protected and is forced to let the user sit out his unfair block while knowing that he was cleared.

The problem with this policy is that it is an attempt to make a blanket cure for an issue that needs to be adressed on a case by case basis. Wheel warring is already a big no-no people should be punished for wheel warring if the situation presents itself, there is no need for another policy that will eventually be gamed. Seraphim 20:53, 13 March 2006 (UTC)

Unless the admin doing the CheckUser is the same admin who unblocked the user originally (In which case why didn't he just do the CheckUser in the first place?), there is no reason (as I interpret this proposal) why he couldn't unblock him again. And, in any case, a CheckUser probably counts as a "significant change" in the situation.
In general, it may be taken as given that in a wheel or edit war both sides are by definition in the wrong (see meta:Wrong version). As with a schoolyard fight, it does not matter who started it or who is right or who gets the last word; all that matters is that the fighting needs to stop. After the fighting has stopped, others may then look into the situation and figure out what actually should be done about the underlying issue. —Ilmari Karonen (talk) 15:09, 16 March 2006 (UTC)
With the new section you added to the page it is fine. :) Seraphim 08:02, 17 March 2006 (UTC)
It doesn't give any more power to the second admin than to the first: it merely means that, if the first admin finds that his/her actions have been overturned, s/he has to find a third admin and explain why they should be reinstated. This is surely a better course of action than just blindly reverting the revert. Physchim62 (talk) 13:30, 12 April 2006 (UTC)


It stirkes me that a) we should try and get a policy to sort the situation of bad administrators as opposed to trying to limit their power. We should sort out the root of the cause no the effects of the problem. That said, I think the proposal is a good idea, but like I said, I think that Admins should have a much more stringent oversight. --Wisden17 19:53, 28 March 2006 (UTC)


There are a number of similar proposals at WP:1WW. I'm not sure how we ended up with two redundant proposal pages, but somehow we did. As there's been a proposal to ratify some version of the 1WW proposal, and as this proposal seems to be fairly stable and enjoy some degree of consensus support, I've suggested on the "other" talk page that both proposals be used as basis for the final policy. Please comment there. —Ilmari Karonen (talk) 12:01, 12 April 2006 (UTC)

It's not important how we forked; these things happen. Nobody's to blame and it's not a problem. I'm merging the two together. John Reid 18:25, 13 April 2006 (UTC)

New discussion

I don't like the term "wheel war" because I believe it is unclear and misapplied at Wikipedia, and I am wholly unconvinced that having a policy regarding "wheel warring" will accomplish anything useful. Admins are supposed to discuss things, be civil, and respect our way of doing things, and any admins that don't shouldn't remain admins. I don't think it's possible to violate any of proposed policies (at least, the ones that have meaningful support) while still discussing things, remaining civil, and respecting our way of doing things. Instruction creep is bad, and so are rules that can be gamed. I believe that the perceived mandate for this stems from conflating a number of unrelated cases:

  1. The unfortunate matter involving editors reverting Jimbo's deletion of a userbox, leading to an unusual degree of scrutiny and handwringing over a matter that would otherwise be an unremarkable deletion quarrel.
  2. The actions of several admins who did not prove trustworthy enough for their ring of keys. With the number of admins we have now, bad admins will always be with us and a certain amount of ongoing housecleaning in this area will be necessary. Bright-line policies like the ones proposed here make this harder, not easier, because bright-line policies are easy to game.
  3. Certain narrow but ongoing policy disagreements on various admin-only functions, among them whether or not rollback may be used other than for strict vandalism, whether and how copyvio and otherwise inappropriate images should be deleted, and how to deal with content that some admins consider harmful to the project.

I would suggest that the core issues of: Jimbo's authority and role in the project, how admins of dubious distinction should be dealt with, and the resolution of the certain other policy matters listed above would do far more to reduce conflict and further our goals than any of these "wheel warring" policies. The Uninvited Co., Inc. 21:07, 17 April 2006 (UTC)

I'd like to thank this editor for his comment. To an extent, I agree. It should not be necessary to have explicit rules governing every area of permissible conduct, especially conduct of members of a group that has been singled out for special trust. I'm afraid I can't agree that it is not necessary. There is a large volume of housekeeping tasks that at present can be done only by admins. The only way to work off that backlog is to entertain admins of less than unimpeachable character. That's a hard, painful statement. I'm very slow to support marginal nominees. I endorse rather summary de-opping of admins who abuse the trust they've been given. But if we gave the mop only to the most trustworthy then it would take months to get a speedy done. So, granted that admins are merely human, the time-tested way to improve performance is to set explicit limits. Another possible approach is to hire admins, pay them well, retain only the most trustworthy, give them maximum discretion, and work them hard.
I agree that this proposal (like most policy initiatives) is driven by specific cases of abuse. Perhaps the driving force is actually the mass deletion of UBX and reversions of same. But it doesn't matter what inspired it; it's still needed. I also agree that this editor has raised some very good points of long-standing organizational problems. The Jimbo is an anomaly; he has himself stated he must fade away but shows no sign of doing so. Admins need more than 1WW; they need actual training. This must go hand-in-hand with a clarification of exactly what an admin's role is, its extent and its limits.
Still, I see real value in a clear, unambiguous wheel warring policy. By analogy, I feel 3RR to be childish; I can't imagine a situation that would provoke me to revert another editor twice in a day, let alone three times. But it has been shown in practice to have a positive effect.
Rules most certainly can be gamed. It's important to me that 1WW be made as game-proof as possible. I believe that the longer the text of a rule, the more prone it is to gaming; every effort to close a loophole creates an opportunity. That's why my original text was but a single statement. That may be too short and other, similar proposals have advantages and disadvantages. I suggest we move forward with a single draft and attempt to write something that will work for good. John Reid 23:36, 17 April 2006 (UTC)

Let's refactor

As of 03:44 Saturday, 2006 Apr 22, the straw poll stands at:

proposal # author support oppose support %
1 Philwelch 4 16 20%
2 Carnildo 3 5 38%
3 Theresa 6 7 46%
4 John Reid 24 9 73%
5 Locke Cole 11 7 61%
6 Kelly Martin 3 3 50%
7 Kelly Martin 1 3 25%

It's my belief that most of us are tending toward the same or a similar restriction on wheel warring. I suggest that a single, improved version may fare better on its way to policy. I do not suggest in any way that the plurality for version #4 constitutes a mandate for my point of view; rather, I only ask that I be allowed to refactor the existing, competing seven proposals into a single draft. This refactor will be open to editing immediately by any editor. I will ask editors to refrain from supporting or opposing the new draft for the time being; instead, to edit the proposal to reflect their specific concerns. I believe the true consensus policy will then emerge, in true wiki fashion. After all, we're not so far apart. John Reid 03:58, 22 April 2006 (UTC)

I don't know whether it's too late to propose this, but I'd like to suggest a much simpler proposal that, as a rule, administrators should not undo each other's admin actions. Exceptions would be where an unambiguous error had been made by the original admin, or where there is some urgency to the situation and the original admin is not online. If you disagree with an admin's action, discuss the issue with him/her, and if you're unable to make headway, take it to AN/I and wait for a consensus to emerge. But do not take it upon yourself to revert the admin before either agreement or consensus is reached. This is more or less what the current blocking policy says about undoing blocks. We could simply extend this to all admin actions. SlimVirgin (talk) 04:20, 22 April 2006 (UTC)
I support John Reid's suggestion. I also have one of my own: archiving the current page and replacing it with John Reid's proposal, a modified version of Locke Cole's proposal and/or a hybrid of the "most popular and accepted points of all 5 proposals?" It might be good to have an "either or" option to avoid splitting support for a proposal. -- Kjkolb 04:42, 22 April 2006 (UTC)
I agree that this is the real way to solve it. Admin actions shouldn't be undone without discussion in the first place. I do admit that the problem with some of the proposals is that it places the power in the hands of the person that does the first revert. - Taxman Talk 15:43, 22 April 2006 (UTC)
Exactly. It's the ones willing to make that first revert without discussion who are causing most of the problems, and yet they are precisely the ones most of these proposals will hand the power to. SlimVirgin (talk) 16:16, 22 April 2006 (UTC)

As the one who started this process, I agree that some refactoring would be useful, although along slightly different lines. I'll show you what I mean this weekend, hopefully. — Phil Welch (t) (c) 05:03, 22 April 2006 (UTC)

Look forward to what Phil and John have to offer. Rich Farmbrough 12:12 22 April 2006 (UTC).

How about doing this on a separate page? That would avoid controversy with people who still want to discuss the individual proposals listed on this one.--Eloquence* 13:13, 22 April 2006 (UTC)

I was tempted to do this; but that's just the kind of forking I think we need to step away from now. John Reid 09:47, 23 April 2006 (UTC)

I try not to get too involved in this sort of thing (not being an administrator, it doesn't really affect me anyway, at least not at the moment), but to be honest, while I favour John Reid's proposal over the others, I don't have any real objections to any of them. I would certainly be happy with any 'final draft' of the proposal that was roughly along the lines of what has been suggested so far – Gurch 15:38, 22 April 2006 (UTC)

  • Well, I've gone ahead, been bold, and refactored. The prior, "7 forks" version is still available in history, of course. I do not think of this as "my" proposal in any way and I strongly encourage participants to Edit this page as you see fit. John Reid 09:59, 23 April 2006 (UTC)

After the refectoring, the "1WW" mnemonic doesn't seem very logical any more, given that the emphasis is not on strict numeric limits any more. I've thus replaced it with "NWW", and have also created the redirects Wikipedia:No wheel wars and Wikipedia:No wheel warring, one of which I'd propose as the final name for this policy. —Ilmari Karonen (talk) 11:27, 23 April 2006 (UTC)

Minor note, I've modified the table at the top to reflect support percentages (feel free to correct anything I got wrong). I need to read over some of the discussion before jumping back in, but hopefully something will come of this. =) —Locke Coletc 03:30, 6 May 2006 (UTC)

I am not infallible

I've just added the following notice to my user and talk pages:

This is not meant to be a criticism of this proposed policy, which I have myself helped draft and generally support, or an attempt to dilute or circumvent it. I am merely choosing to personally assume good faith on behalf of other admins and to expressly announce a presumption of amity as described in the proposed policy.

It should, of course, go without saying that I will personally follow whatever policy is enacted, without assuming that anyone else will necessarily grant such permission to me as I have done above, and that I will in any case do my best to exercise caution in my admin actions, whether or not they may consitute a revert. —Ilmari Karonen (talk) 20:12, 23 April 2006 (UTC)

Admin opposition vs non-admin user opposition

I disagree with the "in the knowledge that another admin opposes it," part. I think Jimbo is spot on at User:Jimbo_Wales/Statement_of_principles (principle 2), where he says "there must be no hierarchy or structure which gets in the way of this openness to newcomers". Admins have the technical ability to do things others can't (by virtue of belonging to the sysop group), solely because it would be a security risk to give those abilities to all users.

Perhaps the example which most clearly demonstrates how the policy poses a problem is:

  • Admin A blocks Anon (IP) N for one week. Admin B receives an email stating I edit from IP N, which is widely shared. I am not in violation. Please remove this block. Admin B declines. Admin B messages on Admin A's talk page, saying You were wrong to block N. Block expires. Admin A blocks Anon N again.

Contrast this with:

  • Admin A blocks Anon (IP) N for one week. Admin B receives an email stating I edit from IP N, which is widely shared. I am not in violation. Please remove this block. Admin B declines. User C messages on Admin A's talk page, saying You were wrong to block N. Block expires. Admin A blocks Anon N again.

A strict reading of the policy makes the first case a violation but the second case not. This means that admin B is given a higher status (as opposed to technical access level) than user C. This applies even if user C is a much better contributor to Wikipedia than admin B. A1kmm 04:43, 30 April 2006 (UTC)

I'd like to say that A1kmm does have a point. Admins are not more important than other users; their opinions carry no more weight than any other user. But they do have more responsibility and inevitably this means that they have more authority. Admins are a class of users who have been entrusted by the community with the power to execute community policy. They are expected to exercise that power for the common good. But they are also expected to exercise it.
NWW is only intended to stop wheel wars. In the second example you cite, it is quite possible that Admin A has acted unwisely by ignoring User C and renewing the block. But since no admin has expressed opposition to the block, this is not a wheel war. Note that in any such case, real or hypothetical, one or more admins may well have violated other policies and guidelines, such as civility, neutrality, or process. Here we only explore possible violations of NWW.
As much as we might like to do so, we can't speak to all issues at once. John Reid 07:21, 20 May 2006 (UTC)

Ambiguity in Examples vs. Policy

They contradict. "An admin action" would seem to mean the act, not the act only by that admin. If A blocks X and B unblocks X then C reblocks X, C knows that B objects to the block and yet he blocks anyways and would be in violation. In the examples, he is not. If the examples are correct it should probably say "...repeat any of their admin actions..." or something of the like (which still leaves the absurb posibility of 900 total blocks and unblocks without a violation, but I'm sure people know that). Kotepho 02:23, 9 May 2006 (UTC)

I understand your point. I say that C's knowledge of B's opposition is irrelevant on the occasion of C's first action. It is only when C repeats the action that NWW is violated. You say that C's first action is a repetition of A's action. Correct?
I think we both understand exactly what is intended but you may have found room for improvement in the text. Edit this page.
(I agree that the text does nothing at all to prevent slow-motion, whole-community wheel wars. I just find it hard to believe that any such can be intelligently prohibited -- nor do I fear them. Sooner, rather than later, matters will come to a head. As admin after admin weighs in on an issue and casts his single stone, the most heavily biased editors will be most eager to take action. In time, only more moderate admins will be left; eventually the pool of uncomitted admins will be sufficiently neutral to reach an amicable compromise. Either that, or one side will so outnumber the other that consensus can be shown to have emerged. Further admin actions against that consensus may not violate NWW but they are subject to other policies. It's the cheapest way of resolving disputes.) John Reid 06:57, 20 May 2006 (UTC)

An issue

For the last 45 days or so, I've not been very active, so I don't know if this has been discussed before. Phil Welch has blocked a user and this was undone by other admins after a discussion on WP:AN just around 10-15 minutes. I believe that undoing an admin action, specifically something like removing a block, should not be considered as wheelwarring if the original block had to do something with the interaction that the admin has with the user. Thanks. --Gurubrahma 11:09, 15 May 2006 (UTC)

From what I understand of this proposed policy (one I haven't been following closely) it would only be wheel warring if Phil Welch re-applied to block. To quote the summary box:
No admin shall repeat any admin action in the knowledge that another admin opposes it, except with community consensus
I think that makes it pretty clear, and sounds sensible IMHO- i.e. it doesn't stop an admin ever reverting another admin, but does stop them getting into a pointless and damaging revert war. Petros471 11:18, 15 May 2006 (UTC)
This is absurd. It's the first admin who undoes the action who has started the wheel war. I'm going to remove that summary. SlimVirgin (talk) 08:22, 20 May 2006 (UTC)
Surely we all realize that different folks have different opinions on this? Many of us (myself included) do not feel than undoing something once fits any useful definition of "wheel war". I think it's better all around to invite others to revert your action if they feel strongly enough to do so. It's the admins who continue to use brute force after being reverted who are the wheel warriors. Too often I've seen people say, "Well, I disagree- but an admin action was performed, so we can't undo it." That's just silly- what's easily done can be easily undone. If we had infallible admins, I might feel differently, but I think we're running low on them (certain admins own opinions of themselves notwithstanding :-) Friday (talk) 18:34, 22 May 2006 (UTC)
Yes, but that is how they start. No one thinks they can't be undone at all, but they should never be undone without discussion and a consensus. If we followed that extremely simple rule, we'd have no wheel wars. Any admin action that isn't right will be quickly overturned and any that are fine will stay. There just simply aren't many things that can't wait until a discussion develops to decide what to do. In this specific case, according to Gurubrahma's summary there seems to have been discussion and an agreement to undo the block. That is how it should be done. - Taxman Talk 19:59, 22 May 2006 (UTC)

Community Consensus

I rm except with community consensus from the policy text. (This was inserted with edit sum Modify summary to match text.) I'm fairly unhappy about it -- I believe in leaving new policy open to editing in general and I don't like to seem the sole defender of my own draft. Please allow me to discuss this edit.

First, the edit sum itself. The policy is deliberately worded on one line. That's not a summary of the policy: that's the text. Everything else on the page interprets or illuminates this; nothing else has the same force as the text. Note the refactored prior discussions under Commentary: some contradict one another but they all illuminate our joint state of mind as we refine this policy. None have the force of policy itself. I realize that other policies are extremely wordy; lately the fashion has been to head some policies with a summary titled " a nutshell". This is not the case here -- and indeed, I think these policies are not particularly well worded. The more words a law contains, the more crevices left to shelter lawyers.

Community consensus is, of course, our highest organizational principle. It trumps all other policies at all times. But it is rather risky to make explicit exceptions to lesser policies. It is all too easy for a malefactor to claim But I had community consensus behind my actions! Consensus is just one of those things, like good music: All know it when they see it but everyone has a different opinion of what it is.

I suggest that the desire to leave loopholes like this stems from the all-too-prevalent feeling that if I don't fix this, it won't get fixed. Unfortunately this is exactly the good intention that leads to bitter wheel wars.

I do think this clause is a pretty big risk -- but if the community insists on it, I'll not demand its removal. I'm just me. John Reid 06:45, 20 May 2006 (UTC)


This is one of the worst recipes for wikilawyering I've seen, because it's so complex. I've removed the policy summary because it's not clear what it even means, and you seem to be trying to come up with a novel definition of wheel war, whereby the admin who starts is may do so, but any subsequent admin is only allowed to undo once? Is that it?

It's particularly important that the wording be simple and clear (and fair) if people might be penalized. SlimVirgin (talk) 08:25, 20 May 2006 (UTC)

Could someone please tell me what exactly is wrong with what the current blocking policy says, viz. do not undo another admin's actions (the exception being if there has been an unambiguous error and the blocking admin is not available). What could be simpler than that? SlimVirgin (talk) 08:27, 20 May 2006 (UTC)
I've reworded the intro in an effort to keep it simpler, and to make it consistent with the blocking policy and protection policy, but the other sections are a nightmare. SlimVirgin (talk) 08:47, 20 May 2006 (UTC)
Well, I thought it was rather complicated myself. "No admin shall" blah blah was too hard to understand at first glance. But I guess that's why admins have to go through RfA and everything. --M1ss1ontomars2k4 (T | C | @) 05:28, 26 May 2006 (UTC)

Wow, too obvious.

The temptation to move this to Proposed wheel warring policy ON WHEELS!!! is so strong! -- Saaber 07:28, 26 May 2006 (UTC)


Slim Virgin deleted the entire policy (!) as shown above with edit sum I strongly disagree with that summary; see talk. I really think I've done enough here and it's somebody else's turn -- but I'd suggest that we either reinsert the policy statement into the policy page or insert a new policy statement.

As I've said before, the boxed text is intended to be the policy, not a summary of that policy. Only this text is intended to have the force of policy; all else is explanation and commentary, with lesser force. Without any actual policy statement, this page now looks to me like a wheel without an axle upon which to turn. John Reid 22:36, 8 June 2006 (UTC)

I'll have a go at it, but this page is mostly dead. =/ Kotepho 22:51, 8 June 2006 (UTC)

Oh, I don't think so. The discussions that led up to this point were lively and drew wide participation. This is a wiki and details are always subject to change but I see broad support. Thanks for helping; I've boxed the policy statement again and fixed a minor typo. John Reid 01:45, 9 June 2006 (UTC)

"Aggravation, mitigation, and defense"

The scope of an admin's action is a critical factor in our evaluation of a charge of wheel warring. Factors in aggravation or mitigation may include, but are not limited to:

  • Time: Actions which are repeated within minutes or hours are more serious than those repeated within weeks; actions repeated within months or years may not be actual violations at all.
If it was years ago, do you even know that it still opposed? Kotepho 23:49, 8 June 2006 (UTC)
  • Range: Actions that affect large numbers of pages, highly prominent pages, or heavily-used Template or MediaWiki namespace pages are more important than those that involve a single obscure page. Multiple cotemporal violations of NWW aggravate one another.
Duh? Kotepho 23:49, 8 June 2006 (UTC)
  • Intent: Actions that clearly demonstrate disregard for community policy are more serious than those taken in clear good faith. An allegation of bad faith intent does not establish a violation of NWW; but an established violation is aggravated by clear demonstration of bad faith. A clear demonstration of good faith may be sufficient to exempt an admin from sanctions.
Again, mostly "common sense" and does not help clarify things. Kotepho 23:49, 8 June 2006 (UTC)
  • Talk: Actions performed subsequent to civil discussion on talk are presumed to be more amicable than those without. Neutral, honest, polite edit summaries and discussion prior to action go far to mitigate violations of NWW. Hostile talk comments and edit summaries (or none) aggravate.
It doesn't matter if you are talking civilly while blocking and unblocking constantly, it is still bad. Not sure what this does to further the policy. Kotepho 23:49, 8 June 2006 (UTC)
  • Coviolation: Violating another policy in the same action aggravates a violation of NWW. The other violation must be established as if there were no violation of NWW; only then can it be considered in aggravation.
Again, duh and unhelpful. Kotepho 23:49, 8 June 2006 (UTC)
  • Repeat offender: Having violated NWW in the past does not establish that an admin has violated it in the present. But if a current violation is established then a prior violation is an aggravating circumstance.
Ditto. Kotepho 23:49, 8 June 2006 (UTC)

The context of an admins action may influence the finding of wheel warring. Few defenses are able to succeed. Possible defenses include:

  • Single action: It is quite possible for any single action to violate some other policy but this is a certain defense against NWW. Violation requires repetition.
True, but isn't this obvious? If it isn't, shouldn't we just make the policy part more obvious than adding this in this "mitigating circumstances" or whatever it is section. Kotepho 23:49, 8 June 2006 (UTC)
  • Necessity: The defense may be made that it was necessary for an admin to violate NWW in order to uphold some other policy or vital principle. Such argument may be considered in mitigation with the caution that it is extremely unlikely to be sufficient defense. Admins are advised to enlist the support of other admins in upholding policy.
Not sure what the point of this is. Kotepho 23:49, 8 June 2006 (UTC)
  • Ignorance: The defense may be made that an admin was not aware of admin opposition before repeating an action. However, we presume that an admin is aware that another admin has reversed an action before he repeats it. Admins who repeat a block should know if the previous block has expired or if another admin has unblocked. Admins do have a positive responsibility to review comments made on their talk pages.
Same as "single action:" above. Kotepho 23:49, 8 June 2006 (UTC)
  • Amity: The defense may be made that an admin repeated an action after coming to an amicable agreement with all admins who initially reversed it. This is a good defense against a charge levied by an admin who did not participate in discussion. If the charge is brought by an admin who did participate in discussion, then the contents of the discussion must be examined for evidence of amicable agreement, absent which the defense of amity fails.
More duh. If you say "oh, sorry, I thought we worked this out." and continue discussion all is roses. Enough with the "defense against a charge levied" stuff, it is too legalistic. Kotepho 23:49, 8 June 2006 (UTC)


  • Move this page to Wikipedia:Wheel war policy and change tag to {{guideline}}. This now has wide acceptance. While there may still be room for improvement in the details it's clear that nobody endorses wheel warring.

On a personal note, I'd like to thank the many editors who contributed to this proposal. This was not created overnight nor was it the pet project of one or two editors. Much was written, much was debated, much was hacked away, and much rewritten. The people who contributed here did not all agree on every point but in the end we all managed to hammer out something we can live with.

I believe an example has been set here that we would do well to follow in our dealings with other troublesome issues. John Reid 03:54, 9 July 2006 (UTC)


I've moved this page to "Wheel wars". Having a page with "policy" in the title when it's a {{guideline}}, not a {{policy}}, is somewhat absurd. --Sam Blanning(talk) 14:39, 19 July 2006 (UTC)

Oh dear, turns out we've already got a page called Wikipedia:Wheel war, so that's too confusing. I'll move to Wikipedia:Wheel war guideline... and I just had to realise that only when I was finished doing the double redirects >_< --Sam Blanning(talk) 14:40, 19 July 2006 (UTC)


Yes, Sam, neither you nor I quite have the right grip on this. What's really needed now is a merge with Wikipedia:Wheel war. The reason a second page on the same topic was started is that the first page was general and descriptive, not specific and prohibitive. Several editors thought that prohibition needed to be discussed in detail but they didn't want to disturb the existing (and well accepted) page. The process leading up to this (second) page was indeed lengthy and not without sweat and backtracking.
Now that both pages are in a reasonably settled state, they should simply be merged. Wheel warring is not a topic requiring multiple pages. John Reid 19:03, 7 August 2006 (UTC)
I'm not sure that the two pages are enough in agreement to be merged. For instance, this one seems to forbid individual admins 'repeating' an action... but would apparently leave open the possibility of various admins reversing each other so long as each does so only once (as discussed in some of the quoted comments). The other page says not to 'revert' an admin action (even once) until consensus is achieved. Likewise, the other page warns against taking a first admin action if there is an ongoing discussion where that action has already been disputed/no consensus yet reached... while this page deals only with subsequent actions and would attach no negative implications to such a 'first strike'. Et cetera. Parts of the two pages agree, but there are alot of details 'around the edges' where they are quite different. Mind you, I agree that we should have just one page... but I think there are still some differing views on this issue that need to be sorted out. --CBD 22:21, 7 August 2006 (UTC)
Nothing apparent about it. The "slow-motion" WW in which one admin at a time reverts an action once and once only is definitely not forbidden. It may be in poor taste but it doesn't violate policy. The theory is that the most extreme partisans will weigh in first; eventually all admins who are unable to compromise will eliminate themselves; the pool of those who remain will become steadily more moderate and a compromise will be reached -- a compromise acceptable to all but the extremists.
You're right; the two pages do not agree. Merge is not a slam dunk. I think the essential insight is that there are many things that might steer close to 0WW and should be avoided; then there are clear violations which must. The key to the merge is to define both the forbidden and the discouraged; and to distinguish them. John Reid 09:00, 27 August 2006 (UTC)

Slow motion WW

See, I don't like the 'it isn't a wheel war if each admin only reverts once' form at all. It seems to me to allow all the negative elements of wheel-warring to take place and only stops when one 'side' runs out of people willing to 'edit war' rather than discussing. On the big conflagrations which led to this policy/guideline effort you'd have twenty admins on either side... and some of those who 'only acted once' were desysopped. I much prefer treating it like any other dispute... if people disagree talk about it and get consensus before taking another action. As to a merge... well both pages agree that multiple reverts are bad, so as you say we could have the page cover the common ground and then list the different perspectives on the less obvious cases. --CBD 16:06, 5 September 2006 (UTC)

This policy-to-be enjoys broad support. It doesn't answer every need nor does it meet everyone's every expectation. Your concern, CBD, was aired at length during the process that led up to the current version. It's legitimate.

I think if you read both "front" and talk sides to this page -- and especially if you review history of both -- you'll see the rationale for permitting the slow-motion, one-shot-per-admin WW. That such a thing is a WW, there is no doubt; that it is highly discouraged, also no doubt. Perhaps ArbCom will even sanction somebody for participating in such a thing; I rather hope the effect of this policy will be to moderate such penalties. That's really neither here nor there. WP:WHEEL has a much looser definition of WW and that's a good thing. Admins should avoid anything that even looks like WW but it's hard to draw a bright line there; this is the function of the words at WHEEL. Absent the bright line, reasonable men might argue on either side when some admin is accused of WW. This page has a very strict definition that everybody can agree on. You may well be WW before you cross this line; you certainly are, once over it.

The trouble with forbidding the slo-mo WW is when somebody asks the question, When is it a WW? If you can answer this question quickly and confidently without using loaded terms like "vandalism" and "good faith", then you are an involved party to a potential WW.

Let's say -- in neutral terms -- that I, Zeus, from my remote and entirely neutral mountaintop observe an editor blocked by Admin A. Is that a WW? No. The disagreement, whatever it is, is between an admin and a non-admin; it's not a WW. Okay, well, the block is lifted by Admin B. Is that a WW? Probably not. It may not even be a difference of opinion, depending on the length of the block, the time elapsed, and whether the blocked editor made a good plea for unblock. Perhaps Admin A would have been happy to lift the block if he'd seen the plea but he's out to lunch. Now, what if Admin A talks to Admin B and restores the block with B's agreement? That's not a WW, is it? No. B no longer opposes. If A fails to talk with B that may still not be a WW; perhaps circumstances have changed. Perhaps the editor in question made a fool of B in the meantime, giving the lie to his apparently good plea but now B is out to lunch. Admin A reblocks and B is content; it's not a WW.

On the other hand, perhaps A and B are known enemies. Let's say they're involved in a prior edit war over Chile or Chili; they have, of course, steered close to 3RR but wisely never overstepped that bound. Coldly civil, contentious words have been exchanged on talk. Now A blocks an editor who gets into the edit war on B's side. B unblocks. WW? Probably so but how do you prove it? Where do you draw the line? A and B aren't blood brothers? A and B have both edited the same article? They have occasionally left messages on each other's talk? If they've left rude messages we don't need to look for a WW; we should sanction them immediately for violation of WP:CIVIL. If not, it's because such messages are polite enough -- therefore, how can you say that either A or B did not act in good faith? The standard is plain and easy to apply: If A knows that B opposes and reblocks, it's a WW. Note that the "reasonable man" principle of common law applies here; Admin A cannot claim ignorance if a good showing can be made that a reasonable admin should have known that B opposed.

Well, why isn't B WW if he knows that A opposes the unblock? Well, we have explicit language for this, too: Do not repeat ... Since B has never yet acted, his action -- however in-your-face it may be -- cannot, when it comes, be WW. It may be hostile, uncivil, unwise, or a clear breach of community trust. But these are all subjective evaluations and don't lend themselves to bright-line distinctions. We here on this page permit B to act once. This gets us out of the deadly trap of endorsing whoever shoots first. Otherwise, A could do any damn thing, announce loudly that he opposed any counteraction, and "own" the article or user until doomsday.

So, if Admin A is wrong to reblock knowing B is in opposition, why do we permit C to do the same thing? Well, first of all, perhaps we don't. Perhaps we sanction C for being hostile, uncivil, unwise, or somehow violating his trust. But if he can make some plausible argument that A is right and B is wrong, then he is entitled to act -- whether you or I agree. By "entitled" I only mean to say that he has not violated 0WW. Who knows what else he's done. It should never be the purpose of a new policy to restate all the other policies that might apply to a given situation. Avoiding sanction under 0WW does not guarantee C a free pass under any other policy including WHEEL. 0WW does raise the bar slightly for the person who wishes to attack C; it's just not enough to yell "He reverted/reblocked/reprotected -- he's WW". So you see, by giving C his one shot, we avoid the second deadly trap of endorsing whoever shoots second.

In similar vein, we permit D to reunblock -- but not B. We avoid endorsing whoever shot first and whoever shot the third time, confirming the first guy. After all, who is "right"? We can't say just by examining the block log. Admin A may have chosen wisely and B may have done so; both are probably a little right and both probably a good bit in the wrong. But we avoid endorsing either party as we would if we attempted to call a halt to the slo-mo WW. If we did yell stop at some point, we would freeze the situation in favor of one side or the other -- and at any point far from the end, a fanatic is in charge of each side. We have cut short some unpleasantness and disruption but paid the price by setting a biased point of view in stone.

Instead, the WW plays itself out. Yes, 20 admins on each side may hit the ball one time each and everybody watches the ball go back and forth. I'll bet that in the course of the slo-WW most of the participants will have violated a small handful of other policies -- and we get to sanction every one of them for that -- and we should. If by some small miracle, all 40 admins manage to stagger through the process without (a) ad hominem attack, (b) failure to assume good faith, (c) failure to discuss contentious changes on talk, (d), (e), (f), and (g), well then, what harm has been done? Not much -- certainly not as much as if the same crew were to squabble over which side crossed the line first. The latter scenario boots the issue up one notch on the metawar ladder: always an evil thing.

Sooner or later, one side or the other will have exhausted all those willing to make complete jackasses of themselves and risk sanctions due to violations of (a) through (z) other policies -- everyone completely unwilling to compromise anyway, even if you chained them to the table. Now moderates -- still holding strong opinions but not fanatics -- come in and do their bit -- but with a smidgen of real compromise. The trick is that we have got all the fanatics away from the table. With them gone, the moderates who remain will work something out; in the very worst case, the last man standing will decide. And it will still be over and every admin will have had a whack at the ball.

I guarantee to any reader that if you do not see the justice in this, it is because you are thinking of a particular WW, one in which you hold an opinion as to which side was "right". You want the 0WW policy to somehow be written so that, if it had been in effect back then, "your" side would have been granted the victory. Instead, 0WW takes the attitude that all WW is wrong but there is no way to call an arbitrary end to a slo-mo war without taking sides. The last thing we want to do is create a system for somebody to game; so we never take a side in a WW. By insisting that every editor has one shot only, we ensure its swiftest possible neutral end. John Reid 00:49, 10 October 2006 (UTC)