# How to handle crankery on MathOverflow 2.0?

On the previous MO we had a thread on Tea called "crank post to flag as spam" in which crank posts were pointed out so people could flag them as spam and have them deleted.

Seeing how we are now using a new and improved system, where users can vote to delete posts (both in the review system and directly, if their score is low enough), do we want to continue with the previous approach?

There are arguments in favor of this:

1. Crank mathematics is in fact the spam of mathematics, and often these posts contain links to viXra papers which make them actual spam.
2. There are only 23 non-diamond users which can vote to delete answers at this time.
3. Spam flags come with the penalty of 100 points of reputation (mostly unimportant because most cranks come as "fresh users", but in some cases they do manage to gain some points before their cranky posts).

There are arguments against this:

1. The old system was ill-equipped when it comes to deleting posts, but this system is much better and improved.
2. Spam is spam is spam. Crank posts against Cantorian set theory, or Cauchy's theorem are not spam.
3. Spam flags do not appear to non-diamond users. Low-quality flags, on the other hand, send a post right into the review queue where anyone who can review can vote for its deletion. (This gives a vastly larger user basis for deletions, but requires people to check the review queues often.)
• With great vengeance and furious anger, as Jules Winnfield once said – Yemon Choi Jun 25 '13 at 10:53
• @Yemon: Well, they do attempt to poison and destroy my brothers! – Asaf Karagila Jun 25 '13 at 11:56

## 4 Answers

I actually think that the "low quality" flag makes the most sense. I just reviewed and voted to delete a post marked low quality, and it was easy and fun. It's also good in that I think many people were hesitant to flag things as spam that weren't traditional spam, but there should be no disagreement about voting to delete things that are "low quality".

UPDATE : To give folks an opportunity to test-drive the "low-quality" system, I have flagged the following bit of spam as low-quality : link deleted

• Andy, isn't that a repeat offender plugging his vixra paper? I recognize the style – Yemon Choi Jun 25 '13 at 20:14
• @YemonChoi : Sure is! Let's vote it into oblivion. – Andy Putman Jun 25 '13 at 20:15
• The problem with LQ flags deletions is that they are only accessible from the review system which means that first we need to clear out the gigantic queue there for this to be an effective mean to fight cranks. – Asaf Karagila Jun 25 '13 at 20:47
• @AsafKaragila : Huh, that's weird. My low-quality queue is empty; it had only a few items in it after the migration, and I dealt with them pretty quickly. Is there something wrong? – Andy Putman Jun 25 '13 at 20:49
• Oh right right, LQ queue is empty. I'm just distracted by the other numbers! :-) – Asaf Karagila Jun 25 '13 at 20:52
• Actually, this is important -- did the post above end up in the LQ queue? If not, then I'm not sure what went wrong. – Andy Putman Jun 25 '13 at 20:53
• It's hard to tell. I didn't see it there (I flagged it as spam (because it contained that shameless link to vixra) and it got deleted, though). If you have enough reputation then you should be able to see other people's action on the review queue's history. I don't have enough reputation here and I don't know how much is "enough", probably 10k. – Asaf Karagila Jun 25 '13 at 21:04
• Hmm...I didn't see in in the review queue's history. Maybe it just got deleted too quickly. – Andy Putman Jun 25 '13 at 21:05
• No, the history should contain all the actions anyone did. I know that on MSE I can see the full history. Perhaps "enough" is 20k, then. Can you get 2.5k reputation really quick? :-) – Asaf Karagila Jun 25 '13 at 21:12
• Hmmm...My LQ review queue's history is very small (it only contains stuff about mathoverflow.net/questions/53669/…). Does anyone else see other things in it? – Andy Putman Jun 25 '13 at 21:26
• Do you see only your reviews, or do you see mine as well? – Asaf Karagila Jun 25 '13 at 21:43
• @AsafKaragila : There are 6 people listed as voting to delete that answer, including you. – Andy Putman Jun 25 '13 at 21:45
• Okay then. I suppose this was the only thing that got to the LQ queue. Which means that it works slightly differently than I have expected (if you flagged the post in your answer as LQ, it certainly didn't get there). Perhaps it's time to make a feature request... – Asaf Karagila Jun 25 '13 at 21:50
• The text for spam flag reads as follows: "This question is effectively an advertisement with no disclosure. It is not useful or relevant, but promotional." It seems to me it applies to posts about crank preprints, doesn't it? – Kaveh Jun 26 '13 at 3:01
• @AsafKaragila iirc the VLQ queue is based on an algorithm. – Manishearth Jun 26 '13 at 9:04

In the past few days some well known crank has been posting a lot of junk, in particular some of it seem to be some sort of an attack on one of the most established users in this community.

I feel strongly that we don't need to allow these posts to exist any longer than the minimal time it takes to delete them. But that would require the posts to appear on the review queue immediately after flagging them. Until something like that is implemented (see Flagging as Low Quality should add the post to the review queue), I feel that we need to resort to spam/offensive flags again.

• Larger issues with a specific user can be dealt with more effectively by moderators, the community can only act on a per-post level. Moderators have much more powerful tools available for such situations than regular users. – user35354 Jul 12 '13 at 12:37
• Even when they are unregistered and keep creating new users? – Asaf Karagila Jul 12 '13 at 12:53
• Moderators have some tools that make it easier to identify if they are the same user. And they need only a few clicks to either suspend or nuke the whole users and all posts belonging to it. – user35354 Jul 12 '13 at 12:58
• I see. In this case it's completely unneeded to use any special moderator tools to know it's the same person... – Asaf Karagila Jul 12 '13 at 13:00
• @AsafKaragila Sure, however mods have tools that help find new iterations of that person quickly. Ask them to keep an eye on it, they'll be able to delete stuff quite easily. And keep flagging. – Manishearth Jul 12 '13 at 14:04
• Asaf, are there currently any examples that have not yet been deleted? – Andrés E. Caicedo Jul 12 '13 at 16:07
• @Andres: I flagged four posts today, and all have been deleted at this point. You should be able to see those deletions on the moderator tools, I believe. – Asaf Karagila Jul 12 '13 at 16:14
• OK, they are already deleted. (Hmm... No, none of them appear on the list.) – Andrés E. Caicedo Jul 12 '13 at 16:18
• (Ah, some sleuthing revealed hypothetical history, funny non-well-orderings, and the like.) – Andrés E. Caicedo Jul 12 '13 at 16:25
• @Andres: I only recalled a few about supertasks and what seems like thinly veiled insults towards Joel, and a couple about some permutations of the rationals... Were there more? – Asaf Karagila Jul 12 '13 at 16:35
• Those are the ones I found. Pointless and boring. – Andrés E. Caicedo Jul 12 '13 at 16:38
• And back. – Andrés E. Caicedo Jul 12 '13 at 16:54
• Someone just upvoted the "question" linked to in my previous comment. I hope that's just a sock-puppet, alternatives are too depressing – Yemon Choi Jul 12 '13 at 18:34
• @MadScientist, Kaveh: The problem here is that we are not dealing with spammers per se. We're dealing with trolls. – Asaf Karagila Jul 13 '13 at 8:34
• @Kaveh: Well, this particular case is for cranks which are also trolls. But yes, the general thread is about cranks. I don't think that anyone that has spent more than two years answering questions on math websites can mistake the meaning of "cranks", and no one that has spent any period of time moderator a community can mistake the meaning of "trolls". We both did these things, so we both must be on agreement on these terms. – Asaf Karagila Jul 14 '13 at 7:51

If you can get and maintain a chat room going (e.g. here), you can simply link to the post there for your fellow users to review and flag. It's pretty effective.

• According to a recent similar thread on meta.MSE, non-diamond user are no longer privy to spam/offensive flags. – Asaf Karagila Jun 25 '13 at 13:27
• @Asaf Aw, that sucks. – badp Jun 25 '13 at 13:28
• I know!!${}{}{}$ – Asaf Karagila Jun 25 '13 at 13:28
• See this, and in particular the first comment by mixedmath on the main post. – Asaf Karagila Jun 25 '13 at 13:31
• @AsafKaragila Well this question is about how to broadcast posts for others to review. I assume you're going to be able to apply the proper policies on a per-case basis :) – badp Jun 25 '13 at 13:34

On Physics.SE and Chemistry.SE we have a policy where non-mainstream science is not allowed. Non-mainstream is defined by "Anything that couldn't be published in a reputable journal". So a question/answer that proposes a new, non-mainstream theory is not allowed. We get a few of these every ~week (sometimes with a persistent user who eventually gets banned)

This is reflected in our help center page:

• Pitches for your own personal theories or work

We deal with mainstream physics here. Anything that couldn't be published in a reputable journal is not appropriate on this site.

(There's also a "Pitches for your own personal theories or work" bullet point under the "Don't ask about" section of our about).

I suggest you edit these lines in: while it may not be a good deterrent, it's useful to have a help center page to back you up when people complain.

On Physics, most users flag crank answers as very low quality or use a custom moderator flag mentioning "non mainstream theory (blah blah)". I personally feel that the custom flag is better, since it tells moderatros exactly what the problem is. (crank theories can sometimes be well-disguised, and in that case the post certainly doesn't look "very low quality"). However, if you wish to lessen the burden on mods, use the VLQ flag, as it pushes the post to the 10k user flag queue. Additionally you can post it in chat and solicit spam flags -- 6 spam flags leads to deletion.

For questions, we have a custom off topic reason "We deal with mainstream physics here. Anything that couldn't be published in a reputable journal is not appropriate on this site.". People can flag questions with the same reason. These will get pushed to the close review queue and can be easily dealt with by the community.

• I don't feel that the moderators should carry the burden of deleting all the crankery. This is something easily done by a community with a large enough user basis. In particular since the moderation on MO is "less subject" to meddling by yearly elections. – Asaf Karagila Jun 26 '13 at 10:09
• @AsafKaragila hmm. How much crankery do you get? 10k users can see VLQ flags. – Manishearth Jun 26 '13 at 10:10
• I don't know if a whole lot. But more than "rarely". I also don't think that the 10k user base is sufficiently large for this. (For example, I'm not a part of it! ;-)) – Asaf Karagila Jun 26 '13 at 10:12
• @AsafKaragila Well, aside from 10k deletion, how will you get rid of such posts? Spam flags? (also, a meta post urging the 10k community to take an active part in community moderation would be good) – Manishearth Jun 26 '13 at 10:14
• No, users with 5k or so can "recommend deletion" on LQ reviews. If the LQ flags send posts into the review queue, that would be sufficient and much better. – Asaf Karagila Jun 26 '13 at 10:15
• @AsafKaragila Right, except that "recommend deletion" does nothing but make the post a priority for 10k users (i.e., a 10k user visiting the queue sees these first). Same problem, then – Manishearth Jun 26 '13 at 10:18
• No. It does delete the post if six users recommend its deletion. You're a moderator, get your facts straight! :-) – Asaf Karagila Jun 26 '13 at 10:20
• @AsafKaragila huh, that's new – Manishearth Jun 26 '13 at 10:21
• No, that's how it always been. It's even more interesting to see that the deletion is "unsigned" there's no trace to who were the users who recommended deletion, unless you dig out the review report. – Asaf Karagila Jun 26 '13 at 10:21
• @AsafKaragila I'm quite sure that wasn't the case initially :P Also, as a moderator on sites where community-deletions rarely happen, I've never seen that happen. – Manishearth Jun 26 '13 at 10:21
• That is how things happened on MSE ever since "Recommend deletion" was implemented. Trust me, I've been following that system for a while now. – Asaf Karagila Jun 26 '13 at 10:22
• @AsafKaragila Oh well. Better update my meta-faq post on this then :) – Manishearth Jun 26 '13 at 10:22
• (And ask a second one for piping VLQ to the queue) – Manishearth Jun 26 '13 at 10:23
• I already made a feature request on the second one. – Asaf Karagila Jun 26 '13 at 10:23
• Also, if you look at the comments on Andy's answer on this page you'll find a link to a review where six users recommend deletion and the post is gone. – Asaf Karagila Jun 26 '13 at 10:24