Normally new users are not allowed to up-vote until they gain a bit of reputation by receiving up-votes or accepts from the community of a site. But this has a weak point that is being exploited:

A number of new users post, answer, and accept each others' off-topic questions before we can close them (in less than 5 min). Accepting each others' answers gains them enough reputation to start up-vote each others' posts.

We are closing them and the moderators clean up afterwards but this keeps happening over and over again and becoming frustrating to the extent that I think a systematic solution might be required. Is there a system to prevent this kind of circumvention of the reputation requirements using accepts by new users?

One possible solution might be to change the system so accepts from users who do not have enough reputation to up-vote do not give any reputation (at least until the answer gets an up-vote).

  • 58
    If you're seeing this on CS or CS Theory, I believe those sites have been subject to a coordinated attack by students from an Indian university. They have been posting terrible questions that are immediately answered by other students (usually using plagiarized content) and then they are voting in a tight ring, inflating posts by 8 or more votes. They just tried to pull this on Stack Overflow, and they'll find they no longer have accounts there. New accounts from that university should now be blocked from posting for a while. I've informed the other moderators. Commented Apr 28, 2016 at 19:18
  • 16
    I temporarily blocked new-user questions from a bunch of vodafone IPs, which should slow this down a bit also.
    – Shog9
    Commented Apr 28, 2016 at 19:23
  • 6
    I always downvote all the things and VTC. Sometime after closed and negative-scored, you can vote to delete.
    – user1228
    Commented Apr 28, 2016 at 19:28
  • @Shog9 I'm just curious - Was this question is one of those idiots?
    – Alon Eitan
    Commented Apr 28, 2016 at 19:30
  • 6
    @Won't, I do that and also vote to delete, but I can vote to delete only after a few days I think, and even then since there is no delete review queue other users who can vote to delete do not see them, so it is really left to moderators to clean up these. If it was a just a few it wouldn't be a problem but 30 questions in an hour is too much to handle manually I think.
    – Kaveh
    Commented Apr 28, 2016 at 19:30
  • 7
    Not that I'm aware of, @Alon. India does not possess a monopoly on bad questions.
    – Shog9
    Commented Apr 28, 2016 at 19:31
  • 8
    Please downvote and flag. The downvotes help with the automatic post blocks, and we'll delete all the plagiarized or extremely low quality content on sight. Commented Apr 28, 2016 at 19:37
  • 2
    @Kaveh holy hell.
    – user1228
    Commented Apr 28, 2016 at 20:06
  • 6
    @AlonEitan - Sorry, my amused tone may not have been conveyed there. Those users are part of this group, and the answers appear to be largely plagiarized (thus the second link in my comment). Same self-answering and rampant plagiarism as the rest, and they're all coming from the same university. What a mess. Commented Apr 28, 2016 at 20:19
  • 2
    related: Sudden increase in off-topic posts on MSE - example showing that attacks of certain scale simply can't be adequately handled by community moderation at smaller sites
    – gnat
    Commented Apr 28, 2016 at 20:25
  • 8
    Not sure there's anything magical about hitting delete a couple dozen times, but... Done, @gnat.
    – Shog9
    Commented Apr 28, 2016 at 20:39
  • 7
    @Shog9 I wish there was something magical, hitting delete a couple dozen times is boring. Commented Apr 28, 2016 at 21:08
  • 1
    @Gilles at Stack Overflow triage does this kind of magic
    – gnat
    Commented Apr 28, 2016 at 21:36
  • 35
    @sho9 I think what they should do make it so that whenever a post is closed for being off-topic and the owner has less than 50 reputation, the accepted answer is automatically unaccepted and the user cannot reaccept it. Basically, what I mean is making "accepting answers to closed questions that are closed for being off-topic" a privilege awarded at 50 or so reputation. Commented Apr 29, 2016 at 1:57
  • 3
    @John, can you rewrite that as a full answer? This seems (and 26 people agree with me at time of writing) to be a good solution.
    – AAM111
    Commented May 5, 2016 at 23:09

3 Answers 3


I think what they should do is make it so that whenever a post is closed and the owner has less than 50 reputation, the accepted answer is automatically unaccepted and the asker cannot reaccept it. Basically, what I mean is making "having accepted answers to closed questions that are closed" a privilege awarded at 50 or so reputation.

  • 9
    Why make this restriction only for "off-topic" and not "primarily opinion-based", "too broad" and "unclear"?
    – user642796
    Commented May 6, 2016 at 3:58
  • 1
    @arjafi I used to think it should only apply to questions closed for being off-topic, but I no longer think that. Commented Jun 2, 2016 at 0:29
  • 3
    Not just the accept vote, but also OP's upvotes on any answers (so closed, off-topic questions can be deleted more easily; effectively "unasked"). In parallel, for similar reasons, a priv at 50 (or whatever threshold) to "upvote questions of 1-rep users". On EL&U, we get a large population of ESL students, effectively as silent lurkers who upvote off-topic questions from 1-rep users (i.e. newbies unaware of ELL.se), as well as any (still off-topic) answers they attract. Which attracts even more of the same. Quite frustrating.
    – Dan Bron
    Commented Jun 12, 2016 at 22:45
  • 2
    @Dan Bron Yes, it should also apply to upvotes. :) Commented Jun 12, 2016 at 22:56

I do think all users should be delayed from accepting an answer for 15 minutes. IIRC there is already a day delay for self-answered accepts.

Perhaps after a certain reputation level it does become only a warning allowing really straightforward answers to be accepted after confirmation.


There are two moderation tools which are quite close to this task.

First is the suspicious-votes page. http://example.com/admin/show-suspicious-votes. It already shows when there's too many votes that a user had given to or received from another user.

Second is the show-user-votes page. http://example.com/admin/show-user-votes/123456. For a particular user it shows percentage of up/down/accept votes from and to other users. Suspicious mass-accepts can already be detected here.

What needs to be done: suspicious cases from show-user-votes page for each user should appear on the suspicious-votes page (or a similar one). There they would have to be manually revised by moderators. The reason is that not all cases of "mass-accepting" or even mutual "mass-accepting" are bad. E.g. 30% of my questions were answered by a single user and that's ok.

Possible signs for a suspicious mass-accepting:

  • One of the following:

    • Most questions of user A have accepted answers given by user B (with total questions above 5)
    • Most answers of user B are given to questions of A and accepted (with total answers above 5)
  • Accepting user or both have low reputation
  • Extremely suspicious when they share the same IP.
  • Accepted answer was given within a short time after the question was asked
  • Question has received close votes
  • 3
    I appreciate that you took the time to answer. However, as one of the moderators of CS.SE (one of the sites that was attacked), I'm afraid it's not that simple. Yes, we were using all of the tools available to us, but for various reasons some of them were not as effective as one would hope in this particular attack. I don't want to get into the details here, to avoid helping others try similar shenanigans, but I think it's worth pointing out that Kaveh, who asked the question, is a former moderator of CSTheory.SE; I'm sure Kaveh knows about the existing tools.
    – D.W.
    Commented May 7, 2016 at 19:16
  • If you're curious to know more about the methods that the attackers were using in this case, ping me in a private chatroom. I'd rather not discuss them publicly.
    – D.W.
    Commented May 7, 2016 at 19:16

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .