Users have found a way to circumvent spam filters and deceive reviewers with less-than-keen eyes for spam, by copying part of a legitimate post and adding a spam link. Can we better control this kind of spam? What possible solutions could be implemented to control this type of spam?

An example (10k only): What are ways to prevent files with the Right-to-Left Override Unicode character in their name (a malware spoofing method) from being written or read?

  • possible duplicate of Really cool spambot behaviour
    – bwDraco
    Commented Sep 7, 2012 at 15:48
  • Note: It's up to you whether to vote to close--I just found this question. I'm not sure if this is the same issue, as it involves taking content from the same question to post a spam answer, rather than taking content from a different site to post a spam question, as the linked duplicate question discusses. The fact that the content is taken from the same question may make it easier to detect through automated tools than the case discussed in the linked question.
    – bwDraco
    Commented Sep 7, 2012 at 15:50

2 Answers 2


On submitting an answer, the system could check if at least 75% (more or less) of the answer exactly matches either the Markdown code or the rendered text of the question or another answer, and reject such an answer with a message similar to the following:

Oops! Your answer couldn't be submitted because:

  • We're sorry, but as a spam prevention mechanism, we cannot accept answers with content mostly copied from (the question | another answer). Please add more content of your own.

As a side note, I will no longer add spam warnings to answers. I will add a comment to the same effect instead.

  • 1
    Then they'll just make sure 26% is randomly generated text, or change the order of paragraphs; this won't stop them if you think about the details. :) Commented Sep 7, 2012 at 19:15
  • This would be prevented if the quality filter is applied to answers. See meta.stackexchange.com/questions/127073/…
    – bwDraco
    Commented Sep 7, 2012 at 19:17
  • 3
    You can keep making posting rules stricter, and spammers can keep getting more sophisticated. Eventually, it will become difficult to post legitimate answers. A spammer could source text from multiple answers, what then? But don't legitimate answers also quote from multiple similar questions/answers at times? etc. Really, the only reliable method until we build Skynet is to have humans review. To that end, auto-flag posts which would trip this filter, rather than blocking them entirely.
    – Bob
    Commented Sep 11, 2012 at 12:02

Well, I really liked your approach of adding a warning label to the post.

It just confused me on first sight, as I assumed the spammer itself included the warning. Which would have been quite weird.

I took the liberty of adjusting your warning label slightly to make it stand out more:

> Spam Warning
> ------------
> This post contains an advertisement or is otherwise promotional, or contains a link to malicious content.  We suggest that you ignore this post.
> It will soon be deleted by the community and/or moderators; you can help us by flagging this post.
> ---

I just don't know if it's such a good idea if everyone starts putting warning labels on other peoples posts.

  • The idea came from this answer. I'll adjust the notice accordingly for future spam posts.
    – bwDraco
    Commented Sep 7, 2012 at 16:00
  • 2
    The community's and moderation team's decision was not to put up these banners, actually. So, no, I'd leave it be. Spam flags escalate pretty fast and we'll deal with them anyway. /cc @dra
    – slhck
    Commented Sep 7, 2012 at 17:58
  • 1
    @slhck is right, actually. We raised this question recently (about whether or not to put spam notices in spam posts) and the consensus overall was not to bother, there wasn't any point, as we have round-the-clock moderation and lots of flagging tools.
    – studiohack
    Commented Sep 7, 2012 at 18:56

You must log in to answer this question.

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