18

There are a number of sites on the SE network who have had apparently recurring, malicious guest posters making low-quality answers and comments, in need of moderator intervention, regularly for some number of years now. I've asked that individual sites be given the option to turn off guest answers permanently, but responses from SE staffers to that have been generally negative.

In the cases we've seen on a number of sites, dedicated guest (unregistered) posters are able to spam the sites in question with posts that are unhelpful and frequently rude. Over the years moderators have attempted to communicate with or advise these posters to no avail (example). To my understanding, these posters can't be sent direct messages or sanctioned without a registered account. The posts in question are hand-crafted, unique, and not automated.

In the linked SE Meta question from the top, one commentator suggested, "If there is another way for you or your moderators to handle that 'one bad actor' then you should try that IMO -- try to work 'within the system' or framework or defaults that the site inherits from SE". Personally, I'm not sure what system or framework exists for that, so it seems like a good question.

What mechanisms currently exist for dealing with dedicated, recurrent spam-posting guests on SE sites?


In comments, there have been requests to see examples of the regularly deleted posts that we're talking about above. I hope that these are appropriate to put here:

Deleted answer from SE Math Educators here

Standards are not the solution. Methods are the solution. Of course even the methods are butt ass hard. But standards? Objectives? AS IF the only thing needed was clarity on goals and it was soooo freeaking easy to just teach the crap.

[Just giving Dan Collins more reasons to ban me. "Dare I disturb the universe?"]

Deleted answer from SE Math Educators here

In the US, we learn rigorous proof in geometry class in Fairfax County public high schools. Then some proofs occasionally in algebra 2 and calculus.

Of course, working problems is way more important. Well, except in geometry. Then it is proof, proof, proof. About shit (Dan, use this naughty word to ban me) that nobody cares about. Like why one triangle is the same as another triangle.

And the poster now regularly self-identifies as the "guest troll" on the site, e.g. here:

SE ME "guest troll" post

(It might bear noting that I'm just a standard user, not a moderator, and I don't have ban or delete privileges.)

4
  • Do you have concerns about specific questions and answers that are not removed? We can probably take care of those pretty quickly if you point them out. I don't see anything objectively wrong with the examples you've provided.
    – Mast
    Commented Jul 25, 2020 at 17:25
  • @Mast: I think the issue is best expressed in the linked question(s) that ask to turn off guest posting permanently, because on many sites dealing with guest posts is an ongoing burden for no value-add. Someone there suggested there was a general systemic way to avoid that, hence this question. Commented Jul 25, 2020 at 17:41
  • Can you update close to (at least) the first example? It refers to an answer, but it has been deleted now and it appears as if there is something wrong with the question (it effectively redirects to the question as it scrolls to the top of the page). E.g. something like "(an answer, now deleted)". Commented Jul 26, 2020 at 9:45
  • 1
    @P.Mort.-forgotClayShirky_q: While I'm uncomfortable with posting deleted answer text, per request, I've added that above. If anyone tells me that's inappropriate I may remove it later. Commented Jul 26, 2020 at 14:13

1 Answer 1

11

If there are rude posts, flag them as such. That will feed into SpamRam and the poster will be blocked at a lower level than their guest account.

You can also report rude posts to Charcoal HQ, that room exists to point flaggers of rude, abusive and spam posts to posts they can flag. The admins there can and will change Smoke Detector so that it catches additional posts automatically if there's any pattern to the rudeness.

13
  • 2
    Is there more detail to what "blocked at a lower level" entails? I can imagine someone being skeptical based on (a) the existence of proxy services, and (b) years of experience with that not warding off a dedicated and malicious SE guest poster. Commented Jul 9, 2020 at 6:15
  • 1
    @DanielR.Collins Exactly how would it be possible to implement an automated system that could block accounts from different IP addresses without there being false positives? (IP-based blocking can prevent legitimate users who share that IP, but that's a different issue.) How could the same user with a different IP be definitively identified? I believe the SpamRam reference is talking about machine learning of the content of the messages themselves (not the poster) that make them considered to be malicious. And that corpus is only built from manually reporting such content. Commented Jul 9, 2020 at 6:24
  • 1
    @JasonBassford: Sounds interesting; I suggest the answer could be improved by including those details. But still, in the given example, the flagging said to feed SpamRam has been going on for years with no discernible improvement. Commented Jul 9, 2020 at 14:50
  • Most unwanted posts are blocked Commented Jul 9, 2020 at 14:57
  • 1
    @DanielR.Collins What metric are you using to say there has been no discernible improvement? If the amount of spam received has increased by a factor of 5 (for example), but the number of spam posts has remained the same, I would argue that it's now 5 times more effective at blocking the spam. (Otherwise, it would be following the same rules as before, and we'd be getting 5 times more spam posts than before, based on volume received.) Commented Jul 9, 2020 at 15:37
  • 2
    @JasonBassford: I'm not talking about generic or automated spam. I'm talking about dedicated malicious individual guest posters (quote, "one bad actor"), identifiable by the content of their posts. Per the links, there are particular individuals posting on SE Math Educators, Music, Sound Design, Photography, etc., for between 1 and 3 years without any apparent barrier to continued rude and low-quality interactions. Commented Jul 9, 2020 at 15:58
  • @DanielR.Collins And how exactly is that any different when it hits the servers here? It shouldn't matter what generates the spam, or how it's uploaded to the site. All that should matter from the perspective of the site users is if it's blocked. Commented Jul 9, 2020 at 16:48
  • 1
    @JasonBassford: Okay, so to answer your prior question, the metric for this question is, we're tracking several specific examples over several years, and they're not blocked. Commented Jul 9, 2020 at 18:11
  • 1
    @DanielR.Collins Then update your question with the specific example of spam (not spam account), and ask why the systems have not been trained to stop that text from appearing. I don't know for sure, but I would not be at all surprised if it would be too much of a false positive, and they require moderator intervention to make a subjective decision. (Also, who's "we," and why have the posts that you think are spam been allowed to exist for years?) If, instead, you're talking about accounts you think are spam posters, I can't see how that hasn't already been asked and answered. Commented Jul 9, 2020 at 20:31
  • 2
    To be clear: In its current state, this won't be the selected answer. The flagging (SpamRam) suggestion is clearly insufficient, because it hasn't been able to do the job in any of the last 3 years in this case. (Furthermore, the answer lacks detail on the theory of why one would expect it to be successful.)... Commented Jul 25, 2020 at 16:46
  • 2
    In addition, re: the Charcoal HQ suggestion, I interacted with users there and was told, "Handcrafted stuff is hit and miss. Smokey finds patterns; if each response is unique then there's not a huge amount we can do". Again, the current question involves an individual troll-type user who hand-crafts a response to almost every thread on a site... Commented Jul 25, 2020 at 16:48
  • 2
    In summary: None of the suggestions here do the job, per past history and experts in those tools. Commented Jul 25, 2020 at 16:48
  • 3
    More from Charcoal HQ chat: "I'm sorry, but I think this troll is out of Charcoal's league." Commented Jul 25, 2020 at 17:30

You must log in to answer this question.

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