Let's talk about handling spam or rude/abusive posts

Question

I keep a pretty close watch on spam and rude/abusive posts on this site, as much as is possible for a normal user without a diamond here, using the 10k tools and some other things that I use to watch for red flaggable posts.

And what I've seen is... sometimes, sub-optimal.

Let me say that I've made a lot of research into how Stack Exchange deals with spam. I've gotten about as involved in the process as possible - from flagging spam all over the network, to destroying spam accounts as a mod on Literature.SE, to reading a ton of posts on Meta.SE about spam... Spam and flags are my specialty.

And so... here are the things that are bothering me.

The bigger issue that I see is people voting to delete spam and rude/abusive posts. I can pull up examples:

...and so on. I don't want to call out any specific users - they're not the only ones who have VTDed - but these are just some convenient recent examples.

So... voting to delete spam or rude/abusive posts is a misguided effort. Not only does it not help, it actively harms the system's ability to block these people from posting more.

When a post is flag-deleted that was red flagged, an IP block is thrown up to heavily rate-limit the person who posted the spam or rude/abusive post, the post is deleted and locked, and the post content hidden behind the revisions link (for non-mods).

When a post is manually deleted - either by a mod or by 20kers - that was red flagged, the post content is hidden by the revisions link.

That's it. Nothing else. The author doesn't suffer any blocks or anything. They don't get rate-limited if they try to post another rude/abusive or spam post. They don't get fed to SpamRam, the internal spam filters that SE has.

Needless to say... that's sub-optimal. We want the people who are trolling or spamming the site to get blocked and rate limited. We don't want them to be able to continually post without getting blocked.

Voting to delete rude/abusive or spam posts is misguided, and could cause some persistent trolls to get away with a lighter block than they would have.

---

The second thing is editing red flaggable posts. While a simple edit won't invalidate red flags on the post (as far as I know), a rollback will invalidate any extant red flags on the post. While this is less common than voting to delete red flaggable posts, this is also something that I have seen and that is less than optimal - a determined troll who is familiar with the system can rollback the post after it's edited and invalidate any flags, making it take a whole lot longer to delete (and get them IP blocked).

---

That's it - have any thoughts? Am I making a mountain out of a molehill? Did you learn anything?

Answer 1

I think the issue here is that you're trying to look at the big picture, training the post system to detect spam across the entire Stack Network and to IP-block and 'punish' spammers to prevent them from doing it again but ignoring the little picture, that getting these posts off our site by deleting them is also an effective way to deal with them.

Given the vanishingly small amount of spam that we actually get (less than one piece a week), given the microscopically small number of repeat-offending spammers we have on SFF and given the fact that spammers tend to use disposable and IP-spoofed accounts anyway (rendering IP-blocking pretty useless), it's largely irrelevant how we get rid of them. The important thing is that we act on them swiftly, especially when they contain objectionable content. I'd rather they were just gone rather than leaving them up in order to help train some nebulous algorithm.

Your recent post in chat is especially illuminating on the subject;

I'll note that when spam or rude/abusive content pops up in the LQP queue I always click 'looks okay' because it should be flag-deleted, not regular-deleted.

People who hang around Charcoal tend to think that systemic solutions are required because they see dozens of spam posts each day but the reality is that these are spread out over an estate of hundreds of thousands of daily posts. They aren't even background noise on most sites and barely register as a blip on SFF.

---

Should we flag them as spam? Sure, that's the ideal way of doing it and I gather our mods periodically delete the accounts of those who've posted spam anyway.

Should we care that people aren't doing it or try to make them feel bad for not following the optimal procedure? No. We have bigger fish to fry than something that causes a minor annoyance for less than a minute once a week.

Answer 2

That's it - have any thoughts? Am I making a mountain out of a molehill?

Yup.

Should people use the spam flag to flag spam? Obviously. That's why it exists. But if people delete it instead, so what? It's not like SF/F gets a ridiculous amount of spam. I struggle to see why Smokey even bothers monitoring our site when spam here is practically non-existent compared to other sites.

If you are concerned about spam and want to make the effort to clean it up, that's absolutely a lofty goal and you should be commended for it. But there are tons of things wrong with this site that people could champion as lofty goals (would you like to talk about tags for a minute?). Your pet project is not everyone elses burden to bear. For most of us, we see a post that shouldn't be here, we vote to delete it, it goes away, we're happy.

And really, if a spammer or two gets through the system and their post gets deleted and they don't get IP blocked ... so what? Exactly who does that really hurt? If they spam again, we can just catch them. If their posts keep getting deleted, they'll get blocked from posting anyway. If they don't spam again, then it doesn't really matter one way or another.

---

Besides. All 3 of your examples clearly indicate they were deleted as spam, and their revision histories claim they were deleted by Community (though they'd received other delete votes as well.) Are you saying none of those posts triggered any further action behind the scenes? I find that difficult to believe.

Answer 3

The problem is that when someone encounters a spam post in the "Low Quality" review queue, they have two options for dealing with it:

1. Scroll down, find the tiny little "Link" link, right-click and select "open in new tab", switch to the new tab, click "flag", select "spam", click "Flag answer", close the tab, and click "Skip" in the review queue.

2. Click "Delete" from the review queue.

Guess which is the lower-friction option?

Until we can manage to train spammers to only post high-quality spam, people will keep taking the easy route and use regular deletion on it.