[Equinox]

It seems that SL is being absolutely flooded with bots. Some of which were actual spam bots, some are just spiders, but in general the formula is about the same for the lot of them. Some of them are more creative in their naming, while other bots are literally just random characters.

While it's easy enough to report spam bots, the spiders are much harder to spot since they never post. I've reported a lot of them myself, and I'm sure other users have reported a fair bit of them as well, but it's a daily occurrence. I've reported several to @mothered directly, used the report button on profiles (though I honestly don't think it works), etc.

I feel like the board would benefit from having the gates closed again. Going back to invite only even for a little bit would at least allow time for the absolute pile of bots there are currently to be filtered out.

[Oni]

Spiders are mostly benign, as they're just reading posts. Spammers are often people that want to access restricted sections, or are trying to get backlinks. As always, we're looking for more staff, but requirements are high.

There is zero chance we'll ever do this again, though. In terms of activity, money, and traffic - it was a big mistake. It also created an elitist mentality that I'd rather quit, than deal with again.

[Equinox]

Spiders are mostly benign, as they're just reading posts. Spammers are often people that want to access restricted sections, or are trying to get backlinks. As always, we're looking for more staff, but requirements are high.

There is zero chance we'll ever do this again, though. In terms of activity, money, and traffic -  it was a big mistake. It also created an elitist mentality that I'd rather quit, than deal with again.

Well I'd say I definitely don't meet requirements for staff, but kind of sad to see it's a more nuanced problem than at face.

Rate limiting perhaps? Or... a more creative solution, have a key that as a hidden input for threads and replies. The user request should send the key back, and if it's not an exact match, boom, bot spotted, don't allow them to post. To have it be even more bot-proof, have the input key rotate, and again, if the POST requests' key name and the one in the SL backend don't match, decline posting.

This way, actual users will still be able to post, and on top of that would not notice any sort of bot verification. And actual bots which are completely automated fail everytime. At the very minimum, discounting spiders, this would prevent actual spam bots from... spamming.

[Oni]

Spiders are mostly benign, as they're just reading posts. Spammers are often people that want to access restricted sections, or are trying to get backlinks. As always, we're looking for more staff, but requirements are high.

There is zero chance we'll ever do this again, though. In terms of activity, money, and traffic -  it was a big mistake. It also created an elitist mentality that I'd rather quit, than deal with again.

Well I'd say I definitely don't meet requirements for staff, but kind of sad to see it's a more nuanced problem than at face.

Rate limiting perhaps? Or... a more creative solution, have a key that as a hidden input for threads and replies. The user request should send the key back, and if it's not an exact match, boom, bot spotted, don't allow them to post. To have it be even more bot-proof, have the input key rotate, and again, if the POST requests' key name and the one in the SL backend don't match, decline posting.

This way, actual users will still be able to post, and on top of that would not notice any sort of bot verification. And actual bots which are completely automated fail everytime. At the very minimum, discounting spiders, this would prevent actual spam bots from... spamming.

Believe it or not, most of these are actually humans, or bots with human intervention/Captcha-solving. We also cross-reference registrations with StopForumSpam, as well as a couple of other things. A good portion of the people we ban successfully answer registration questions and fill the Captcha at registration and between postings.

People that are going for backlinks (normally humans/agencies), will also sneak their link into otherwise normal posts. In some cases, they'll even wait an entire day before editing the link into posts from the previous day (you replied to one yesterday). Anti-bot measures won't stop these people. We do catch many of them, but sometimes a bit late.

[Equinox]

Believe it or not, most of these are actually humans, or bots with human intervention/Captcha-solving. We also cross-reference registrations with StopForumSpam, as well as a couple of other things. A good portion of the people we ban successfully answer registration questions and fill the Captcha at registration and between postings.

Yeah, there's businesses that you send the captcha to some indian guy who solves it. Used it before. And I believe there's a lot of intervention--still though some of it is just blatant spam bots which are nothing but python scripts. Anyways, I'll close the thread for now.