All my posts get flagged

[lechbialek]

My posts keep getting flagged for review and now I can’t post at all, I get the message below… please help

“Your content can not be submitted. This is likely because your content is spam-like or contains inappropriate elements. Please change your content or try again later. If you still have problems, please contact an administrator.”

 

[Famine]

My posts keep getting flagged for review and now I can’t post at all, I get the message below… please help

“Your content can not be submitted. This is likely because your content is spam-like or contains inappropriate elements. Please change your content or try again later. If you still have problems, please contact an administrator.”

It's simply because your posts contain phrases that are, or you're acting in a manner that is, perceived by the spam filter as being like spam/spammers. Indeed the word "spam" itself is perceived by the spam filter as being spam-like, so that post also got flagged - as will your subsequent post if you quote this one...

I reviewed your most recent two posts before I approved them and I can totally see why! Unfortunately I can't tell you why, because anything that helps spammers evade spam filters is bad

It will likely persist until you have reached a threshold number of posts (which, again, I can't tell you), but I wouldn't worry about it; if I see your posts in the Approval Queue I'll just approve them.

 

[iyohato]

Indeed the word "spam" itself is perceived by the spam filter as being spam-like

LMFAO WUT

Unfortunately I can't tell you why, because anything that helps spammers evade spam filters is bad

Do spammers even bother to look up how to evade a specific site's filter, though? I'd have presumed most of them are partial-automated or being done by people in third-world countries that don't speak English fluently, where the alternative is shovelling **** off the streets for a penny a week.

Often poking around server logs you'll find tons of attempts that never go through, such as trying to login with credentials that don't actually exist or attempting to post new threads in whatever forum ID 1 (or the first forum in the index listing) is.

Sites that transcend across Internet culture like 4chan and Twitter tend to lend themselves to specifically-tailored scripts and people who explicitly target those platforms, but most of the time forum/blog spam is done the same sort of way and rarely do they actually "adapt" that well, since a simple catch-out like that throws the majority of them off and the few that are left are so minuscule (a sortable Q&A that pertains to the forum's topic, with a question like "Which cars are German and which are Japanese?" would be time-consuming for spammers to Google the answers to, in which they'd probably just go find a more vulnerable place to post their crap that's quicker to sign up on.

 

[Famine]

LMFAO WUT

Exactly what I said. Spammers literally put the word spam in their spam - like reverse psychology or something. There's been a whole bunch recently actually putting "SCAM ALERT" into their spam thread titles:

Reverse spam psychology, I guess.

Do spammers even bother to look up how to evade a specific site's filter?

Generally anti-spam measures are not site-specific but supplied by external businesses. One extremely common one you may have heard of (just about every Wordpress site out there uses it) is Akismet, which is actually crowd-sourced: the websites that use it also feed back into the software so that it gets better at it.

Anything that helps spammers evade its spam filters is therefore bad, as it affects every site served by that filter.

Those that are operated by humans do like to act in certain ways, and the filters - and the moderators - usually catch them before they can spam.

I don't know how many spam accounts are caught in the first few steps of the process, but the forum filter behind those catches about 20-30 a day. We get maybe one or two a month that actually makes it onto the forum, but GTP's users are pretty sharp about reporting them (overly so sometimes! But that's okay too).

Unsurprisingly, sometimes humans who aren't spammers act in the same ways, and the filters usually catch them. That's where the approval queue comes in, for staff to manually check - and it's way more of a bind for us to check 30 messages a day than it is for a user to wait six minutes for their post to be approved.

I'd say its false positive rate is about 0.2% based on the last week of approvals I've seen. Maybe less.

 

[iyohato]

Exactly what I said. Spammers literally put the word spam in their spam - like reverse psychology or something. There's been a whole bunch recently actually putting "SCAM ALERT" into their spam thread titles:

Lol, I guess they were taking an attempt at this: https://en.wikipedia.org/wiki/Bayesian_poisoning

Generally anti-spam measures are not site-specific but supplied by external businesses.

Yeah, there's also stuff like the StopForumSpam database. In one sense it's useful but at the same time, I do recall in 2017 where three people (one of which was admittedly me) submitted the IP and email of someone we didn't like (who seemingly got a kick out of crapping up sites they didn't like running an old forum software they developed, which utilised the SFS API) to their database, altering some entries in their database to fake a spam post from that person. Petty, I know, but it's possible to abuse such external services if you know what you're doing.

Not to mention someone once put me in there in 2016 when nuking my account on some board, which meant I had to change my username and email at the time.

Anything that helps spammers evade its spam filters is therefore bad, as it affects every site served by that filter.

I guess this is one of those cases where not allowing random people to see the code is an advantage.

Those that are operated by humans do like to act in certain ways, and the filters - and the moderators - usually catch them before they can spam.

Yeah, sometimes even down to picking the same kind of names...

Unsurprisingly, sometimes humans who aren't spammers act in the same ways, and the filters usually catch them. That's where the approval queue comes in, for staff to manually check - and it's way more of a bind for us to check 30 messages a day than it is for a user to wait six minutes for their post to be approved.

^_^

I'd say its false positive rate is about 0.2% based on the last week of approvals I've seen. Maybe less.

Nice