There's a lot of spam posted to the forum these days.

It tries to catch spammers at registration time, before they ever post. When a user registers, their username, email address and registration IP address are checked against the StopForumSpam database, and then put in a queue for me to review if any of those things match. I look at the queue each morning and approve or reject manually.

If something gets through the net and the team or I clean up the spammer using the spam tools, the data is then sent to StopForumSpam to improve their database, to let it get better at catching potentially malicious registrations.

In the year or so that we've been using it (since we moved to the new forum software), I think I've rejected just one legitimate account by mistake (the user registered again, which is possible), yet I've been able to successfully stop hundreds, if not thousands at this point. There are usually 5-10 a day that get caught in the net.

 

This system is okay for stopping posters (bots) from other boards, but isn't preemptive. Considering the pattern of content for spammers, surely there's a way to detect from the content? Anything with a link in the first post should be questioned. Maybe in the sign up, specify any accounts posting a link in the first post will be permabanned. That'll offer good bot protection as it'll take a human intelligence to understand and adapt accordingly.

 

so... how about using those google captcha? the one that you just click the checkmark?

maybe make less than 5 post new user must use the captcha?

 

I'm assuming these posters are people (or smart bots) capable of clicking Captcha's. It's not so much whether a new user is a robot or person, but whether their content is spam. Thus, to me, it seems the content is where we should be looking. Spam content == banned account.

 

I think there's already something about links in the first few posts needing approval

 

I thought it used to be link privileges were denied noobs. Well new posters are managing to include spammy links, so that's not in effect. I'd also argue it's not ideal as new posters quite often want to talk about something they've read elsewhere, hence making a post. That said, limiting links to stop spam might be worth the aggro for noobs.

 

Unfortunately XenForo has nothing built in for link limits on user class.

What can be tweaked are the thresholds for StopForumSpam to reject or enqueue.

 

I've upped the number of messages users have to post before the beginner spam checks are bypassed to 10 (was 5) and I've blocked URLs from being posted by that group too. If that cuts most of it out then great. If not, we'll give Akismet a go.

 

Security question: who's shifty geezer's dad?

 

You get more registrations than I do of course but since I've had xenforo I've had one spammer get through stopforumspam. I reject all flagged accounts for email and ip and accept probably 95% of flagged accounts for username flags after checking ip source.

They last 6 months probably 90% of email and ip flagged spammers have come from India for some reason.

I get around 1-5 flagged registrations per day.

 

Anecdotally, but when I bother to check the IP source, it's usually either India or somewhere around the Phillipines. Probably paid spam farms with human operators. I think most of them aren't actually trying to get people to click their links, either. Rather I think they just want the link to be indexed for SEO reasons. Some spammers put links in profile fields for example.

Regardless, much sadface all round.

 

Yep, which makes bot checks useless.

TBH I don't see how it's worth spending money on spam. Surely no-one clicks these links or anything? Is there any actual benefit to gettinf spammed links everywhere? Heck, Google even penalises search results - I've had people requesting their spammy link be removed from an old forum I ran so they don't get penalised.

 

btw the rampant IP banning is annoying. I cant access Guru3D and a bunch of other sites, including PSN when im on mobile 4G or 3G.
Heck, i even cant comment on Eurogamer for months a while back.

too many ip range got banned. Getting dynamic IP that is "clean" is very hard.
not a problem at home though. I rent a "clean" static ip for me for 5 dollar a month.

so... if B3D will use lots of ip banning, make sure it still allows PM or something. so people that live in spam (and virus/malware/spyware*) country like me can have communication via PM or something.

*paying government insurance? you need to give your cellphone number. Then in 5 minutes you will get a bunch of scam SMS from legit sounding sender (they spoof the number to company name).

 

I don't ever ban IPs from accessing the forums, it's just a signal used at registration time to catch spammers. If you ever find yourself unable to browse Beyond3D for whatever reason, hit me up on Twitter or on the Beyond3D IRC channel and ask me to take a look.

 

This is the worst I've ever seen here at B3D, but maybe I've been lucky and the mods have managed to clean up previous instances before I could happen on them.

 

Time to learn Korean I guess!
There is nothing interesting to read in General Forums in English any more

 

Spam is generally dealt with manually by clicking a spam button, then okay, which bans the user and removes all their posts. However, there's no clean-up option for posts in the general discussion fora - there's no spam option on your posts here for example. And there's no manual ban or spam option on user accounts. Ergo there's literally nothing Arwin or I can do with this sudden attack (of completely pointless, zero conversion) spam. I may have managed to scare them off with some warnings. Until Rys is around (and it's a Bank Holiday here so he's probably partying hard at the Norwich Bacon Sandwich Festival) this spam shall remain.

Feel free to look it up on Google Translate...

 

Btw, I've never encountered Korean spam other than in B3D. Why? Is there a K-pop sub forum hidden somewhere? Are you guys secretly have Korean ancestor?