In a discussion about enhancing Trust Level 1 (TL1) requirements to combat spammers, Stranik suggests adding a mandatory post count requirement, noting that current rules (reading posts, time spent, topics entered) can be bypassed by patient spammers. However, Stranik also questions the effectiveness, as savvy spammers might adapt by writing plausible posts. awesomerobot agrees, warning that requiring posts could incentivize meaningless contributions from legitimate users and that sophisticated spammers may post clean content initially, then edit it later to include spam. codinghorror counters that there’s little evidence spammers are systematically bypassing TL1, and argues that raising the bar further may not be worth it — attackers typically target easier platforms, and Discourse’s existing requirements are already quite stringent. The consensus leans toward caution: while new measures might deter some spammers, they risk unintended consequences for genuine users and may not meaningfully improve security against determined, human-run spam operations.
At the moment, the following settings are available for achieving TL1:
tl1 requires topics entered
tl1 requires read posts
tl1 requires time spent mins
Maybe it makes sense to add one more: how many posts the user must write in order to reach TL1
We have some behavior. Spammers have learned how to get around TL1, they just wait, read.
The first post will show all at once. Although, perhaps it does not make sense. They can likewise find out about this behavior. And they will simply write normal answers and wait for the switch to TL1.
The existing system allows you to filter out 99% of all spammers. But there are very smart. But then it is probably difficult to do something. Just thoughts.
Right, I think we’re finding that these are often human spammers who will always learn workarounds, and requiring a post to get permission might also result in normal users posting meaningless things to meet the requirements.
We’ve also seem situations where a spammer will make something that looks like a normal post (which they might have even copy & pasted from somewhere else in the forum!), and then come back and edit it later to add spam. These spammers can be very persistent.
I’m not really seeing any evidence that spammers are “figuring this out”. The reality is that any new effort bar — and the bar is already pretty high as-is for Discourse — means attackers will move on to easier targets of which there are billions.