In WoW, when you report an RMT bot, it's posts disappear from your chat log, and you never see another from from it for the remainder of your login session. Saves blacklist space.
Printable View
In WoW, when you report an RMT bot, it's posts disappear from your chat log, and you never see another from from it for the remainder of your login session. Saves blacklist space.
Here you can report them all you want, even if these messages are OBVIOUS as they are, they do absolutely nothing, be it on the pfinder or in shouts / cities. And these users are never banned even if the insinuations / annuendo are just in your face / easy to get.
Nothing is done, absolute impunity as long as you just don't show it blatantly in a screen or stream it. All is permited.
Transmog Catalog.
Banwaves are better if botting or other cheating is involved. For just banning advertisement bots it barely matters. Except that instant bans will cause more dump accounts to be made (which can be undesired). But even then, a week is a lot. A bot can become profitable within a day. (maybe the banwaves are daily, but the report is weekly, but i suspect this is not the case)
I think in this case they should just ban the message based on repetetive content (machine learning can go quite far here and generate some relatively reliable algorythms based on reports). The moment these spam messages can be identified quickly, it doesnt matter who posts them, they can just be hidden. And then they might not even need a ban to begin with. If the message itself just gets muted, it does enough. And even if there are some mismatches, as it doesnt instantly cause a ban, it barely matters. In most cases the matching factor is going to be a domain name in the url, and that is exactly the factor you want to block.
And this can even work for the raffle scam that directs to a fake forum. As again, the domain names are a easily identified factor.
Hashing the message doesnt work as they always have a suffix to break it (while machine learning can reliably purge that info and still hash it for quick checks anyway). This is one of those things machine learning should be able to handle relatively easy. The more difficult it becomes to circumvent such system, the less advertisements you get. And sure, some will go through it. And they might find a workaround relatively fast at first. But again, the more their domains become recognised, the less workarounds will work. At some point the advertisements will just become a garbled mess which obviously wont make it work anymore. The ease for reading towards people is a weakness on its own.