@Lurkykiller: Actually, combining Muty’s points 2 and 3 would create a situation, where realms like Living Flame would trigger somewhere around 160 000 “red flags” every single day. If we assume checking each flag takes a human GM 1 minute… That is 160 000 minutes of work / day for a single server. 160 000 minutes is… divided by 60… round up… about 2667 hours / day… 8 work hours / day… Round up… 334 GMs doing nothing but flag checks all day, every day (including weekends)… and that for just a single server.
If you are wondering, why? The answer is simple. Muty’s system is not able to identify an illegal 1 gold deal from a legal 1 gold deal. Goldsellers would simply overwhelm the system on any non-locked realm.
You can not even assign a legal maximum value to say, linen cloth, because it does not have the same value on every realm (unless you insist on maximum being the vendor value (7 coppers, if I recall correctly)). In which case, due to the AH cut, all sales create an actual loss and most people will just vendor EVERYTHING they do not need now / very soon themselves. When it comes to more valuable items, things get even harder. A legal deal at 500 gold could be illegal at 450 gold on another realm. A concept that an AI can not grasp with ease. I suppose a very well written AI could spend weeks, maybe months trying to crossreference different realms, but it would be neither fast nor cheap.
As for the 24 hour autoban system… That would get abused, too. Do not like someone? Create a throwaway account with fake personal info, bot on it, send gold to your “enemy”, when you know the target is away from his/her computer longer than 24 hours, report the throwaway account from your main account (or better yet, ask your guildmate(s) to) and “Voila!”, instant, effectively untraceable ban achieved. Of course, the target can complain and the sentence might eventually get overturned by a human GM, but even that takes time, most likely days at the very minimum, given the limited amount of staff Blizzard has left.
In short, overextensive automated punishment systems are really easy to either abuse, wreck or evade or all of those and are therefore a REALLY BAD idea.
Flagging suspicious activity is perfectly fine, issuing automated permanent bans is a big no-no.