
Originally Posted by
delsydsoftware
In the turn-based strategy game I'm working on now, the AI generates a threat list based on several factors. The threat list has a large influenece on the AI's decisions. The threat list code worked great, but I made a simple mistake. The code was pretty generic, so every AI player could access the data. I forgot to tell the AI to ignore it's own threat rating. So, a particularly agressive AI would eventually decide that IT was more of a threat to itself than anyone else. So, it would launch an attack against its own country,which would increase its threat level some more. Eventually, the country would nuke itself out of existance.