Thanks! It's been really refreshing for
me to have a community to talk about this stuff with and interact with. As a one-man dev, even with 3 other alpha testers and a dozen beta testers, it's hard to get feedback from a large number of people like what I've been getting since the game's release. The game is the better for it, I think, and I'm keenly aware of that.
I'm like you, a lot of times the mod-based AI in games is better. I really liked Sorian's AI mod in SupCom, but it just got too slow (thirding the speed of the game in many cases) because he was limited to the threading model of SupCom itself. And even though that AI was some of the best I had seen in an RTS game, it still ultimately became predictable -- of course, this was with versions from over a year ago now, so it may have evolved well beyond that since then. He had a lot of intriguing ideas about intel gathering, etc, that he was working on at the time. Very impressive stuff, especially given that he was getting his results using a traditional decision tree for the most part.
I originally started out wanting to do something along the lines of what he was doing, but then kind of backed into the approach I wound up actually using and was surprised to see how well it worked. Once I started getting good results after just a few days of programming, I knew that was an approach I wanted to really explore more thoroughly. I've got more articles planned for that AI series, and I had hoped to have one out this week in fact, but I just ran out of time. Soon. My hope is that maybe some other devs will also find my approaches to be useful, and will integrate that into other games. There's some impressive AI in a lot of FPS games and some other genres, but I have been underwhelmed by the offerings in RTS. Hence AI War.
Anyway, you got me going there, but thanks for stopping by and chatting!
Chris
