I was thinking about this in light of a map I'm making named Battlefront (currently in preproduction, I'll be looking for terrainers soon). I wanted to change the formula of the genre for the better. One of the things I wanted to do was smart AI divided into regiments and squads. How I was going to do this with computer AI confounded me for a time.
Then it came to me that if I was going to control them precisely I would need to be able to command each unit individually. So 1-pixel locations enter the picture. By following each soldier with a 1 pixel location I can track each and every one of them. By combining several different locations on top of this I can ultimately control entirely what the squad does. With a lot of triggers I can make the group work as team completing objectives, covering fire for other teams, and helping out the players by both getting out of their way as they run back to base and getting in the way of whatever is trying to kill the player.
Suddenly the options are limitless. The only drawback is that due to the limited number locations being placed on the map these squads will probably be smaller in numbers. This matches up with my original goal for my map though
I will work hard to produce a proof-of-concept map ASAP.
None.