If we assume that all bots cannot be seen further than the 12 bot lengths or whatever that's standard now, then the problem becomes considerably easier.
So the idea is that a bot has a minimal (and maybe maximal) level to detect R,G, and B values? How do we handle bots that are actually all black? Are we assuming they're invisible? Do we calculate a color at a certain depth and have the difference of the bots color to that?
Just thoughts. I'll admit it would be neat to have a bunch of white dots floating around because the bots are all perfectly blended into their environment.
We'll need a formula that factors in relative motion, color of the bot, background color (which is effected by depth I'd assume), and how sensitive the looking bot's eyes are. Unless I've missed something.