This is cool! Through a bizarre coincidence I’ve been working on something with quite similar mechanics. “Robots” battle on a grid; you write an algorithm that helps your robot beat the others.
The big difference is it’s designed to be played live, between two players, like a game of chess. You get a limited time to write a function that allows the robot to make a decision on where to move based on the content of the squares around it. Your function is then called four times, interleaved with the opposing player’s.
It’s meant to evoke that manic hacking-to-a-deadline feeling, where you have to balance careful thought with a scramble to finish your implementation before the time runs out. You don’t have a very sophisticated API available, and your robot is almost blind, so you have to do the high level thought yourself.
There’s a very crappy single-client-only proof of concept, here. Hopefully I’ll finish it one day:
Wow, that’s great! I had a cool idea similar to this a while back:
You know games like StarCraft, where one player directs a bunch of units? So the game would have a “general” who would control all the units. The units would each be an individual player who coded its AI beforehand. During the game the player/programmer would also be able to enter commands in real-time to influence their unit. At the same time they would listen to the general for strategic instructions.
You know those movies/anime where you have one guy giving orders, and everyone else hacking away? It’d be like that. :-D