[ohhaimarc]

This case is a perfect 'recipe' for reinforcement learning. Let me know if you want help here.

[benawad]

I considered going down the ML route, but didn't know where to start. I'd love to hear how you would approach it.