ML project: Rock paper scissors question -- what kind of learning?

Hey all, what kind of learning do you want to setup for this? I’m thinking supervised or reinforced. Between those two I’m thinking reinforced, because it is reward based.


This was a fun project but it’s not the machine learning that’s popular now like neural nets and nearest neighbors. If you google around there used to be a RPS tournament online where different bots would compete and most used some kind of Markov chain algorithm, at least for part of their design.

You have an advantage though and you can look at how the bots you are competing against in this project are programmed to play and act accordingly. It’s more fun than it should be to watch the anti-bots not lose to the bots.

Ok hmm… I’ll have to think about my approach some more. Thanks!