# research-idea-boardgame-ai-gym

Open AI Gym allows you to create virtual agents that operate in well-defined environments.

Idea: Implement a environment structure for implementing boardgames.

I'm interested in this as a modelling problem and figure out better ways to model a game.
Other folks can work on the winning part :)

A generic framework for creating "simple card games" (multiple players, known start-end conditions) would be a nice thing to have.

Anyone can implement a game in such a system. Once someone has implemented a game, others can write agents for it and start competing.

I was thinking of doing card games like bang! for a start. And then maybe try Uno? (I don't want to limit it to just the standard playing cards, but even doing that would work?)

Would such a thing be useful to people who work in AI?