This is an OpenAI Gym environment. The environments provided are such that two adversarial agents can be trained to try to win against each other.

Play Quidditch as the seeker chaser keeper beater golden snitch!

Outwit the seeker and fly around as long as you can without being caught.

Play Quidditch as the seeker. Catch the snitch as fast as you can.

You have two adversarial agents--the seeker and the golden snitch--opening up some fun possibilities! I'll set up a new environment where the actions for both the seeker and the golden snitch can be provided.