Stone Ground Hearth Battles

Summary Stone Ground Hearth Battles (https://github.com/JDBumgardner/stone_ground_hearth_battles) is a project to create a simulator for Hearthstone Battlegrounds, and apply reinforcement learning to create a superhuman agent.  We currently have a nearly complete simulator, and can train a beginner level agent. Origins Stone Ground Hearth Battles started around the beginning of 2020, suggested by a friend jazzcs who was an avid player of the game, with a peak ELO of ~10k (we also considered TFT, but we thought simulating it would be much harder).  I also had a couple of other friends who were learning to code, or learning machine learning.  The project was initially conceived as an attempt to apply the same techniques as the Alphastar/Dota2 papers to Hearthstone Battlegrounds, creating an AI agent that can crush all human opponents. Hopefully, along the way we would learn something, regardless of whether we succeeded. Simulator Our first step was to begin developing a simulator.  Sinc