Reinforcement learning chess - CM50270