Professor Santos discovered the SIMPLE library last year and was interested in seeing if it could be continuously used to implement and understand new games. In video games, reinforcement learning is commonly used to train artificial intelligence to perform tasks and interact with the game environment. However, this approach is limited to single-player games. When it comes to multiplayer games, you are fighting against an opponent, not the game itself. The answer to this problem lies in SIMPLE, which stands for Self-play In MultiPlayer Environments. This is a package that trains the model by pitting the current version against previous versions of itself. This allows for continuous learning and improvement that would otherwise be impossible in multiplayer games.