As for poker, Google DeepMind selected heads-up no-Restrict Texas Keep’em as its benchmark for this experiment. Game Arena is operating like a heads-up poker tournament in between main AI styles, with effects feeding into a public leaderboard. Google DeepMind is growing its Game Arena System to benchmark AI models in https://knoxoxwsn.tinyblogging.com/getting-my-game-arena-to-work-83937642