The last two weeks were pretty exciting for chess. Google DeepMind published a paper detailing how they created a chess engine, AlphaZero, that was able to crush the top computer program, Stockfish, beginning only with knowledge of the rules of the game. The games were pretty remarkable. Unfortunately, to many interested chess players, the details behind the algorithms — the methods AlphaZero was employing to make their moves — remains a complete mystery. I’m hoping this post will give an overview of how AlphaZero was coming up its moves.