Skip to content

Mastering Othello: A Computer Scientist’s Triumph Over Complex Strategy

  • by

In the realm of strategic board games, Othello stands out with a staggering number of potential board arrangements, making it a formidable challenge to solve. But now, the game has met its match in the form of a computer scientist’s ingenuity. Othello, also known as Reversi, has a playing field consisting of an 8×8 grid and was first introduced in England in 1883. Despite the game’s popularity, which soared in the 1970s and led to annual World Championships, its complexity has long puzzled players and theorists alike.

The essence of Othello lies in its ability to flip the opponent’s pieces—turning black to white and vice versa—which can dramatically alter the state of the board. Human players often find it difficult to anticipate moves far in advance, but computer algorithms have been outperforming humans since the 1990s.

Enter Hiroki Takizawa, a bioinformatician at Preferred Networks, who has achieved a milestone by solving Othello. He improved an existing algorithm named Edax and tailored it specifically for Othello’s intricate problem-solving. Takizawa divided the challenge into more manageable segments, starting with analyzing positions after 14 moves and progressively tackling fewer empty spaces.

His workhorse for this achievement was MN-J, a supercomputing cluster housed at Preferred Networks, which includes the energy-efficient MN-3, once ranked the world’s top supercomputer. Through brute force computation, Takizawa demonstrated that with perfect play from both sides, Othello inevitably ends in a draw. This breakthrough lays the groundwork for a computer program capable of executing flawless Othello gameplay.

However, Takizawa is mindful of the skepticism surrounding computer-generated proofs, as system errors can occur. Yet, he’s confident in the reliability of the results, given the error-checking and correction features of the supercomputing cluster used.

Looking ahead, Takizawa ponders whether chess, with its mind-boggling 10^43 possible positions, could be the next game to fall to computational prowess. However, such an endeavor would require not just brute force but also significant theoretical advances.

For the tech enthusiasts in their forties and fifties who grew up during the rise of computer games and appreciate the intricacies of algorithmic problem-solving, Takizawa’s accomplishment is nothing short of inspirational. It’s a reminder that the games we love may hold complex challenges, but with the right combination of technology and intellect, even the most daunting puzzles can be solved.