go

go or i-go, a board game popular in Japan that probably originated in China or India as long ago as the third millennium B.C. The board is marked by a grid of 19 horizontal and 19 vertical lines to form 361 intersections. Of the 361 pieces, 181 are black stones and 180 are white. The player with the black stones begins by placing a stone on any intersection. The players alternate turns. The object for each player is to control the most territory on the board while capturing as many of the opposing player's stones as possible. Stones are captured and removed from the board when they are completely encircled and are deprived of any access, either directly or through a chain of like stones, to a free space. A game is over when all the empty spaces on the board either are controlled by one or the other player or cannot be controlled by either player. The winner is the player who controls the most open spaces after the stones captured by the opposing player have been substracted. The complexity of go has made it, like chess, a subject for artificial intelligence research, but it was regarded as a more difficult challenge than chess. In 2016–17, however, versions of Google's DeepMind AlphaGo defeated top go players.