Jonathan Schaeffer is a computer scientist at the University of Alberta best known for two milestones in the game of checkers. In the 1990s he and his team built Chinook, the first computer program to win a world championship title against humans in any game. And in 2007 the same team announced that they had weakly solved checkers: with perfect play by both sides, the game is a draw.
Chinook’s rise put it against Marion Tinsley, the greatest checkers player in history, who lost only a handful of games across decades of play. The 2007 result, published in Science as “Checkers Is Solved” (Schaeffer et al.), capped roughly eighteen years of nearly continuous computation across many machines. The team searched and verified enough of the game’s positions - checkers has on the order of 500 billion billion of them - to prove the value of the starting position.
Schaeffer’s work sits in a direct line from Arthur Samuel’s 1950s checkers program, which coined the term machine learning, and it stands as one of the most thorough demonstrations that a non-trivial game can be exhaustively settled. He combined large-scale search, endgame databases, and verification rather than the neural-network self-play that later powered systems like AlphaZero.