Digital computers applied to games
In a chapter of Faster Than Thought, Turing set out rules for choosing a chess move that "could without difficulty be made into a machine programme", and printed a game played by those rules against a weak player.
Why it matters
A chess-playing procedure was written down and tried against a person before any computer played chess: the same chapter reports that the first attempt on a machine, in November 1951, only solved mate-in-two problems.
The chapter is credited jointly to Audrey Bates, B. V. Bowden, Christopher Strachey and Turing; a remark further on names Turing as the author of the chess section. He starts from the rule of following every continuation to the end of the game, which in chess is "of merely academic interest", and turns to what a player actually does: look a move or two ahead, value the positions reached, chiefly by material, and follow the more profitable moves further. The printed game was played by these rules, with square roots rounded to simplify the calculations. The chapter gives its own reason for games, and it is not a test bed: the principal motive was "the sheer fun of the thing", and games show most clearly how poorly a machine takes an overall view of a problem.