A zero-player game or no-player game is a game that has no sentient players.
In computer games, the term refers to programs that use artificial intelligence rather than human players.
Conway's Game of Life, a cellular automaton devised in 1970 by the British mathematician John Horton Conway, is considered a zero-player game because its evolution is determined by its initial state, requiring no further input from humans. In addition, some fighting and real-time strategy games can be put into zero-player mode where one AI plays against another AI.
Types
There are various different types of games that can be considered "zero-player".
Some games, such as Conway's Game of Life, evolve according to fixed rules from their initial setup. Others such as Snakes and Ladders evolve according to chance, but similarly the players have no decisions to make and have no impact on how the game progresses.
A more complex variation on the above is the case of artificial intelligences playing a game. Humans may have a challenge in designing the AI and giving it sufficient skill to play the game well, but the actual evolution of the game has no human intervention.
For solved games the optimum strategy for all players is known. Players can maximize their chances of winning by following these strategies, and any deviation would be sub-optimal play. Tic-tac-toe is a trivial example; two players who follow the easy-to-learn optimum strategy (rather than making sub-par decisions of their own) will always find their games end in a draw. More complex games have also been solved, for example checkers, but in this case learning the optimum strategy is beyond human capabilities. Solutions for even more complex games, such as chess or Go must exist (as per Zermelo's theorem) but they have yet to have been computed.