>>10242687Watson: database search algorithm with limited context-based keyword analysis. Later versions got feedback from other contestant answers' to further refine searches.
Deepblue: min-max algorithm with database search algorithm of game tables.
Alphago: backpropogation training on a database of games to train a policy network and a value network which are used in conjunction to simulate games to produce a probability field of best moves. On powerful enough hardware the training data can be built from scratch by just providing the rules of the game.