Learning to Forecast by Explaining the Consequences of Actions

I explain a method to learn to achieve goals in games. In very complex games, such as the game of Go, a search intensive approach is intractable. A knowledge intensive approach is best suited. I represent knowledge using the combinatorial game theory and first order logic. I give a method to learn to predict the consequences of some moves. Each rule is learned using a single example, by explaining the transformations involved by the move. This learning method has been implemented in Gogol, a Go playing system. It learns to evaluate games representing the achievement of some interesting goals of the game of Go. This method is presently applied to other games and to decision making in the management of a firm.