This document discusses alpha-beta pruning, an algorithm used to prune branches in decision trees during game tree search. It uses alpha and beta values to discard branches that cannot possibly change the outcome. The algorithm is demonstrated on a sample decision tree for a game with nodes A, B, C, and D showing how alpha (the best value for maximizing player) and beta (the best value for minimizing player) are updated during the search to prune branches until a terminal node is reached.