alpha/beta pruning

alpha/beta pruning

(games, algorithm)An optimisation of the minimaxalgorithm for choosing the next move in a two-player game.The position after each move is assigned a value. The largerthis value, the better the position is for me. Thus, I willchoose moves with maximum value and you will choose moves withminimum value (for me).

If it is my move and I have already found one move M withvalue alpha then I am only interested in other moves withvalue greater than alpha. I now consider another of mypossible moves, M', to which you could reply with a move withvalue beta. I know that you would only make a different replyif it had a value less than beta. If beta is already lessthan alpha then M' is definitely worth less than M so I canreject it without considering any other replies you mightmake.

The same reasoning applies when considering my replies to yourreply. An alpha cutoff is when your reply gives a lower valuethan the current maximum (alpha) and a beta cutoff is when myreply to your reply gives a higher value than the currentminimum value of your reply (beta).

In short, if you've found one possible move, you need notconsider another move which your opponent can force to beworse than the first one.