MonteCarlo Backgammon

Monte Carlo Backgammon is a neural net based computer backgammon player which has learned to play backgammon by self play.

Training neural nets to evaluate backgammon positions is a process whose results speak for themselves. After all, almost all of the strong computer backgammon programs in existance are neural net based.
Monte Carlo Backgammon is no exception, and the strong neural net technology inside it makes for one feisty (and truly enjoyable) opponent.