58 ;;Quote: TD-Gammon is a self-training neural network for backgammon that outperforms other programs and, sometimes, human experts

59 ;;Quote: the goal of temporal difference methods is to match the learner's current prediction for a pattern with the next prediction at the next time step

59 ;;Quote: deep search can not be used for backgammon because there are several hundred possible combinations per ply

61 ;;Quote: TD-Gammon is a multilayer perception network with 40 hidden units and backgammon feature encoders

61+;;Quote: with just raw-encoding, TD-Gammon was a strong intermediate after 200,000 training games

65 ;;Quote: TD-Gammon is successful because of the randomness of backgammon and a fairly smooth outcome function

65+;;Quote: even with a random initial network, TD-Gammon would terminate in, at most, several thousand moves

67 ;;Quote: human experts use TD-Gammon to evaluate the best move for a position by playing the position to completion several thousand times