| Checker Problem |
Barclay Cooke Computers are taking over our lives. They seem to be everywhere. But, four or five years go, when they first programmed Backgammon, they were not too sophisticated. I recall demonstrating one for a New York department store and two of its rather bizarre plays remain in my memory.
In one game, I had rolled an eleven and had run with one man. The computer rolled 5-5, hitting me and making its one point. The other two 5's were played in its outer board. I entered on the three point and left my man there. The computer now rolled 3-2, hit me from the six point with the 3, and then went to safety, piling three men on the one point, putting three men out of play far too early. In a subsequent game, the computer made a move that guaranteed it would lose a triple game regardless of what my next roll was.
Nowadays, I am sure such nonsense would not happen. Just recently, a computer known as BKG 9.8 beat the current World's Champion, Luigi Villa (so designated because he won Monte Carlo last year). In Scientific American Magazine, the new-found prowess of this current computer is written about in glowing terms. The author, Hans Berliner, who is also a chess expert, admits that it did have the best of the dice against Villa; nevertheless, he is justly proud of its victory.
However, there is one critical position where I disagree totally with one of its moves, a move heartily praised by Berliner. See what you think. I have a friend in Los Angeles, Danny Kleinman, whom I consider to be the finest analyst and technician in the game. If he says I am wrong and that the computer is correct, I will gracefully yield, but until then I feel strongly that its play was far off base here.
This is the position:
|
|
|
Now, if instead Black were to play 22/20, 6/4, 6/2 (specifically rejected by Berliner), he of course partially destroys his inner board. But his two defensive points still hamper White in White's board and Black has a spare man which he can play from the 20-point. Moreover, in a back game, there is always hope as long as the one point is still open, which is the case here.
Hans Berliner sits at the screen
of his backgammon computer.
By this play, Black keeps his game somewhat flexible, He may soon have to abandon the 20-point entirely. But these men may ultimately remake his six point giving him once again a formidable board. He will still hold the 22 point to bother White for a long while. Also, if he chooses, Black will be able to make his 1-point giving up the back game.
The gammon threat against Black is largely an illusion because his holding the 22-point is so strong. Therefore, it seems criminal to all but concede two points here by the cowardly play of placing three men on the 22-point. If Black wins this game it wins the match and I think it showed very poor judgment in its handling of the 2-2.
Incidentally, I am not second guessing here, because the computer actually won this very game, and thereby the match, in a straight race without hitting White. However, such a result in no way justifies its play. The cruel and unfair part of backgammon is that in more than any other game, the wrong play so often turns out to be the winner. What do you think? Have I convinced you, or do you go along with the computer? Danny will have the answer for us!