Champ To Play Electric Chessman

CAN a computer program beat the world chess champion? The answer is almost certainly ``no,'' for now - but the operative word is ``now.'' That probably explains why titleholder Gary Kasparov jumped at the chance to play a two-game match this coming Sunday against Deep Thought, the world's strongest computer chess program. ``This is basically a historical countdown,'' says New York master and chess promoter Shelby Lyman, who arranged the match. ``Deep Thought now looks at three-quarters of a million or more moves a second. In a couple of years, it will be a billion.

``There's no question that Kasparov is stronger now, but within five years, he is doomed.''

Murray Campbell, a member of the five-man team that created Deep Thought, agrees with Mr. Lyman's long-range forecast while conceding that the machine ``probably isn't quite ready for Kasparov'' at its present stage.

``But we're going to play anyway,'' he says. ``It should be a good learning experience.''

So are the results of Sunday's games a foregone conclusion?

``In competition, nothing is a certainty,'' says Lyman. ``Kasparov will only have been in New York two days, and maybe he'll still have a residue of jet lag. He's never played a computer of this strength ... and he's a bit nervous about not wanting to lose.''

Lyman also noted that the computer, with its blinding speed, is favored by the relatively speedy pace at which this match will be played, compared to the usual tempo of serious international competitions. The latter are contested at the rate of 40 moves for each player every 2 1/2 hours. This match, however, is being played under ``sudden death'' rules, with each player getting 90 minutes for the entire game.

But Mr. Kasparov has played plenty of games at fast time limits, and the consensus is that this will give him plenty of time at the current stage of Deep Thought's development.

The ability of computers to play chess has increased dramatically in the last few years (see accompanying story), and Deep Thought - developed by a group of graduate students at Carnegie-Mellon University in Pittsburgh - is clearly at the head of the class.

For more than a year, the computer has been playing in regular tournaments with enough success to maintain an official United States Chess Federation rating of 2551. That puts it among the top 30 or so players in the country, and is comparable to the ratings of some grandmasters - though not yet the very strongest ones.

Last November, Deep Thought upset the veteran Danish star Bent Larsen en route to tying for first place in a strong tournament in Long Beach, Calif. Just a few weeks ago it stood off New York grandmaster Robert Byrne in a two-game match with one win apiece. Mr. Larsen was one of the world's top players in the 1960s and '70s, and Mr. Byrne was once a world championship contender. But if the computer were a prize fighter, one would say it was being rushed along too fast - that it should be tested against younger, more formidable opponents before being thrown to the world champion.

``True - but Deep Thought can't get hurt!'' Mr. Campbell parried in a telephone interview from IBM's Thomas J. Watson research center in Hawthorne, N.Y., where he now works. ``Kasparov is on a US tour now, the opportunity presented itself, and we couldn't turn it down.''

THE match is being held at the New York Academy of Art, with one game in the afternoon and the other in the evening. Kasparov will sit at a regular chessboard, while members of the computer team sit at a computer terminal. Moves will be relayed back and forth via telephone lines. The Deep Thought team will be ``acting as robots and making the moves it tells us to make,'' as Campbell puts it.

Lyman came to prominence as host of the PBS-TV coverage of the Bobby Fischer-Boris Spassky world championship match in 1972, and has continued televising and promoting chess events ever since. He says he got the idea for this match about three years ago when he watched a top computer in action.

``I knew that within a few years it was going to be close. So I said, `Why not a match every year between the best computer and the world champion?'''

But if Lyman and Campbell are correct, only a few more years may be left when such a match will be worth contesting.