May 12, 1997
Chess | Robert Byrne
At the End, Deep Blue in a Walk
By the standards of top chess, it was a complete breakdown. As the grand finale of a historic contest, it was just no contest.
Garry Kasparov opened himself up yesterday to an attack that no leading player ever lets himself fall into. As other grandmasters, members of the press and a big crowd of spectators watched in stunned disbelief, Deep Blue overwhelmed the world champion without even heating up its circuits.
Some of the assembled grandmasters believed that the man who knows more openings than anyone had forgotten the correct way to play the opening he himself chose for this important battle.
Others speculated that when he made the move that everyone else rejects, he must have wrongly thought he had worked out a way in his pre-game preparation to hold off the brutal attack it invites and win.
Perhaps the cause of the final debacle can be found in the ferocious resistance that the incredible computer had been putting up against him.
In Game 5, on Saturday, Kasparov looked pitiful when the machine's inspired, intrepid defense transformed a position in which Kasparov thought he was winning into a draw.
Afterward, he was visibly shaken. Perhaps with that draw, his hopes -- his self-confidence -- turned to ashes.
In yesterday's game, Kasparov chose an opening, the Caro-Kann Defense, that he rarely plays. It is a conservative, firmly grounded approach, but it is vulnerable to a few fatal traps that most grandmasters know how to avoid. The variation Kasparov went into exchanges off the main center pawn with 3 . . . de 4 Ne4 and prepares a solid, slightly cramped formation difficult to break through.
The wandering knight after 5 Ng5 anticipates 5 . . . Ngf6, when the e4 knight would have moved away to avoid an exchange of pieces that would have eased the pressure on the black position.
For some time now, Caro-Kann players have played 7 . . . Bd6 to gain more development before trying to drive off the advanced white knight with . . . h6. But Kasparov either forgot the main line of this analysis or played 7 . . . h6 to taunt the computer and lure it into sacrificing a knight with 8 Ne6!?
After 10 . . . Kd8 11 Bf4, the black king was caught in the center where the attacking white pieces could get at it. Kasparov prepared some development with 11 . . . b5, but this let Deep Blue play to open more lines with 12 a4.
Deep Blue's 17 Bf5! simply threatened to cruise through the black position. The best Kasparov could do was 17 . . . ef 18 Re7 Be7, but that was as useless as anything else. It was not a question of material -- that was approximately even.
But Deep Blue's 19 c4! forced the opening of more lines in the neighborhood of the black king and Kasparov could not organize a defense. On 19 . . . Nb4, there could have followed 20 Qf5 bc 21 Ne5 Bb5 22 Ng6. Kasparov had had enough.
IBM Chess Machine Beats Humanity's Champ
In Saturday's game, Kasparov followed his regular match strategy of playing subtle hypermodern openings whose laid-back, nonaggressive early moves would create situations without pronounced features where the machine might lose its way. It didn't. Yes, it conceded the bishop pair to Kasparov as early as 4 . . . Bf3 5 Bf3, but it mobilized very efficiently and after 18 . . . Ng4 had an even game.
Deep Blue may have been too optimistic in playing 25 . . . Nc4, and Kasparov obtained the better piece play after 33 Rg6. Yet in the end, the I.B.M. powerhouse brilliantly brought up its king to save itself from what seemed like certain defeat.
Kasparov's g pawn was heading toward the queening square when 48 . . . Kb5! 49 g7 Kb4! stopped it cold. On 50 g8/Q, there would follow 50 . . . Rd1 51 Kc2 Rd2 with a draw by perpetual check.
And if Kasparov had tried 50 Rb3? instead, he would face immediate checkmate.
Thus came the draw and what must have been a depressing letdown. Surely, 24 hours later, Kasparov had not shaken off its effects.
Copyright 1997 The New York Times Company