27 MAY 1995, Page 60

CHESS

Chips down

Raymond Keene

WHY ARE MIND GAMES, and chess in particular, important to us? Throughout the history of culture, prowess at mind games has been associated with intelli- gence in general; and mind games do have an extraordinary pedigree.

According to Dr Irving Finkel of the Western Asiatic Antiquities Department at the British Museum, game boards have been discovered in Palestine and Jordan dating back to Neolithic times, around 7,000 years BC. Astoundingly, this predates our current knowledge of when writing and pottery were introduced in those societies. It is likely, since many of the board games were found in tombs, that the shades of the departed had to play a game with the gods of the underworld to ensure safe conduct into the afterlife.

Board games are no longer regarded as a sort of IQ test for the dead, but they do retain their potency as symbols of intelli- gence. Goethe, for example, once describ- ed chess as 'the touchstone of the intellect'; Leibnitz liked it because it 'served to per- fect the art of thinking', while Lenin called it 'the gymnasium of the mind'. Human beings like to consider themselves the most intelligent beings on the planet, but that position is now under threat from comput- ers. Indeed, last year in London, Garry Kasparov himself was sensationally beaten by a computer, the Pentium Genius.

However, I have some good news. Despite the awesome power of computers, the best human players can, on occasion, play perfect chess. Philosophically speak- ing, the game of chess at the outset is a draw. In order to win at the highest levels one side or the other must take a risk to unbalance the position. This uncertainty principle brings with it exposure to loss. In my view, in the future, whatever level chess computers reach, championship matches between humans and computers will be open and level, with both sides scoring, so long as the humans remember that the basic barrier of the draw is an element working in their favour against the colossal calculating power of the machines.

The sole precondition for this mental arms race between humans and computers to be balanced is that chess should still continue to attract young players, fresh blood, who regard chess as a worthwhile career. If it fails to do this, the machines will take over completely. This has already happened in draughts, a related mind sport, where the best player in the world is now the Canadian computer program Chinook.

During August 1994 in Boston, Chinook won the world championship in draughts, replacing the human champions for the first time. The number of different posi- tions in draughts appears at first sight to be astronomical: 500,995,484,682,338,671,693. Chinook, however, was able to excel in draughts because of its brute force approach. That number, colossal though it might appear, is infinitesimally small com- pared to the possible number of chess posi- tions, which is approximately that number again but, staggeringly, multiplied by itself: in other words, 1020 for draughts as opposed to 1040 for chess.

It is my belief that the very strongest humans will never lose consistently to com- puters at chess so long as we collectively maintain the will, desire and commitment to beat them. Instead there will be a mutu- ally beneficial and challenging mental arms race, a symbiotic relationship with both sides learning rapidly from the other.

Indeed, last week in Cologne the pendu- lum swung the other way and Kasparov gained his revenge, defeating the Pentium Genius in a two-game match. Having won the first game, Kasparov followed my advice in the second by using the barrier of the draw to exploit the computer's inability to form threatening long-range plans. The result was a tedious game of manoeuvre which meant that Kasparov wrapped up the contest 11/2--1/2 in his favour. Paradoxically, in his win, Kasparov threw caution to the winds and only won because the machine's penchant for snatching material got the better of it. We join the game at the critical moment.

Kasparov—Pentium Genius: Cologne, May 1995.

(Diagram) Black should now play 25 ... Qa2, maintaining its queen on the critical a2-g8 diago- nal. In that case 26 Rxf7 Qxf7 27 Rd7 Qf5 28 Qxf5 gxf5 29 Rxb7 would lead to a likely draw. Instead it is seduced by the possibility of further plunder. 25 ... Qxb2 26 Qc4 Any human player would swiftly appreciate that this pin, plus the second one which swiftly arises by force, will be deadly to Black's cause. However, the evil final consequences of all this were probably beyond the computer's horizon. All that it could see was that it was two pawns ahead with no instant dan- ger threatening. 26 ...Raf8 If 26 ... Qf6 White would play 27 Rxb7 with the threat of Rdd7. 27 Rxf7 Rxf7 28 Rd8+ Bf8 29 Bh6 Qa3 30 Qe6 Now Black is trussed plug and switch. Of course, Kasparov does not trade on f8 prematurely, even though this would win Black's queen. First of all he reduces Black's defences by advancing his pawns. 30 ... Qc5 31 h4 Qb4 32 f4 Qbl+ 33 Kh2 Qb4 34 Kg2 Qa3 35 h5 gxh5 36 f5 Qb4 37 Rxf8+ Qxf8 38 Bxf8 ICxf8 39 f6 Black's position is now desperate. It played 39 ... Rxf6 and then resigned.