Programmed to Win | | But not yet. Fortunately this game |
| was only the first in a six-game |
| match. Kasparov won the second |
Joseph McLellan | | | | game, partly because Deep Blue |
KASPAROV VERSUS DEEP BLUE: | | This was quite a departure from | | had technical problems, and for the |
Computer Chess Comes of Age | | the exhibition in 1985 when Kaspa- | | remainder of the match observers |
By Monty Newborn | | rov had played simultaneously | | [id:27943] enjoyed the remarkable spectacle of |
Springer. 322pp. $29.95 | | against 32 computers and won all 32 | | Kasparov gradually learning how to |
| | games. [id:27939] , he had lost one | | cope with this unique adversary.The |
SOMETHING historic and | | game to a computer in 1994, but that | | final score was 4 to 2 in Kasparov’s |
traumatic happened in Philadel- | | was in speed chess, when he had | | favour, but this does not reflect the |
phia on February 10, 1996. For the | | only 30 minutes to make all his | | of the adversaries: |
first time, a computer playing under | | moves. Computers have a significant | | Kasparov was better than Deep |
standard match conditions (two | | advantage at that speed, and nobody | | Blue but nowhere near twice as |
hours to make its first 40 moves) | | took that game very seriously. | | good. The six games between |
defeated a world chess champion. | | But playing under standard world | | Kasparov and Deep Blue, with an |
The news that an IBM program | | championship rules? That gets to | | extensive commentary that fills |
called Deep Blue had beaten Gary | | the ego. Kasparov had summed up | | some 43 pages, are the heart of this |
Kasparov was prominently featured | | the problem in 1989 before begin- | | book, but it offers much more than |
in the media and [id:27935] people | | ning a match (which he won easily) | | its title suggests. It will not quite tell |
around the world - not merely | | against a predecessor of Deep Blue | | you how to design your own chess- |
those who regularly follow chess | | called Deep Thought: “I don’t know | | playing computer program, but it |
news, but literally millions who | | how we can exist knowing that | | discusses the technical and philo- |
were interested in technology, in | | there is [id:27940] .” | | sophical aspects of this activity in |
competitive activities or simply in | | Monty Newborn, a major figure | | considerable depth, as well as its |
humanity’s position as the lord of | | in the history of computer chess and | | history, beginning with the |
creation.The game was carried on | | an ideal choice to write this | | theoretical work of Claude |
the Internet and attracted some | | definitive study, thinks we had bet- | | Shannon and Alan Turing. |
1,200 “hits” per minute. | | ter get used to the idea: | | [id:27944] there are nearly 100 |
The human race had been sym- | | “For the first quarter-century of | | games, tracing the development of |
bolically and collectively humiliated | | progress in computer chess, comput- | | computer chess skills from the mid- |
by an inanimate object, found | | ers were clearly inferior (to good | | 1960s to last year - including |
second best in the faculty that we | | human players). For the last five | | several games with a computer |
[id:27936] - our ability to solve prob- | | years, they have been battling on a | | easily won by Bobby Fischer. Some |
lems through applied reasoning. | | relatively even footing with the top | | of the early games must be among |
Since the day when the legendary | | players, and the two combatants will | | the most flagrant examples of |
“steel-drivin’ man” John Henry just | | probably remain fairly equal for the | | [id:27945] ever preserved in book |
barely won his competition with a | | next several years. But the day is | | form, but it is fascinating to watch |
steam drill and “died with his ham- | | not too far off when the best players | | computer programmers learning |
mer in his hand,” the human race | | will no longer be serious compe- | | from one another and from their |
has become used to the [id:27937] of | | tition. Computers will simply | | own mistakes, gradually refining |
machines. Fork lifts can pick up | | consider too many possibilities and | | their software and improving their |
heavier loads and automobiles can | | set up positions that are too | | hardware until their computer can |
run faster, but we still enjoy weight | | complex for [id:27941] .” | | consider millions of possible |
lifting and footracing on our limited | | So what? Is anyone bothered by | | positions per second and seriously |
human scale.The important point, | | the fact that a hand-held calculator | | threaten [id:27946]. There is a certain |
after all, is who sits in the driver’s | | can find a square root faster than | | comfort in observing the blunders |
seat. [id:27938] here we had an | | any human? The answer is that if | | perpetrated by early programs and |
assemblage of wires and silicon | | finding square roots were a com- | | even echoed as recently as Deep |
chips taking the initiative, setting | | petitive activity, as chess certainly is | | Blue’s last game against Kasparov. |
itself up in opposition to its creator | | from the human point of view, some | | At least for now. |
and soundly thrashing him. | | of us would be [id:27942] . | | |
| | | | ‘GuardianWeekly’, April 13, 1997 |
| | | | |