Kasparov-Deeper Blue, Match 1997
After A Decade, IBM Victorious
Deeper Blue's triumph changes how future generations will view
chess as sport and, ultimately, its long-standing place in civilization.
On the 23rd of September, four months after defeating Kasparov,
IBM Research company spokeswoman Melinda McMullen announced that they are scrapping the
project and have no plans to revive it. Instead, the company and the
team of scientists wanted to "move on to other grand challenges",
presumably commercial ones. "What we told Garry Kasparov is that we
do not anticipate a rematch in the foreseeable future." The next day
I am very disappointed at IBM's statement.
From the beginning, I believed in this concept as a
scientific experiment coupled with a way of bringing chess into
almost every home in the world.
I am amazed to read that they wish to "quit while they are ahead."
We stand at one match all and there is enormous wordwide enthusiasm
for a tie-breaking third match.
IBM could continue this great experiment and at the same time
they could address the many questions left unanswered since New York.
In this regard, IBM's Dr. Tan promised to publish the computer's details
for the scientific world. I trust this will happen because 300 million
chess players worldwide await the answers.
This action has the public appearance on an investor cashing in their
chips on the stock market - take your profits and run. It puts a lie to
the scientific experiment and good of mankind theory they have espoused.
It is not too late to announce a sporting match under great conditions
which would be a public relations coup for IBM.
IBM's is beholden to its commerical interests, not its chess interests.
The computer hosting the Deeper Blue software is the RS/6000 SP,
IBM's latest answer to supercomputers. Soon after the match, IBM
was advertising it as "The Computer That Beat The World Chess Champion."
For the relatively small prize fund ($700,000 to the winner, $400,000
to the loser), this is astonishingly cheap marketing.
Not that that's entirely a bad thing.
The RS/6000 SP is much cheaper than the old supercomputers,
with comparable processing power due to its well-tuned
parallel processing architecture.
Applications that require such enormous processing power include
large corporate data-analysis and weather forecasting.
Some noise has been made about the possibility of a commerical
"Deep Blue Jr." being retailed by IBM, using only 1 of 32 processors
that the original machine used in the match, but retaining the basic
chess algorithms. While some stories of playing speed chess against
such a machine have appeared infrequently since last May,
their authenticity is questionable.
What is without question is that IBM's correct business strategy was to take the
victory, the publicity, and go home. They have nothing to gain by accepting
any of the numerous match challenges, whether from the Woman's World Champion
ZsuZsu Polgar (who predicted her "woman's intuition" would carry the day!!)
or from the oodles of other chess-playing machines; they have much to lose.
In this sense there is a parallel to NASA's attempts to put a man on the moon
in the late 1960s. When the project began, the entire scientific community
understood this was a once-in-a-lifetime event, and so the elite joined forces
to make history. Once Armstrong set his foot down on extraterrestrial soil,
there was little more history to be made, and the elite found other pursuits
more worthy of their talents. NASA's lists of failures since then include
fatal launching pad fires, SkyLab's premature re-entry, the Challenger disaster,
and the $1 billion dollar Mars probe that was "lost" (exploded).
What did IBM accomplish? In the eyes of the world-wide public,
regardless of their level of interest in competitive chess,
it is now generally understood that a human, no matter how dedicated,
cannot beat the best machine. To quote Jaron Lanier in Harper's magazine,
"In the old days, there was an idea that when you got good at chess,
you knew you were at the outer reaches of a certain type of thought.
There was an Olympian quality to it. After [Deeper Blue's victory],
you can no longer feel that way." Or as the Boston Herald proclaimed
after Game Six: "You Lose, Man".
And so future generations of prospective
chess players are less likely to become seriously involved with the game.
All good players I know play to prove their superiority over another
competitor; it's the thrill (or denial) of victory that keeps them
motivated. With the dominance of machines now apparently established, youngsters
today will have a decidedly different perspective than I did growing up.
This perspective will definitely damper their motivation for the study and
mental preparation that are integral to the sport of chess.
The appeal of competitive chess is lessened for them, and I think it will
contribute to an eventual sharp decline in the quantity and quality of
tournaments and matches in a few decades.
|Tale of the Tape
||IBM's Deep Blue
Positions Evaluated Per Second:
50 B Neurons
32 P2SC Processors
Garry Kasparov was assisted by his head coach and second GM Yuri Dochoian and advisor Frederic Friedel.
His ELO tournament rating at the beginning of the match was 2795.
Deeper Blue was assisted by a core technical team of six men,
aided by a production staff financed by IBM.
Chung-Jen Tan is one of the chief architects responsible for the parallel processing operating system
used by Deeper Blue.
Feng-Hsiung Hsu was responsible for the parallel-search algorithms and chess-related algorithms.
He has been with the project since 1986.
Murray Campbell has worked with Deeper Blue since 1989, when it was Deep Thought. He is a former chess
champion of Alberta, as well as chess author.
A. Joseph Hoane Jr. and Jerry Brody were software developers on the project.
The IBM team has had various grandmasters as consultants.
Joel Benjamin, one of the better American players in the past decade, was enlisted for the 1996 match,
and his involvement continued for this match. He is credited with "teaching" Deeper Blue
some of the more delicate positional qualities of chess.
The machine's ELO rating, based soley on the
performance in this match, was 2852.
The time control was 40 moves in two hours followed by 20 moves in one hour
and 30 minutes for the rest of the game thereafter. This is slightly-more-than-slightly
faster than championship-level chess of ten years ago.
|Kasparov, Garry - Deeper Blue
Deeper Blue- Kasparov, Garry
Kasparov, Garry - Deeper Blue
Deeper Blue- Kasparov, Garry
Kasparov, Garry - Deeper Blue
Deeper Blue- Kasparov, Garry
|Reti (1. Nf3)
Reti (1. Nf3)
I was not impressed by IBM's Web coverage, nor by its team's tight-lipped
responses and unavailability regarding reasonable technical questions.
The show put on for the "audience"
(thirty-five floors removed from the actual game,
but really an exclusive closed-circuit TV view accompanied by commentators)
Attendance numbered between four and five hundred; the press, which apparently
was at the same floor as the game, numbered between two-and-a-half and four hundred.
National and World Championships in Moscow were always played
in front of a not-very-silent crowd numbering between two and three thousand,
but understand that Moscow has earned its reputation as the hotseat of competitive chess.
Of the main panel of commentators,
IM Maurice Ashley was refreshingly engaging,
particularly informative to the chess expert as well as interested layman.
The software developed for the panel was seemlessly
exploited by a range of chess experts, and the interaction between verbal description
of possible variations and graphic display approached perfection.
Garry appeared in the press center shortly after the last game had finished so
disastrously for him. He took his seat next to IBM team leader C.J. Tan and
looked grimly up into the air till it was his time to speak.
Monty Newborn tried to introduce Kasparov's team but Garry stopped him:
" my team doesn't need to be recognized. Deep Blue's team does."
C.J. Tan thanked Garry, "a brave man to participate in this great experiment."
C.J. Tan continued "the computer played grandmasterly chess. Where do go from
here? We will continue to work with Garry but perhaps on a less competitive
level." Garry's response: "the competition has just started."
Garry started off with:"I have to apologize for the performance today.
The match was lost by the World Champion with very good reason."
Garry elaborated on the loss of the match, that game two had been crucial.
"It was beyond my understanding the moves Deep Blue played. It plays a move
like Be4 but misses Qb6. All the print-outs of game 1 - 6 should be published.
I have no idea what was happening behind the curtains. I'm ashamed I lost this
match, since the machine had too many weaknesses. I want to understand how Deep
Blue won the match. I want to look at the print-outs. The computer still has
generic problems and the way to prove that this was not a single event -
that the computer has a great mind, is for Deep Blue to enter competitive
chess. Make IBM a player not a sponsor."
Garry was asked if he was suggesting that IBM had been cheating on him.
He replied: "I'm suggesting, many things happened well beyond my understanding
and it should be tested on other machines to see if they show up with the same
kind of answers."
Garry about game 6: "The best player in the world cracked down under pressure.
After the computer takes on e6 (8.Nxe6), I can resign. I didn't feel like
playing. I'm human. What happened in the previous games was beyond my
understanding and I was scared."
Garry won't mind to play another match against Deep Blue but one of the
conditions he already stated is: "Only when IBM is not the sponsor.
There is no room for friendly relations and nice talks.
That was one of my mistakes in this match."
Commentary by The Week In Chess by Mark Crowther
Ways Deep Blue is Celebrating Its Victory
As presented on the 05/12/97 broadcast of LATE SHOW with DAVID LETTERMAN
10. Nailed R2D2
9. Just for the hell of it, told the I.R.S. computer to audit Bill Gates
8. Went online pretending to be a transvestite: had cybersex with Eddie Murphy
7. Spent a romantic evening at home with his new girlfriend, Jenny McCarthy
6. Got drunk and beat the crap out of a Nintendo
5. Called up Moviefone and bought Garry Kasparov twenty tickets for everything
4. Doing a guest spot on "Friends" as Monica's "brainy" new boyfriend
3. Enjoyed a refreshing game of ping-pong
2. Downloaded some pictures of Teri Hatcher and gave the ol' mouse a workout
1. He's going to www.Disneyworld.com!