Dutch Treat
by Hans Ree

This article appeared in NRC Handelsblad on May 5, when the
score between Kasparov and Deep Blue was 1-1.

FRUIT FLY 

Chess computers have often been called the fruit flies of the
science of Artificial Intelligence. The real fruit fly is a handy object
of genetical studies. It reproduces fast and it is a simply structured
creature. Studying the small will give insights into general
mechanisms that play their role with the big as well.

The match in New York between Kasparov and Deep Blue can be
seen as a fruit fly too: the fruit fly of the study of the international
media and how easily they can be manipulated.

The German weekly "Der Spiegel" had a cover story: "Duel of the
superbrains." Quote from Kasparov: "I will defend the human
race." "Newsweek" wrote about "The brain's last stand," "Time"
quoted Kasparov as follows: "Maybe the biggest triumph of the
Creator is to see his creatures re-create themselves." (into
supercomputers, Kasparov meant.) Thousands of other examples
could be supplied. All over the world the media sang the song that
they had sung last year, the song of the final battle between human
and machine and of Kasparov as the last stand of the human race
against IBM's Superbrain.

How do they do it, Kasparov and IBM? How is it possible that it
proves to be so easy to let everyone write down the same
nonsense? It almost seems as if chess is the measure of all things.
It would be a very arbitrary measure. The American game of
checkers has already been conquered by the machine. There was
one human being left who could beat the computer in this game
and he has died. On the other hand, the Japanese games of shogi
and go are still far too difficult for the computer. Deep into the
next century the computer might solve a cryptogram; this will be
hailed as the conquest of another last stand of the human race, and
there will be many left after that.

What we know about the relative strength of human chessplayers
and chess computers is this: computers are slowly but surely
getting better all the time and now they regularly beat good
grandmasters. Still there are about a hundred people in the world
who would beat the best computers with a convincing score of let's
say 5-1.

It has been said that Deep Blue is much better than all the other
computers, but how true this is we have no way to judge, because
during the last two years Deep Blue has been avoiding all
competition.

Since it was beaten in 1995 in the Computer World Championship
by the program Fritz, Deep Blue has not played against other
computers, nor did it play against humans other than Kasparov. At
least not in public.

It is understandable and wise that IBM gives as little information
as possible about the true strength of Deep Blue. Nothing could be
gained by doing that. Is the match between Deep Blue and
Kasparov an honest affair? Putting the question seems natural
though almost immoral according to some. We hardly have any
material on which to base a considered answer. Last year Kasparov
won with 4-2. If he had lost, he would not have been able to play a
return match this year. But if he had won 6-0, he wouldn't either.
IBM is Kasparov's sponsor. Not only does the company provide
the prize money for this match, IBM and Kasparov have also big
plans to collaborate in the coming years on the Internet with the
Club Kasparov.

It is not in Kasparov's interest to humiliate his sponsor by defeating
its machine by a big margin. On Sunday he resigned the second
game in a position which maybe was drawn. I do not think that he
did this on purpose. Probably he did not see the move that might
have saved him. But hard to find it was certainly not. The least you
can say is that Kasparov did not look very hard for it. After this,
the score was 1-1, the best possible score to generate publicity for
the match.

No chessplayer would blame Kasparov if he would follow the
maxim of professionals who used to hustle their rich customers in
the chess cafs of old: you have to win, otherwise they won't
respect you, but you should not win by too high a margin,
otherwise they won't come back. It is conceivable that among the
rich patrons there is someone who would might win a game on his
own. It is conceivable that Deep Blue is such a patron. But it is
almost inconceivable, though true, that a match between Kasparov
and his sponsor's creature would be described all over the world as
the final battle between man and machine.

Kasparov is a real master in the rhetoric that is required during
media events like this. At the end of last year in Las Palmas he
spoke about his first match against Deep Blue. Was that only a
superfast calculator? Kasparov seemed to hesitate. Maybe there
was more... "Now and then it was as if I could smell a whiff of real
intelligence," he said. Fingers under his nose, eyes intent, as if a
great insight had descended on him right at that moment. This was
not the case. The remark about the whiff of intelligence I had seen
published earlier. It was a standard bite for copy-hungry
journalists. Now in "Der Spiegel," in a piece written by Kasparov
himself, I saw the remark again. It will return to haunt us many
times, this whiff of artificial intelligence.

A good friend of Kasparov once explained to me that the 1996
match between Kasparov and Deep Blue had generated hundreds
of millions of dollars worth of publicity for IBM. The value of
their stock had significantly gone up the week of the match.

One does not have to be morbidly paranoid to suspect that with this
kind of money involved it would be too much to ask for total
openness and absolute honesty, both of IBM and of Kasparov. As I
said, material that would enable us to come to a considered
judgement as to the seriousness of the match is practically
unavailable. The media piously make the sign of the cross and
conveniently act on the assumption that in big business everything
is always as it seems. A story about sponsorship and the prices of
stocks is dull. There are so many of these stories. Silly as it is, the
story about the human brain's last stand in the fight between man
and machine is much more attractive.

Postscript

When I wrote this article I was obviously convinced that Kasparov
would win the match and obviously I was wrong. Ah well, we all
make mistakes. The mistake in the sixth game that Kasparov
himself called decisive and that he indicated made his position
resignable, he made on the seventh move, after about eight minutes
of play, in a position that he has studied extensively while
preparing for Karpov. Nobody is infallible.

I was wrong as to the final result, but I think that some aspects of
my article are still valid. I still do think that about a hundred
players would beat Deep Blue in a match. Many of them are eager
to prove it, but unfortunately it seems unlikely that they will get
the chance.

This column (apart from the Postscript) first appeared in the Dutch
newspaper NRC-Handelsblad May 5, 1997. Copyright 1997 Hans
Ree, All Rights Reserved. 