"Kasparov versus Deep Blue: Computer Chess Comes of Age" by
Monty Newborn, 1997 Springer, Hardcover, English Algebraic
Notation, 322pp., $29.95

"Kasparov v Deeper Blue: The Ultimate Man v Machine
Challenge" by GM Daniel King, 1997 Batsford, Softcover, English
Figurine Notation, 112pp., $14.95

The matches between IBM's Deep Blue Computer and world
champion Gary Kasparov in 1996 and 1997 were the focus of
intense interest worldwide. How would the silicon monster fare
against one of the greatest chess geniuses of all time? If Deep Blue
won, would that mean the demise of chess as we know it?
Although the 1996 and 1997 matches did not provide definitive
answers to these and related questions, recent publications about
the matches have helped chessplayers to define their inquiries more
clearly and obtain some helpful information in return.

When our review copy of "Kasparov versus Deep Blue: Computer
Chess Comes of Age" by Newborn arrived, the initial reaction was
that it was too little too late. You see, the 1997 match had recently
concluded, and this book was about the 1996 match. Was it even
worth reviewing or was it already out-of-date?

A closer look revealed that this handsomely produced hardcover
book was much more than a story about the 1996 match; it was a
scholarly work that presented a detailed history of computer chess,
from its beginnings in 1950 through the 1996 Kasparov-Deep Blue
match. For those readers who are technically inclined, the text is
replete with discussions of algorithms, alpha-beta searches and
other such things, all of which were incomprehensible to your
humble reviewer.

The techno-babble aside, Newborn tells the story in very readable
fashion. Ten chapters and five appendices will make the reader
very well acquainted with the trials and tribulations of the
development over the last half-century of computers that play
chess. We are taken on a journey beginning with the invention of
chessplaying computers, the problems faced by the developers and
programmers, the improvements and ultimately the 1996 match
which resulted in Kasparov's victory, but not before he was
defeated by Deep Blue in Game 1.

British GM Daniel King's book is a different matter. This was one
of those instant-press books, rushed out within a week or so after
the conclusion of the 1997 match. It too however has much to
recommend it.

The first forty or so pages of King's book deal with the history of
chessplaying computers and, to a larger extent, the prior encounters
between Kasparov and the silicon monsters. Then each game of the
six-game match is presented, well annotated, with two or three
pages of introductory remarks and insights that give the reader a
nice feel for "being there."

Understandably, the emphasis on the history of man versus
machine is much less in King's book, and it does not pretend to be
otherwise. His objective was to appeal to serious players and the
more casual enthusiast, without getting bogged down in very much
technical computer discussions. From that perspective, he has
succeeded.

From Chapter 10 ("The Future") of Newborn's book...

"It is most natural that man finds a way for his newest, greatest
tool, the computer, to compete, and the game of chess is the ideal
medium for the competition. Chess may serve the computer
industry as auto racing has the auto industry. Many advances in the
auto world were first tried on racing models and then after
refinement incorporated into commercial vehicles. This may be the
pattern in the computer field, too, where techniques used by
computers to play chess are on the cutting edge of developments in
complex problem-solving. Of course, there are many ways in
which computers can compete. We have seen robot mice working
their ways through mazes, computers competing with one another
as artists and musicians and even as poets. There have been
programming contests involving hundreds of participants. But at
least for the near future, chess is likely to remain the primary
battlefield and testing ground for computers.

"The lessons of computer chess are relevant to a large class of
important problems in computer science, problems that depend on
search such as automated reasoning, molecular synthesis,
scheduling problems, and even in the design of computers
themselves. In automated reasoning - as was touched on in Chapter
4 when discussing automated theorem proving - a computer is
given a set of statements about some problem and asked to draw
some conclusion. The computer might be given information, for
example, about the objects in a room and asked whether a robot
can navigate from one side to the other. In chess, the computer
searches for the optimal move; for the robot, the computer also
searches for the optimal move, though it is a very difficult type of
move. In molecular synthesis, one to design a molecule with a
particular property, and the computer must explore millions of
possible configurations of atoms searching for just the right
combination. Scheduling problems occur everywhere, from the
airline industry trying to optimize flight times to universities trying
to minimize final examination conflicts. In computer design, there
are VLSI layout problems and logic circuit minimization problems
for both combinational and sequential circuits. Search is a
fundamental tool for solving such complex problems, and those
interested in solving them have much to learn from the creators of
chess-playing systems.

"We are at a point in the history of computer chess where the
abilities of the top players and the strongest computers are
comparable. For the first quarter-century of progress in computer
chess, computers were clearly inferior. For the last five years, they
have been battling on a relatively even footing with the top players,
and the two combatants will probably remain fairly equal for the
next several years. But the day is not too far off when the best
players will no longer be serious competition. Computers will
simply consider too many possibilities and set up positions that are
too complex for mere mortals to cope with.

"Will the top players improve to meet the challenge? They
certainly will improve some. Track and field records have
continually improved because of better training techniques and
because of human breeding; the same is happening in the world of
chess. Children now have chess programs available that play at the
master level. They serve as outstanding sparring partners and
teachers. In the past, it was a rare occasion for a beginner to have
an opportunity to compete with a master, but now anyone can do
so. Further, just as athletes seem to breed athletes, chess players
tend to breed chess players. However, when compared to the rapid
improvement by computers, man's progress will be slow and
modest. There are limitations on the information processing
capabilities of the human mind that cannot change overnight.

"Will Kasparov learn to defeat future programs as he seems to have
done with DEEP BLUE? To some degree, and in the short term,
this will happen. The style of computers is currently too rigorous
and gives itself away after a number of games. Nevertheless, there
seems to be a limit on how much such learning can yield.
Currently, players with rating less than 2400 have little chance
against DEEP BLUE no matter how many games they play. For
Kasparov to have no chance, DEEP BLUE has a long way to go,
but the day keeps getting nearer."

These books are both worthwhile. Those interested in a
comprehensive history of computer chess, and, who are not
intimidated or put off by the injection of technical discussions, will
surely like Newborn's book. And for an affordable account of the
second match, Daniel King has definitely done a credible job.