Even before the first modern computer was built, humanity dreamed of using such a machine to create the perfect game player-an opponent that is a reflection of us yet undeniably superior in its level of play. Both Kasparov versus Deep Blue by Monty Newborn and One Jump Ahead by Jonathan Schaeffer explore aspects of this important piece of computing history. Newborn's book summarizes the highlights of computer chess history-setting the stage for the 1996 ACM Chess Challenge. Schaeffer, on the other hand, chronicles his personal experiences over the early '90s when he and his team created Chinook, the world champion tournament checkers player.
In Kasparov versus Deep Blue, Newborn presents a bird's eye view of computer chess's development that includes Garry Kasparov's 1996 victory over Deep Blue by a score of 4-2. Newborn's book resembles a media guide for the computer chess fan, tracing the important events over the history of the sport. The book includes the moves from and commentary about important games between man and machine -- and sometimes machine and machine. An excellent high-level view of the subject, it is ideally suited to readers interested in both computer science and competitive chess. For the reader new to either subject, it serves as a motivation to learn more. Fortunately each chapter includes a bibliography for those who wish to do so.
Overall, Kasparov versus Deep Blue asserts that computers will inevitably reach a level of play that overwhelms even the world's best human chess player. Given this thrust the book is, to some extent, the victim of bad timing. In February of 1997 the Deep Blue team and Kasparov met again, with Kasparov losing the match 3.5-2.5 in six games. Unfortunately, Newborn's book went to press before this quintessential event in the history of computer Was the victory as close as the score indicates? What role did Kasparov's emotions play in the final outcome? Although Newborn predicts Deep Blue's victory, his book ends on a hollow note because, in hindsight, it lacks a discussion of the many new questions that victory raises.
In contrast to Newborn's high-level account of computer chess, Schaeffer's One Jump Ahead is a highly personal account of his team's quest to create Chinook, the world champion checkers program. Schaeffer, mentioned briefly in Newborn's book, begins his own story in 1989. After becoming frustrated with computer chess, he decides to explore computer checkers -- thinking at first it is a solved problem. The opportunity to read about what happens next is well worth the book's price
Whereas Newborn gives a spectator's view of 50 years of man-machine competition, Schaeffer writes from the viewpoint of a creator of such programs. From beginning to end, he accurately capture the excitement of working on such an ambitious project. At times the book reads like a novel, presenting characters and situations as offbeat as those in the chess world are austere. These sometimes wacky but nonetheless believable slices of life make the book a page turner.
Just as the Deep Blue team went gunning for Kasparov, Chinook's team targeted Marion Tinsley -- the virtually unbeatable Goliath of competitive checkers. The differences between Tinsley and Kasparov typify the differences between the cultures of chess and chess players. Schaeffer uses Tinsley to exemplify those differences, remarking on them and offering his own theories for why the two games enjoy vastly different reputations and levels of interest.
Chronicling the Chinook team's experiences as they worked toward their goal, One Jump Ahead describes how the team studied competitive checkers so that they could learn to solve the game. As Schaeffer and his fellow programmers follow the path they have set, they find themselves becoming part of the dwindling checkers fraternity. Once faceless opponents become dear friends. The resulting character development provides the most interesting part of this engaging book. Schaeffer's excellent writing brings these people to life -- making us feel like we know them as he did.
Emphasizing the teamwork necessary for such projects, One Jump Ahead makes it clear there are indeed human beings on both sides of the board. Schaeffer makes the case that any computer game uses a combination of programmer-developed strategy formulas and huge databases of beginning and end-game situations -- most of which result from human-versus-human play. A computer checkers or chess machine is not a cold, evil entity but, at best, a technological conglomeration of the finest strategies and tactics humanity has devised so far. Systems like Chinook and Deep Blue are not merely scripted machines, but rather vessels that let their creators play by proxy. Ideally, these systems perform well enough to make the opponent believe he or she is actually competing against another living being, which calls to mind Turing's principal.
Monty Newborn's Kasparov versus Deep Blue convincingly argues, with numerous games as proofs, for the inevitability of the computer becoming the world's best chess player. One Jump Ahead takes this idea several steps further and reflects upon the deeper ramifications of this trend. What does it mean that machines can now outplay the best humans? Has humanity really lost something, or must we simply adapt?
One Jump Ahead is one of those rare books on computing that both informs and entertains. Similar to Cliff Stoll's The Cuckoo's Egg (Pocket Books, New York, 1990), it can be appreciated by anyone sincerely interested in computers or checkers, no matter what their level of play. Many in computer science and engineering are asked exactly what they do on a day-to-day basis. Schaeffer's book provides excellent answers to such questions. It explores most of the issues with which an engineer, whether in industry or academia, must deal: solving a problem piece by piece, trying to justify the time and money necessary to continue a project to completion, and, most importantly, the relationships gained and lost during the course of such endeavors.
For anyone considering pursuing an education in computer science and engineering, One Jump Ahead is a must read. Not only does it wallow in the project's good times, it exposes, without embarrassment, some of the bad. Schaeffer's candor makes it easy to identify with him, his team, and their opponents. His unique perspective makes One Jump Ahead first and foremost a book about people working with people rather than man battling machine.
Kasparov versus Deep Blue, on the other hand, is an excellent introduction to the subject or computer chess. It provides a whirlwind tour of the subject's history and provides plenty of directions for future study. In many ways it is a piece of history -- an artifact of a time when the idea of a computer beating Kasparov was a bold prediction rather than yesterday's news.
University of Alberta