A week before his showdown with
Deep Blue, Garry Kasparov sat down to discuss the significance of the rematch.

Q:What is at stake here? Is it about more than money?

GK:It's about the supremacy of human beings over machines in purely intellectual fields. It's about defending human superiority in an area that defines human beings.

Q:Why should people who aren't necessarily fans of chess be interested in the event?

GK:Because it tells us where we stand in a world of intelligent machines. There is always a deep fascination in watching a battle between two different and rival systems. When Fischer played against Spassky it was the free world against communism. People who knew nothing about chess were deeply concerned about the outcome. People are even more concerned about the results of the battle between man and machine.

Q:Please describe your training routine for this match. How long have you been preparing? Where? What does the training consist of?

GK:How long have I been training for this match? One could argue all of my life! Actually, I have done intense general chess training for the last two years and in the last four weeks I have trained specifically for this match against the computer. It was done in Podolsk, which is just outside Moscow. Interestingly, I make a lot of use of computers in my training. I think the other side works more intensely with human beings.

Q:Do you prepare differently for a machine than for a human opponent? How so?

GK:My preparation is completely different. There are very distinct differences between human opponents and computers. Against the former I like to prepare daring and innovative strategies. Against the computer that would be fatal. I have to rely completely on my experience and intuition, try to probe for long-term weaknesses rather than to launch aggressive attacks. I have to prepare myself to be cautious and patient, which is not the most natural thing for my personality.

Q:What did you learn about Deep Blue during last year's match that you can use to your advantage this time around?

GK:The last time I was surprised by the strength of the machine -- its ability to play as though it had a plan and how it understood the essence of the position. I also felt the full awesome force of its tactical strength. On the other hand, there were positions in which it was surprisingly weak. I know that I must try to reach such positions -- again, positions I obviously understand much better than a machine. This time I know what to expect. Mainly, I know that it is going to be tough. I mustn't take the match too lightly.

Q: How does Deep Blue compare with your top flesh and blood opponents?

GK:In many ways it is more difficult to play against this machine. It never tires, never makes tactical mistakes from which you can profit. You have to be on full guard every move of the game, which means it is more exhausting. It never gives you a break. I think Deep Blue is stronger than all but a handful of top human players.

Q: Isn't this a no-win proposition for you? I mean, if you triumph, well, of course-- you're the world champ playing an overgrown Erector set. And if Deep Blue wins, well, the world champ was beaten by an overgrown Erector set. Why do you do this?

GK:Believe me, Deep Blue is much more than an overgrown Erector set. It represents the results of more than ten years of research by a number of brilliant scientists. IBM has spent millions of dollars funding the project. If I am able to beat the machine, then I think I shall be very proud of myself.

Q:What are the implications of all this? What would a Deep Blue victory signify? One day, do you think it will be impossible to beat these machines? Will that be the end of chess?

GK:A victory by Deep Blue would be a very important and frightening milestone in the history of mankind. People in future generations would look back and say this was the moment when for the first time a machine was superior to all human beings in a purely intellectual field. I am sure that one day it will happen. I am just trying to push that day as far into the future as possible. When it does happen, it will not spell the end of chess. Computer programs today can play better than 99.999% of all human players, but still millions go to chess clubs to enjoy the game.

Q: For many people the Deep Blue match is fraught with symbolic importance. What will the match say about the relationship between man and computers? Why is it important for you to win?

GK:The match draws attention to a very important questions that will confront us in many different areas in the not-so-distant future. Deep Blue shows us that machines can use very different strategies from those of the human brain and still produce intelligent behaviors. If you watch the machine play -- and especially when you play against it -- it is very difficult not to think of it as being intelligent. Man will have to accept that using the specific faculties of the human brain is not the only way to solve intellectual problems. It is important for me to win the match because then I will have the feeling that at least for the time being the human brain still has the edge.

Q:How long do you think you can keep on beating the computer? Why?

GK:I hope for a good many years. I hope that when my son becomes interested in these things I will still be on top. But I know that one day I and all my colleagues will be beaten. It is only a question of whether that day is two or 20 years away. A year ago I predicted that in 2010 we would have a computer that can beat the world champion in serious, tournament chess. Now I think it will be somewhat earlier. But even the leading experts, the scientists who have pioneered the field, cannot agree on the figures.

Q: Judging by last year's match, how does Deep Blue need to change in order to beat you?

GK:I hope you will understand if I don't give you a completely candid answer to this question. The other side is going to have to find out by themselves. I have some ideas that I think are quite interesting, and maybe some day, when I am no longer world champion, I will probably cooperate in a project to make a machine play better chess.

Q:Do you think a computer will ever be able to "solve" chess? That is, will computers ever develop flawless strategies for ever situation?

GK:Not in the history of this universe. In chess the numbers are simply too big. Currently, scientists are solving some endgames by examining every one of the billions of positions that are possible with six pieces on the board. With all 32 pieces, the task is completely impossible. There are more -- vastly more -- possibilities in a game of chess than there are atoms in the universe. So chess will never be "worked out." However, while computers may never be able to develop "flawless" strategies, they will probably develop strategies that are far superior to anything the human brain can devise.

Q:Do you think scientists will ever make a computer that passes the Turing test - that is, develop a computer that can pass as a human in an e-mail conversation? Why?

GK:The Turing test depends not only on the computer but on the human tester. There have for a long time been programs that can fool inexperienced testers for a limited period of time. The question is "Will we always be able to find methods to reliably detect a machine at the other end of the line?" I know that for the restricted area of chess this is extremely difficult. Try watching games on the Internet and pick out the ones that are being played by computers. Some years ago a friend showed me the games that Deep Blue had played in a grandmaster tournament. He had hidden the names of the players and I was supposed to pick out the computer in each round of the tournament. It was quite a difficult task. And it is becoming more difficult each year.

Q:To prepare for matches what kind of diet and exercise program do you adhere to?

GK:I do a lot of physical exercises, including swimming, running, weights and other athletic training. I think it is very important for a top chess player to be as physically fit as possible. At the very highest levels, games can often be decided by whether a player was in good physical shape or not.

Q: Chess is considered an intellectual game. But does it have to do with intelligence, space and matrix perception, subconscious mathematical abilities, mastering of the development of possibilities -- or a combination of the above (and which)?

GK:Chess is initially a logical, calculating, mathematical game that makes use of the left side of the brain. But as a player becomes stronger he is using more and more faculties that are located in the right side hemisphere. There is a vast amount of pattern recognition involved. A strong grandmaster knows literally ten thousands of chess "words," little chunks of chess knowledge that tie up with other chunks much like words of a natural language. The strongest players understand the patterns and motifs most deeply and are able to use them most effectively. Naturally, space and matrix perception are also involved, although I am not completely sure what these terms signify in the context of a chess game.

Q:It is often said that politics, diplomacy, and war (even human relations) are "like chess." Is that an accurate assessment?

GK:There are many aspects of chess that resemble life to an astonishing degree. Naturally we should not try to explain everything in chess terms, but it is useful to look at chess for parallels. On the other hand, I have found that life and society have influenced the style of the great masters. In my chess column for a number of international newspapers I take a close look at the way in which the 13 world champions were each formed by the era in which they lived. It is a fascinating subject.

Q:Excelling in chess is said to presuppose mastery of the perception of time and movement. How important is the time factor in the game?

GK:It is not clear to me what you mean by the "time factor." Naturally the use of the time allocated to each player for the moves of a game is important. Also, "timing," the moment when, for instance, you decide to go for the break, is crucial. I remember that in two world championship games against Karpov I played a pawn advance that nobody thought could be played so early in the game. It completely unnerved my opponent and I was able to win one of the most beautiful games of our matches. The timing was perfect. But you probably mean "perception of time and movement" in a more philosophical sense.

Q:Time is first experienced, then measured. Everyone has a so-called "internal clock" -- a grandmaster must surely have a "grand clock." What makes a watch so important to you? What do you enjoy in your preferred wristwatch?

GK:In a chess game you have a fixed amount of time for a certain number of moves -- for instance, two hours for 40 moves. We chess players have to allocate the time we spend on each phase of the game very carefully. When there are simple and obvious moves to be made we do this in seconds. In complicated, critical positions we may spend an hour or more brooding over a single move. At the end of a time control we sometimes run into a "time scramble," playing many moves in just a few seconds. The most successful players know how to allocate their time effectively.

For me my wristwatch has something of a symbolic character. It serves to remind me to use my time properly. My colleagues tease me and say they can tell exactly how critical the position is by watching what I do with my wristwatch. When we have settled into a game, off comes the watch and it sits on the table next to the board. When I have essentially wrapped up the game, I seem to put it back on, which is a signal for the opponent that it is all over.

My preferred wristwatch gives me a sense of security. Naturally, I also like it to be technically perfect and an object of beauty. Men hardly wear any decoration or jewelry apart from fine watches.

Q:Space, time, movement, matrix perception; in all those areas a computer excels -- probably even surpasses a human being. Yet no computer can (at this stage at least) be said to be graced by a personality unless one counts that of its programmers. How do you feel about your forthcoming game with Deep Blue? Can a machine ever become a grandmaster -- or is that equivalent to saying that a Ferrari can win an Olympic gold medal running against, say, Carl Lewis?

GK:Computers may excel in the fields you have mentioned, but that is why it is very satisfying to me -- and probably to a billion other people -- that we can still prove that the human brain is superior. Deep Blue is already a grandmaster; it certainly has the playing strength to qualify for the title.

If Deep Blue beats the world champion, it is a different situation to a Ferrari outracing Carl Lewis. We humans know that there are many animals and machines that are faster, stronger or more agile than we. But none is smarter, more intelligent. In this area, we have enjoyed a monopoly, which is now being threatened in a specific area. There is a big difference between Carl Lewis being outrun by a horse or a Ferrari and the world chess champion being beaten by a machine.

Q:Alan Turing devised a test according to which if you can converse with "someone" in the next room and cannot understand if it is a machine or a human then (if it is, indeed, a computer) it can be said to possess a kind of self-consciousness. Does Deep Blue meet the test, or is playing chess not enough to determine that?

GK:Chess cannot replace the entire Turing test. But it is very difficult, almost impossible, to determine today whether you are playing against a human or a computer if you cannot see your opponent. Computers seem to be making plans and coming up with creative "ideas."

Often you get the feeling that the computer is trying to trick you, that it is "enjoying" the position, or even that it is laughing at you. Of course it is not, it is just processing billions of numbers. But its prodigious calculations translate into a behavior that is very hard to distinguish from the activities of a human chess master. There is a PC chess program that will even chatter during the game, making remarks like "Hey, that was a pretty good move for a human being!" or "What's up, Garry, you've run into a problem or something?" when you are thinking too long. I think pretty soon computers will be able to pass the Turing test for chess with flying colors.

Q: Albert Einstein, who changed modern physics forever, is said to have replied thus to Heisenberg's' Uncertainty Principle: "I can't perceive God playing dice". What is the area you fear to tread?

GK: Assuming you mean in areas of chess, I don't know that there are any I fear. Deep Blue in tactically complicated positions is certainly a brute that you must fear as an opponent. But it is also very exciting to watch the machine unravel the secrets of a complex position. Extremely exciting. There are also computers that have worked out certain endgames completely by analyzing every single position that can occur in that endgame and storing the results in a database. This enables machines to play the ending perfectly, "like God," as Ken Thompson, the scientist who pioneered this area, once said. Some of the five-piece endings and certainly the 6 piece endings that are currently being analyzed fill us with awe. Humans will never be able to understand the perfect strategies that computers have developed. Thankfully, these endings very seldom determine the outcome of a game.

Q:Will you use a similar strategy this time around?

GK:To a certain degree, yes. But I will be doing it better, with a greater understanding of the strengths and weaknesses of my opponent. At least in theory. Maybe the Deep Blue team has made some giant leaps forward that we know nothing about. Then, my only chance is to learn as much as I can about the machine during the first two or three games. That is what happened the last time.

Q:The people at IBM say Deep Blue will be twice as fast this time around. Does that concern you?

GK:If that were the only progress they made I would not feel too concerned. The way computers play chess, you need very major speed-ups to increase the playing strength significantly. But we know that the Deep Blue team has also had at least one full-blown chess grandmaster working eight hours a day for ten months tuning the evaluation functions of the machine. If he has succeeded in doing a good job -- and I think I know pretty well what needs to be done to help the machine -- then things might not be at all easy for me.

Q:In an earlier interview you mentioned the role of human intuition in your victory. Will this always keep us one step ahead of machines, at least in regards to chess?

GK:That's exactly what this match is all about, to find out whether intuition and experience can outweigh the superior calculation powers of the computer. Even the greatest experts in the field, the people who have worked out the theoretical background and those who are actually building the machines, do not agree what the outcome of the experiment will be. The only way to find out is to conduct the experiment -- build the machine and play it against the strongest human chess player in the world.