END OF GAME COMMENTARY
INCLUDING
KASPAROV'S PRESS CONFERENCE
KASPAROV Q+A WITH YASSER AND MAURICE.
Mr. Seirawan Beautiful.
Mr. Ashley Well, in what is obviously a thoroughly hopeless position for DEEP BLUE, the team decided to resign knowing full well that Garry Kasparov was going to carry out a vicious attack on the other side. We may have liked to see that king hunt, but they knew that Garry, when the fingers are drumming, and the watch is on, he knows what's going to happen.
So now we will wait for the press conference, as you said. And Garry Kasparov will also be out here, where we can prod him with some questions and congratulate him. So if you keep your seats, we will be hearing that momentarily.
Well, Yas, give us the inside scoop that you're going to have written about in your magazine Inside Chess. How do you see this match from start to finish?
Mr. Seirawan Well, overall, first of all, as one of those who predicted 4 to 2, I thought that Garry would do it without losing a game. So I was shocked -- and I mean shocked -- when Garry did lose the very first game.
I think it was a brilliant match. I think that DEEP BLUE was a very, very legitimate contender. And I must say that I am very reassured, when I think that the computer can analyze 50 billion positions every 3 minutes and the human being can still win, well, that's great for the human intuition and the greatest thing in the world, the human mind.
Mr. Ashley Do you think that Garry has his bead on the computer now, he's seen how it plays, that the computer will have a tough time matching up against Garry Kasparov with its own limitations of playing so poorly in certain positions?
Mr. Seirawan Indeed. I think that this match, at least from my perspective, means that we've got to see another 4 years of programming, software programming, to do all the tweaking because what we've heard already is they've got the hardware now. They've got the hardware to make a phenomenal amount of computations, calculations. And they're going to improve that. But really, to get the maximum out of their hardware, they're going to have to improve their software.
Mr. Ashley So the question of how to make the computer think more selectively, the big issue here for today and what we've learned is, the human being's ability to do that allows us to stay ahead of the computer.
I was very surprised, but also pleasantly so, toward the end when Garry limited his style somewhat from his attacking style to a more positionally based squeeze style. It seemed to be very effective. And I imagine all the grandmasters who have to play computers from now on will be emulating Garry Kasparov.
Mr. Seirawan Well, indeed. Garry learned. And he learned very, very quickly. From game 1 to the next, I think he looked stronger and stronger. There was those 2 series of draws, game 3 and 4. But if you remember, he had the computer on the ropes through most of those 2 games. And I was going to say once again that Garry approached the position uniquely.
Also, just to say something about Garry's style there, Garry had a very, very difficult task that he overcame incredibly well. And that task was that he had to change what was the most come comfortable and natural mode of play for himself. He loves tactics. He relishes a good tactical slugfest. He goes on the offensive at the drop of a hat. He loves attacking his opponent's king.
And he had to put all of those weapons and hide them in his back pocket and play a much different kind of chess, and he did it extraordinarily well. Not an easy thing to do at all. It's just like asking a pitcher to pitch a new style of throw.
Mr. Ashley Or pitch left-handed when you're a righty. Not typical Garry Kasparov. We see him there at the press conference with the IBM team. He'll be there for about 20 minutes. I hope they can pipe some audio in.
Mr. Seirawan I'm not a hundred percent sure if they'll be piping audio in. Perhaps I can get an answer to that.
Mr. Ashley If they can, I'm sure they will and we will be hearing questions shortly. And we'll definitely want to hear Garry Kasparov's opinion on the match so far.
Looking at the last 2 games, if we zeroed in solely on the last 2 games, this computer was made to look very bad by the world champion. Just the last 2 games, it really didn't look like a strong chess player in the last 2 games. Is that Garry's fault?
Mr. Seirawan Indeed. I think that's Garry's fault. Garry in a sense solved the riddle of the computer. He did it very, very effectively. But again, you know, it's kind of sad to say, but Garry makes a lot of his opponents look bad. You know, he really does. He's marvelously capable of that. And a lot of the opponent opponents he beats are great chess players.
So I can't read too much into this. I'll just say that DEEP BLUE has learned a lot. The programming team has learned a lot. It was a marvelous victory for them that very first game. C.J. Tan was explaining that for himself personally it was a celebration of the Chinese New Year, and he couldn't have got an better present from game 1.
And now we have piped in audio.
MR. DiBLASI We have a very busy schedule for Mr. Garry Kasparov and we'd greatly appreciate it if you would finish up the pictures, please.
Mr. Ashley You can hear some of the audio, and now we get a video feed of Kasparov looking somewhat subdued, but certainly happy.
MR. DiBLASI Ladies and gentlemen, it's a pleasure to welcome you to this press conference. I'm Joe DiBlasi, the executive director of ACM. And ACM, as the entire public, has only 4 words to describe this week Let's do it again.
Now it's my pleasure, because we don't want to take a lot of time here, to introduce Mr. Garry Kasparov.
(Applause.)
Mr. Kasparov Thank you. It was very appropriate statement. I have no doubt that it's just the first match among many of them that will take place in the future. I strongly believe that competition man versus machine could be a crucial part not only for the world of chess but also for those that would like to see how the future relations between man and computer will develop.
I would like first to congratulate IBM team for its outstanding achievement because it's a serious opponent. It's a really serious opponent. I won this match 4 to 2, but it was as tough as a world championship match and, believe me, I played very very seriously.
Now, if somebody gives me such a harsh time at the chess table, that's a real opponent. And it's the first time a machine programmed by other people, most of them who don't play chess, could create such a fight against a world champion.
Some people probably will be disappointed by game 6 or game 5 and will tell that machine in many positions is still stupid and can't understand simple things that any human player would understand instantly.
But, you know, first you have to create such a position and, second, machine has so many other advantages that it's very difficult just to watch it during the game not to miss any tactical strengths that could throw you off the board instantly.
I really enjoyed this event. I did not expect that it would be that tough. I was told that it would be a good fight, but, okay, take it easy. And probably I was very lucky to lose game 1 because that was the best warning. Otherwise, the same disaster could happen later.
And again, after game 1 I -- match was short and I had no more reasons to be warned about the strengths of my opponent.
I think the games of this match will be deeply analyzed both by the chess players and by the scientists because we know a lot now about the way machine thinks in the game of chess.
And also probably we could draw many conclusions how the machine would react in similar situations, not in chess, because I believe that first time in the history of mankind we saw, definitely I saw, something similar to artificial intellect because that's -- it has a great capacity to reduce number of mistakes and to certain level match human intuition.
What I do by just feeling that it's right or wrong machine feels or machine finds by just making these billions, billions and billions of calculations. As I said, it's the first event. I think it will take place again. I think we have to work out the rules for this these matches.
I'm sure that there will be other chess players, strong players, that would like to take a challenge. But from this match I believe there are very few chess players in the world that can really take this heat and to play such a machine. The field of the potential competitors for such a match is really limited.
But I would wait for next challenge because it's also fun and it's -- it gives me also a great opportunity to study game of chess. And I learned a lot not only to play against machine but also learned more about the game of chess because we have to go very deep. We have to mobilize more potential of our brains that we normally do in the game of chess. And it's really exciting experiment, and I'm looking forward to repeat it.
Thank you.
(Applause.)
MR. DiBLASI Before I introduce the DEEP BLUE team, let me first give a special thanks from ACM's standpoint to the gentleman at ACM that has made this all happen, Mr. Monty Newborn.
Monte?
(Mr. Newborn rises to applause.)
MR. DiBLASI And now it's my pleasure to introduce the lead individual for the IBM DEEP BLUE team, Dr. C.J. Tan.
(Applause.)
DR. TAN Thank you, Mr. DiBlasi.
As I said at beginning of this event, without Mr. DiBlasi and ACM's sponsorship, this historical event would not have happened.
First of all, I would like to congratulate Mr. Garry Kasparov for wonderful experience that we have had with him this week. And he has played excellent, excellent chess. This is opportunity that comes for many of us once in a lifetime. We really appreciate it.
And secondly, I'd like to congratulate our own DEEP BLUE team. They have done a wonderful job. This is a brilliant effort they have put together last several months.
This is really indeed a historical event. Many records, historical records, have been broken during this event, both for chess as well as the technical community. We have learned a lot from this. We have certainly found out a lot of weak points and strengths of our system.
And our goal for coming here was really to test our system and learn from this experience and take this back with us so that we can continue to advance our research and computing technology in general so that we can use this to further develop our system to serve mankind.
Thank you very much.
(Applause.)
MR. DiBLASI We're just going to take a few questions now. We have a roomful of people next-door and we have the ACM awards dinner, at which we are going to give the award and prize to the champion and IBM's DEEP BLUE.
So I'll point to anyone who has a question.
(Question from the press.)
Mr. Kasparov Yeah. I think against machine psychology, or strategical psychology, is more important because I pointed out a few times that machine, you know, could go up to unbelievable rating, like 3,000 in open positions, and it could play very weak in the positions where it doesn't see a plan.
I understand that now to play the machine, I mean, you have to limit its unlimited potential to find combinations and to threaten your king or, you know, your other pieces.
I tried to select the openings where machine didn't have a clear plan, especially if it could create any threats against my king or other weaknesses in the position.
Now, you know, the simplest conclusion, you should be very careful to create any weakness. You should be very careful to have hanging pieces in the position. And you should play more solid and positional chess because any mistakes, even positional mistakes against such an opponent, will be punished more severely than against a human opponent.
(Question by the press.)
DR. TAN I believe the question, is there any intelligent way of capturing Mr. Garry Kasparov's style of play into DEEP BLUE so that it will play better next time. Is that right?
Okay. 1 of the most difficult problem, as some of us pointed out during this week, is to capture the knowledge of some -- of an individual, especially knowledge as sophisticated as Mr. Garry Kasparov's, and translate that in algorithms or programs that computer can play with that.
That has been our research effort. A lot of intelligence has already been put into the machine. Machine doesn't have intelligence. All the algorithms and intelligent knowledge have been programmed into our system by our scientists.
(Question by the press.)
Mr. Kasparov The question is about, you know, willingness of other grandmasters to participate in such an event, because the match was really tough, and whether they would like to play and to take risk of humiliation. That's the question.
As I said, I don't think there are many grandmasters that can take this challenge seriously and to hope for real victory.
But as for, you know, elite that could play and potentially win this match, I don't know. I think that some of them will be very cautious and probably will try to avoid, because nobody is challenging anybody yet. Probably other players will be less inclined to do so.
But from my point of view, this challenge is inevitable and whatever the other players, other top players would like to do in the future, I feel it's my responsibility to accept any challenge that would come from a stronger team.
The good thing about such an event is that it is never too late to learn. I believe for the match that there are so many things to learn about the game of chess.
Mr. Kasparov I think the main distinction between us and computers, at least you can learn and I learned a lot from game 1 and game 2. And I think after the last two games, it was a result of me learning and playing the positions and playing the moves that are most unpleasant for the machine.
It is difficult for the machine to learn and it is impossible to learn during the match because this great team needs time to go back to lab and study. It is more complicated for the machine to learn from these games. It will take more time. That's why it requires, I think, more time to make its homework.
But again, the crucial thing is that I learned quickly and I found the right strategy and in fact I avoided the things that must be avoided.
Game one told me enough is enough. Don't create witnesses. Don't play an open position where it has chances to touch your king and I avoid it.
In the next five games I avoided the direct clash. The positions were open but still the possibility of machine to create something really unpleasant for my king and other pieces, these possibilities were very, very limited.
(Question from audience.)
MR. TAN Just to answer your question very quickly, I believe personally that when they have five or six different position here, one of the reasons we lost this match in the sense of chess is what we don't have the chest knowledge and experience that Mr. Kasparov has.
However, our goal for this project to start with was really to use chess as a test case, as a benchmark, as a measure of how we can solve complex problems by computers and learn from them so we can improve our design of computers and we can apply this knowledge to solve other problems that are more related to our particular lives.
(Question from audience.)
Mr. Kasparov I'm very happy. I won a very important event. Maybe the most important event in my career, which I can compare only to match in '85 when I won high title against Anatoly Karpov.
As I said, this match marked a new era in chess because chess now is a big thing. Chess now is in the mainstream because all in chess we have such a competition that could bring the whole country, this country and the whole world to watch and to pay such close attention. I feel like I did a good job for chess first and probably for mankind.
(Laughter.)
(Question from audience.)
Mr. Kasparov I never believed it was a real threat. I still believe we would rather cooperate with the machines and implement other way of thinking and to benefit from that rather than to take a very hostile view.
But it will take more time for us to understand all the lessons of this match because I think it's raised so much issues in the relations between man and machine. And I think that from now on we have plenty of opportunities to use this material and to find more about future prospects.
MR. DiBLASI Thank you very much for coming. From ACM's viewpoint we will do it again. Thank you.
(Applause.)
Mr. Seirawan So we will be joined by Garry Kasparov and by the IBM team here very, very shortly. They will be on their way.
I just wanted to say that it was a great pleasure for both Maurice and myself to serve you with our commentary. I was very delighted with yourselves as an audience and from both Maurice and myself, we thank you.
(Applause.)
Mr. Seirawan It was our great pleasure.
(Applause.)
KASPAROV ANSWERS QUESTIONS WITH YASSER AND MAURICE.
Mr. Seirawan We invite everyone to have a seat. Garry, you did it.
Mr. Kasparov Not only did it. It is ultimate. But believe me, it was tough. It was quite easy in last game. I won it, and probably some people would say, you know, not worth talking about that because machine can play such weak positional game, but it was really, really difficult, and as I said at my press conference, it is an outstanding position of IBM team because it is first time -- probably much before any of us expected the computer gave such a difficult time for world champion. And congratulations.
(Applause.)
Mr. Seirawan I would like to say that Garry was saying he didn't expect such an extraordinary challenge, and from ourselves, neither did I. That first win was incredible. The first time in history that we had had a reigning world champion in tournament regulation type controls. How did that feel to have that?
MR. TAN We felt just great. How many of you scientists here like us have a chance to play against Mr. Kasparov and actually beat him in a game?
Mr. Ashley I would very much -- you know, people always say it seems as if these days Garry needs to lose a game if you are going to awaken the sleeping giant. We saw this with Anand. You lost a game and started running away with the match. This is your motivation, spot somebody a game.
Mr. Kasparov It looks so. I can hardly doubt the results in this match, but I needed this loss. It is the best for me that it happened in game 1 because I played too much games against other machines. I discussed the potential of the machines with people that had a great knowledge of the machines, and none of us could predict that it would be difficult. You know, I played game 1, you know, I didn't pay enough respect for my opponent, you know. Otherwise, I would play queen D five and repeat the moves simply, but as I said, I needed that because otherwise it would happen in another game. It would happen in game 2 or 3 or later on because after this loss I was -- I was concentrated and I mobilized more potential than I could imagine and I played very carefully and I didn't give this machine a second chance. Some chance in game 4. Really, you know. I controlled the match, and that was the key of my success. I think that is the main problem for all human beings that would play later on. We have to keep control of the game, and most of the game in fact in chess are lost because we just are not capable of controlling ourself, controlling the game, you know, through the whole event. But, I managed and I was lucky I lost game 1.
Mr. Ashley Games five and six, it looks as if DEEP BLUE didn't play as well. I would like you to address that question. Yas has made the point it didn't look as if it played as well, but Garry had a little part to do with that. Your feelings on your control of those games.
Mr. Kasparov I had had a chance to explain my views on the machine's strength before the audience here. That is the machine does not play at a fixed rating. It has big ups and downs, and the strength of the machine could vary from, let's say, 3,000 rating to 2200. It depends on the position. And I was in game five and in game 6. I was capable of creating positions where machine had very little chance to demonstrate its strengths, and that's why I think the fact that I could learn quickly, contributed mostly to my success in the last two games.
Mr. Ashley Game five seems like a paradox to me. Open position, it looks like a lot could happen and yet DEEP BLUE still played poorly. I would like somebody on the team to address that. Maybe Murray Campbell, who is also a chess player, could address this issue.
Game five did not look like the position where Garry could dominate as easily as he did in the pace of four moves.
MR. CAMPBELL We are surprised too. It did go down without much of a fight after it got into that end game and we are not quite sure exactly what is missing from the program. That is something we will have to take back up when we get to the lab.
Mr. Seirawan I would like to take this opportunity. I know we have many in our audience that would like to ask a few questions. We certainly won't keep you all up here for a great deal of time, but at this point I would like to invite the gentleman who just raised his hand.
(Audience commentary.)
Mr. Seirawan The question was to Garry. Congratulations, if there is one player in the world who you had your choice who you could play against, who would it be?
Mr. Kasparov As world champion I never had the privilege of selecting my opponents. I had to take challenge. But probably now the most exciting would be the next year of the revised version of DEEP BLUE.
(Applause.)
(Audience commentary.)
Mr. Seirawan What are your thoughts on why males dominate the game of chess, Garry?
Mr. Kasparov It's beyond my scientific knowledge.
(Laughter)
(Audience commentary.)
Mr. Ashley Yas, the question is are you ready for DEEP BLUE?
Mr. Seirawan I think, first of all, that Garry's statement is accurate, that there are very, very few players, I would say probably 60 players in the world right now, that could defeat the computer or be a legitimate match. I know a lot of them, a lot of them would not want to play the computer because the computer doesn't sweat. It doesn't get nervous, and a lot of players would have difficulties not having a human opponent. As for myself, I would love the challenge.
Mr. Ashley I must say, to follow up on that point, in London last year there was some difficulty -- not last year, the year before last actually, some difficulty with Ivanchuk not wanting to play against the computer. He said it was against his religion, so yes, some of the top players not wanting to play.
(Audience commentary.)
Mr. Ashley How confident were you going into the game?
Mr. Kasparov Unfortunately too confident. I never had any doubts, even probably little doubts. I denied after game 1, but I never had serious doubts that I would win the match, but I didn't think it would be so difficult, and I still believe that even for me there are plenty of things to learn and even if next year the reverse version of DEEP BLUE, I am sure I could also do my homework and to improve my chance for that match.
Mr. Ashley You made an interesting statement that you learned a lot of chess from that match. That is hard for somebody like me to understand, the world champion growing such after a match that he won so convincingly.
Mr. Kasparov When you play normal human chess, we do not have the same level of concentration, made in the world championship match. When you prepare for the game, you throw the pieces, I go here and he goes there and we will play. Now you have to play triple attention for any position where something is hanging. You can't just decide I go this way because position looks good. If it looks good, it doesn't work anymore. It can look good but in five moves you are out of business. That's why during this match I learn how to prepare differently for the games because you have to change the way you are preparing for the game, not only just due to some pure technical reasons, but also you have to play only very sound positional openings. You have to say, uh-huh, weaknesses avoid it. You can't create these weakness because you will be punished for that. This kind of chess disciplines chess players and I'm very happy I played the match because now I have a different feeling on certain positions. I would play more accurately and I will pay certain attention to things that we normally ignore because it was irrelevant. Computer is teaching us this factor could be very, very important, don't lose your concentration and don't underestimate certain factors.
Mr. Ashley The world champion is getting tougher from playing a machine.
(Audience commentary.)
Mr. Kasparov I created it into my mind, an opponent, even though it is a computer, this opponent had its own psychology. For instance, it played much better complicated positions. I tried to go away from the image of the machine or from some silicon monster sitting somewhere, hiding somewhere. It was an opponent, a legitimate opponent, a very strong opponent, and before each game I was trying to make an opening or strategy for the game based on my knowledge of this opponent, and I knew that I could learn much better because my opponent would need more time to learn and to come back with really sophisticated counterstrategy, but I created an opponent in my mind and it was very helpful.
Mr. Seirawan I would like to say at this point that I would like to invite Joe DiBlasi, the chairman of the ACM, sponsor of this competition on stage. And please a big round of applause.
Applause.)
Mr. Diblasi On behalf of the ACM, I think it is appropriate that we thank you all of you people who came today and so many of you came for all six games. The public's response to this has been tremendous and I think all of you can join me in my sentiments, let's do it again.
(Applause.)
Mr. Seirawan Thank you all.
(Proceedings concluded at 725 p.m..)
[ The Game | The Program | Commentary | Grandstand | Deep Blue ]
[ IBM home page | (C) | (TM) | IBM Small Planet Pavilion | Internet 1996 World Expo ]