ROBERT J. T. MORRIS: Good afternoon. As you know, we're at the last and determining game of the Kasparov vs. Deep Blue match. Today promises to be very, very exciting. We can take as evidence what we've seen during the last few days. Now, before we get into the match, I would like to run a couple of minutes of scholarly discussions about parallel processing and how far it has come. Will you run the tape, please. "From New York -- do you want cream and sugar with that? -- it's the Late Show with David Letterman. YASSER SEIRAWAN: I love stuff like this. Over the weekend world chess champion Garry Kasparov was defeated by the IBM computer Deep Blue. Deep Blue defeated Garry Kasparov. In a related story earlier today the New York Mets were defeated by a microwave oven. Letterman: This computer Deep Blue is a very, very intelligent machine. This machine, to give you an idea of how smart it is, it has the good sense not to pick up a drag queen on Santa Monica Boulevard! So far, ladies and gentlemen, this is the very best rehearsal we've ever had. "It's the Late Show with David Letterman. Letterman: Garry Kasparov, the world chess champion is playing a computer, programmed to make 200 million chess moves a second. And I believe the series is one win apiece and one draw, and it's the best of, I don't know, what is it, Paul, 18, 20, 24? Best of six? Whoever wins the most after six games? Or are you just making up a number? But any ways, I've been reading a little bit about this computer. They're calling it Deep Blue. Not only does it play chess. This is an amazing piece of modern technology. Look, we have some tape. Look at what else the computer does, in addition to playing chess. (Picture of computer with a ping-pong paddle.) (Picture of a computer driving a sled dog team.) (Picture of a computer vaulting a high jump.) "From NBC studios in Burbank, the Tonight Show with Jayleno!" Len owe: Over the weekend, Garry Kasparov playing chess matches against the IBM supercomputer. We were at the bar knocking down brewskies, shouting, guys are going nuts watching this chess match against the computer. I guess Kasparov won the first match and then the computer won the first game. Kasparov did not take it well. Did you see today how he reacted? I thought this was bad sportsman ship. Here he is playing the computer. There's the other man representing the computer here. Now, watch when he loses. "Check" watch what Kasparov does here? (Shows man unplugging cord.) (Audience laughter.) Len owe: I just hope that brings closure to this! (Audience laughter.) "From the NBC studios in Burbank, the Tonight Show with Jay Len owe." Len owe: And world class chess champion Garry Kasparov playing the IBM computer Deep Blue in game three today of their best-of-seven match. Deep Blue, that's an odd name for a computer. Sounds more like some Smurf porn star, doesn't it? I guess today was a tie. I guess Kasparov won, right, then he lost, and today was a tie. We've been glued to the TV watching! The sitting with the guys, the chips, doesn't get any better than that. If you saw today, did you see what happened? Kasparov switched games on the computer. Very clever strategy. Show this. Look, Garry is playing. Show the game he switched. Look, they switched to Candyland! (Audience laughter.) People think Candyland is a piece of cake! "From the NBC studios in Burbank, the Tonight Show with Jay Len owe!" Leno: Chess champion Garry Kasparov says very impressive is the IBM Deep Blue computer. Have you been following this thing? Showing signs of actual intelligence. You know, why waste time building computers that can beat Grandmasters in chess? Why? Why doesn't IBM put some time and effort trying to make a computer that won't break down while you're in Sears trying to buy a package of underwear? (Audience applause.) (Audience laughter.) Have you been following this chess match? People just on the edge of their seats watching this! It's getting really intense. Did you see Kasparov today? It looked like it was getting to him. It looked like the pressure -- show the match. Show this clip from the match. There's Kasparov. (Fans yelling from the sideline, Kasparov in a grimace.) Once those fans start trash talking like that... ROBERT J. T. MORRIS: I would like to introduce to you Monty Newborn. Monty is the chairman of the ACM chess committee and Monty is overseeing the match. Monty? MONTY NEWBORN: On behalf of the ACM, the Association for Computing, it's my pleasure to be here and participate in this exciting event. This is going to make chess history. We're anticipating an incredibly exciting afternoon. We have the world champion with 2 1/2 points. We have the IBM chess program, Deep Blue, with 2 1/2 points. And we're about seven or eight minutes from starting this game, which will last about four to five hours. Perhaps at this time if anybody has to go to the bathroom, this is a good time. Nobody will leave once we begin. Not that you can't, but nobody will leave. We have three commentators that will bring this game to life in a way that won't be clear by the time it's finished whether it's a football game, a baseball game, or a chess match. I assure you, it's a chess match. Our first commentator is Mike Valvo. Mike? Mike has been the arbiter at the last match between Kasparov and Deep Blue in Philadelphia last year. He's an International Master. He's one of the toughest guys at playing chess with his eyes closed. He can play between ten and 15 and maybe as many as 20 people with his eyes closed, and beat most of them. MIKE VALVO: All these flowery words, and a few minutes ago he said that we were reduced to weather forecasters in that last game. MONTY NEWBORN: We were talking off stage, and this last game was such an incredibly complex game that it was my understanding that nobody quite understood the whole thing, although the commentators were pretty close, I'll give them credit. Our second commentator is Maurice Ashley. Maurice is an International Master. He's a New Yorker. He was a commentator at our last match. And he's the setup man in this combination. You watch how he sets up everybody on stage here. He's terrific. Our third commentator is Yasser Seirawan. Yasser has been the United States chess champion three times. He's been one of the top players in the United States for well over a decade now. He served as a commentator as well in the last match in Philadelphia, and I'm sure that he'll add the last dimension in chess expertise to this panel of outstanding chess players and commentators. I wish you all a very exciting afternoon and I'm going to turn the floor over to Maurice Ashley, who will continue and entertain you for the rest of the day. Maurice? MAURICE ASHLEY: Thank you, Monty. Welcome once again to game six, the final game of the Kasparov vs. Deep Blue Rematch. I don't know about you, but I'm excited, we're excited, the whole world is excited. And no plug-pulling will happen today, I can assure you. There's no tomorrow. MAURICE ASHLEY: There's no tomorrow. This is it. And Kasparov has it seems been rattled by what is going on here. All of us are amazed by how Deep Blue has played. The score is tied 1-1, that is in wins, and three draws. So the winner today takes it all home. And Kasparov today is Mother's Day, which I wish a happy Mother's Day to you all. Happy Mother's Day. Kasparov's mother Clara is here. We often watches Kasparov play, "here" meaning the Equitable Building. I would like to welcome those over the WebTV and the Internet who are watching. We are in the Equitable Building now, the auditorium. On the 35th floor, Kasparov is preparing to enter to play against Deep Blue, and operator Joe Hoane, I believe it is, is awaiting Kasparov's arrival. And about Kasparov's mother watching, there's got to be some pressure on Mother's Day, Yaz, knowing that her son is going into battle. And in addition to that pressure, Kasparov has set this whole thing up as he's defending mankind. Mankind against the ever-encroaching speed and complexity and problem solving ability of the computer. Well, true justice in the human race is enough pressure for Garry, Yaz. What do you think his chances are for today? YASSER SEIRAWAN: Well, that's the 700-thousand-dollar question. That's the winner's purse. And quite frankly, Clara, Garry's mom, is nervous not only for herself and her son but it seems for just about everybody on the whole team. I'm nervous with her. Before the match I thought Garry was going to win very easily. I really thought that the computer hadn't made as much progress as it has, and, I don't know, but I have a gut instinct that Garry is going to win it. A real gut instinct. But he has put himself under a lot of, lot of pressure, playing the black pieces in the final game. He's only one once against the computer in 11 tries -- actually not in 11 tries -- but when he was black, he only managed to beat the computer once in the whole series of games. And by the way, Garry does appear to be late, and I don't know if it's a psychological ploy -- (Audience laughter.) MAURICE ASHLEY: That ain't going to work with this thing. Mike, are you as stunned as everybody else as how well Deep Blue is doing? And before you answer, Garry Kasparov has seated himself at the table. Are you stunned that he is taking his watch off preparing for battle? Are you amazed that Deep Blue has done this well. MIKE VALVO: I'm amazed and I think the computer has taught us new ways to do some things, too. It's just incredible how it drew that game. It just came out of nowhere. MAURICE ASHLEY: "That game" meaning yesterday's game? MIKE VALVO: Out of the blue -- a bolt out of the blue! That was supposed to be a joke. YASSER SEIRAWAN: A bolt out of the blue. MAURICE ASHLEY: Kasparov here you see readying himself. The press has been tremendous this match, and there are several press members here on the 35th floor. They also have a press room on the 49th floor covering the match -- MIKE VALVO: 50th floor now. MAURICE ASHLEY: 50th floor now. Hundreds of journalists. From just all over. This match has taken on epic proportions and this game will decide it all, Kasparov readying himself. The last game Kasparov had, Yaz, he had a good position, he had some strong winning chances. Do we expect to see some of what happened in the last game? YASSER SEIRAWAN: Well, Garry has been playing what I would best describe as second-rate openings. He seems to be quite cautious or is simply afraid of the preparation that Deep Blue and his team, and it seems to me that there was a whole group of Grandmasters helping Deep Blue, and so he's played some second rate openings, not his first lines of defense. We are expecting e2-e4. DB MOVE: 1 e4 MAURICE ASHLEY: And in fact the first move of this epic first game has been played. Deep Blue has played e2-e4. GK MOVE: 1...c6 DB MOVE: 2 d4 MAURICE ASHLEY: He has repeated the move c7-c6 which caused Yaz to grown because usually he plays c7-c5. Deep Blue has prepared and played d2-d4. GK MOVE: 2...d5 DB MOVE: 3 Nc3 MAURICE ASHLEY: This move d5 is in fact different from what he had done in game four. Kasparov had played d7-d6, a bit more cagey, a bit more cautious, and now he's going into what looks like a Caro-Kann. YASSER SEIRAWAN: Exactly. The first two moves for Garry, c6 and d5, it's an opening I play all the time, in fact. What we saw in game 4 was not the move d7-d5, this really obscure unusual move d7-d6. And like I said, Garry got a good game in that particular game. But, again, it's not first-line openings by Garry Kasparov. Now this move d7-d5. And let me just talk about this for a moment. Garry Kasparov is not a Caro-Kann player. The Caro-Kann is an opening that is very quiet, that is to say it's a very positional-based opening. Tactics usually come much later in the game and it's very often that in a Caro-Kann style of play, the whole game is a strategic, positional game. I had hoped as you saw me, I tried to make a prediction that Garry is going to play his favorite Sicilian, c7-c5, which he didn't do. So I don't know what Garry was expecting with the Caro-Kann. GK MOVE: 3...dxe4 MAURICE ASHLEY: Kasparov has captured, which is the most frequent move in this position. DB MOVE: 4 Nxd4 MAURICE ASHLEY: Deep Blue has responded instantly by playing Nxe4, and now Kasparov -- GK MOVE: 4...Nd7 MAURICE ASHLEY: That's a move that aze arch-rival Anatoly Karpov often plays. Mike, Kasparov has not played one opening that we would call a real opening, a dip cal Kasparov opening. DB MOVE: 5 Ng5 MIKE VALVO: By the way, Garry has played the Caro-Kann before. I have seen some games that he has played. So it's not totally new to him. It's very interesting that he is playing this line because this could be a very complicated line, very topical. So the computer is going to have a lot of what we call book moves, a lot of storeed positions. This is the most topical thing in the Caro-Kann today, this particular line. So the computer could play quickly for the next 15 moves. MAURICE ASHLEY: Well, Yaz, this last move, Ne4-g5, to many beginners' eyes, we know the principle don't move a piece twice in the opening if you don't have to. And here this knight has moved from e4 to g5 it seems without any provocation. Why don't you explain to us why this is. YASSER SEIRAWAN: Do I have to? (Audience laughter.) MAURICE ASHLEY: You're the Grandmaster. YASSER SEIRAWAN: I don't want to reveal any of my secrets now. No, chess theory has evolved over a long, long, long period of time. And this move Ne4-g5 has been played after thousands and thousands of previous games where they played Nf3, Bc4, Qe2, and other such moves. But the idea of the move knight g5, in principle, is to create an early attack against the f7 square, forcing black to play e7-e6, and then the knight usually tries to gain control of the e5 square. It's become a favorite weapon, and especially in the mid-eighties it was played almost constantly for white. As you mentioned, Anatoly Karpov is a great defender of the Caro-Kann position, a very illustrious career with the black pieces, and I think it very likely that we will see one of those openings that are analyzed out for 15 or 20 moves, because it's going to be now very hard for Kasparov to avoid those lines. In these types of positions, you don't want to play anything original, because you could get into a lot of trouble early. I think that he's going to play one of the main lines and be satisfied with the resulting position. MAURICE ASHLEY: But isn't this the kind of position that often tends to a draw? A lot of people say the Caro-Kann is a drawish opening. If you want to win, you've got to play something sharp like the Sicilian. Does he want to draw today, Mike? Is Kasparov happy with a draw today and result in a drawn match? MIKE VALVO: Obviously he's not going to be happy, but he doesn't want to lose, either! (Audience laughter.) So he's going to play a nice, solid opening. He's noticed by now that he's doing the best in the endgame against this machine, although the machine seems to be getting away at the last moment every time. He still is doing better in the endgame than any other phase of the game. Yesterday he did well in the beginning, then the computer really fought back hard and, gosh, it looked like the computer was getting an edge, and we went into an endgame, and it looked like Garry was just creaming the computer, going to queen a pawn, and all of the sudden they agreed to a draw right in front of our eyes. Nonetheless the endgame offers the best chances for Garry and that's where he's heading. He's hoping to steer through a middle game, beat back the attack that Deep Thought -- Deep Blue -- used to be Deep Thought -- that Deep Blue is presenting in front of it, and hoping for the endgame. So we may have a very clear-cut, easy-to-follow kind of game, and that's what the audience seems to appreciate. There was one game, I think it was game four, where everybody followed from beginning toned the whole game and I suspect this will be an easy game to follow, very strategic in nature, simple. It will be Yasser's kind of game, he likes this kind of game, he's going to enjoy it, he's going to say yes, the Caro-Kann is vindicated, but the rest of us know it can't be true. YASSER SEIRAWAN: Thanks for that set up. Let me just say on behalf of myself and other Caro-Kann players, yes, in general we take the perspective that white with the opening move has the opportunity of building up an advantage. So the Caro-Kann really is an equalizing weapon. But it doesn't necessarily mean that just because I play the Caro-Kann I'm playing for a draw. The Caro-Kann is a very solid setup for black and if white overpresses, he easily ends up a victim. GK MOVE: 5...Ngf6 MAURICE ASHLEY: Kasparov has in fact moved, playing his knight on g8 to f6. MIKE VALVO: You know, it's interesting. People may be wondering -- DB MOVE: 6 Bd3 MIKE VALVO: -- why Garry is taking so long to make almost forced moves. Why do you think this is the case, Yasser? YASSER SEIRAWAN: In this particular situation the opening is now become established. He was unsure of what Deep Blue was going to play on the third move. Deep Blue had lots of choices, the advanced Caro, the Panov Botvinnik, with the exchange of pawns on d5. So he wasn't absolutely sure he was going to get this position. He's now got the position. And he's preparing himself for the kind of battle that we've been talking about, a strategic battle. MAURICE ASHLEY: After Ngf6 Deep Blue has responded instantly with Bf1-d3, developing the bishop, putting it on a very solid square. Potentially Kasparov might castle king-side, so the bishop is well placed for that. GK MOVE: 6...e6 MAURICE ASHLEY: Opening a line for his bishop. And again Deep Blue is clearly in its opening book because it is playing very quickly. DB MOVE: N1 f3 MAURICE ASHLEY: Kasparov trying to get his bishop quickly into the action, we anticipate the bishop on f8 moving shortly. We would like to note that those who are following this over the Internet. Some who are unable to follow live but must follow using a web server or just follow using a chessboard. So we will try to be as visual as possible, be your eyes and ears, so to speak, because they're just following it in the written text. At the moment, then, we should say to our left we are in an auditorium in the Equitable Building. To our left is a video screen that shows the current position at all times. Behind us is a Fritz 4 computer playing program, very strong, and it has helped us to do the analysis over time and I guess to sort of understand Deep Blue in a way. But we find that Deep Blue is a bit biased in its opinion. MIKE VALVO: Not today, it likes black better. MAURICE ASHLEY: Today it's liking black. We will explain what these mean, some bar graphs and evaluations that we will explain shortly. And to our right is another video screen showing Kasparov and a variety of things including the Deep Blue team, the audience, a number of things at different times. So, back to the game position -- GK MOVE: 7...h6 MAURICE ASHLEY: Instead of bringing out his bishop with Bd6, Kasparov has instead -- DB MOVE: 8 Nxe6 MAURICE ASHLEY: Capturing on e6 instantly and Kasparov shook his head for a moment -- GK MOVE: 8...Qe7 DB MOVE: 900. GK MOVE: ...fxe6 DB MOVE: Bg6+. GK MOVE: ...Kd8 MAURICE ASHLEY: Kasparov is shaking his head as if something disastrous has happened, his king being chased around the board. Is it possible that Kasparov has played incorrect theoretically? YASSER SEIRAWAN: Yes, he has. He blunered. What he did is he transposed moves. What I mean by that is this position is quite well known, and you had witnessed me playing the move Bf8-d6. The idea being that after Bd6, it's standard for white to then play Qe2, and then after h6, this sacrifice Nxe6 doesn't work because black has the move Kf8 later. MAURICE ASHLEY: You mean after Nxe6? YASSER SEIRAWAN: Capturing the knight, there's the check, the king can go to f8. But playing h6 one move earlier, the sacrifice that we've now seen, h6, is possible. As far as I recall, there was a famous game between Granda Zuniga, Grandmaster from Peru, vs. our very own Patrick Wolff. And it was a very difficult game for black to play and it became recognized that the move h6 was wrong. And Gary, assist -- Garry, as you saw his reactions, the moment that Deep Blue played Nxe6 so very quickly and reached the position they now have on the board, he was in just terror, distress. Because he's -- he recognizes that he's fallen for a well-known opening trap. MAURICE ASHLEY: Is this over? Is it that simple? MIKE VALVO: No. MAURICE ASHLEY: I mean he's up a piece for a pawn. YASSER SEIRAWAN: Right. MAURICE ASHLEY: His king is in a sorry state right now on the d8 square, to be sure. Is it just over? I remember another game between wolf and Epishin in the same line, and I believe it was somewhat difficult to just press on the attack necessarily. I know it's a strong position for white, it looks like a dangerous position, and a lot of players would be licking their chops to have a position like this against Kasparov with his king so disgusting on d8. (Audience laughter.) But I mean a piece is a piece after aall -- after all and maybe he can work his way out and not only that, one thing I'm intrigued by is Deep Blue, in what is a so-called trappy line, standard opening trap programmed into it, did not play instantly but spent a lot of time before finally deciding to play this move. DB MOVE: 11 Bf4 MAURICE ASHLEY: This traps the king on the d8 square. Now the king cannot move and Kasparov immediately has to defend. Kasparov cannot be a happy man, Mike. He played an opening that is normally not his mixier, and now he's going to have to suffer fo who knows how many moves. MIKE VALVO: I don't remember Bf4 being a correct move in that position. YASSER SEIRAWAN: In fact, I believe it is. The idea is the bishop on g6 is a very powerful bishop because it ties down black's ability to develop his king-side. If black had the chance, he would love to play the move g7-g5 and then Bg7. What white has done is sacrificed a piece for a long-term initiative. Okay, what I mean by that is the initiative means the ability to make threats. Deep Blue has an ideal attacking formation, quick development, the king is safe. Black has all kinds of problems. How is he going to develop his army? The idea of the move Bf4 is, if we go back for a moment and we consider another possible idea for white, is what Garry would like to do is play Nf6-d5, followed by Qe7-f6, getting out of the way of the bishop on f8 and trying to get rid of that bishop on g6. So the idea of Bf4 is to immediate Nd5 with Bg3, when Qf6 is no longer possible because of Bh4, winning Garry's queen and the game. MAURICE ASHLEY: How is this possible, Yaz? I know we often try to play different openings to fool our opponents, but how could the "best by test" in the world, Garry Kasparov, make such an academic blunder? I'm trying to understand it. This guy's knowledge is encyclopedic, much less -- for him to play so basic and so wrong. YASSER SEIRAWAN: And indeed Garry spoke about that yesterday. He spoke about his memory as being one of the best chess memories in the world. One of the things that -- and in fact I find most upsetting about this particular position is, if Garry Kasparov were to lose today's game, it's entirely conceivable this whole sacrifice and so on is just in Deep Blue's library, opening library, and it's done nothing -- it may turn out it won't even have to play an original move if Garry chooses one of the variations that it has been programmed as a win for itself. Which would be very unfair, not only to the Deep Blue team and its research, but to Garry Kasparov as well, because all he's doing is losing to analysis by his own colleagues. MAURICE ASHLEY: To be fair, though, Garry did not have to choose an opening that he doesn't normally play. I know a friend of mine, whenever I go into a big competition, my mentor, fellow name Willie Johnson, always says, "Maurice, play what you know." And it's good advice. It's served me in good stead. You go into situations that you're familiar with, instead of going into something new and you don't know what's going on, you start thinking for a long time as Kasparov is now. He's shaking his head. He's perturbed already. Already, the game just started. We are only on move 11, and Kasparov can normally whip off 15 moves in a flash, we're on move 11, and he's suffering already for no good reason. MIKE VALVO: You know, and it's exactly the kind of position that he didn't want the computer to have. MAURICE ASHLEY: How so? MIKE VALVO: Wild, complicated, tactical. He just didn't want this. And I wouldn't say that Garry's forte is defensive chess. He's a good defender, but he's a much better attacker. MAURICE ASHLEY: Right. MIKE VALVO: And the thing that occurred to me is the two games where we had extended opening lines, game two and this game, in both cases he used Anatoly Karpov's lines, which is kind of strange. Why not just play c5, like Yasser said, play your own stuff, go with what got you there, as you say. YASSER SEIRAWAN: What brought you to the party. MIKE VALVO: That's what we came to see. We didn't come to see him trying to trick the computer. We came to see him take the computer head on, and I had hoped that he would do it in this final game, he would realize that up to now it hasn't work, this is the time he has to do it, he's pulled all his tricks, now let's go with our strengths, your strengths to my strength. MAURICE ASHLEY: Is he so terrified by Deep Blue and what it might know that it's just completely thrown him off his game? YASSER SEIRAWAN: It seems -- MIKE VALVO: It seems so. He said yesterday "I'm not afraid to say I'm afraid." You said are you going to play powerful in the last game, and he said, "I'll play the best moves that I can." MAURICE ASHLEY: Garry is known for his bravado -- GK MOVE: 11...b5 MAURICE ASHLEY: He's going to put me at a loss for words in a minute, and that's not easy to do. b7-b5 has been played. b7-b5, Yaz. MIKE VALVO: Could we see the computer's clock? I want to see if this has all been book. YASSER SEIRAWAN: I think the computer has played very quickly, -- MIKE VALVO: I was wondering if they were out of book. MAURICE ASHLEY: Deep Blue has used only six minutes to play 12 moves. Kasparov has used 15 minutes, but the time spent was really on only one move and that was the last move, b7-d5. Every other move he actually played quickly. But h6 took him a little time to play, a little time, and then suddenly the sacrifice, Nxe6? A. This mover, Yaz, b7-b5, what's this about? Seems to me to develop -- he should be wanting to develop pieces, not push pawns. YASSER SEIRAWAN: What Garry has to do obviously is develop his pieces. If you look at his army, the two bishops on the back row, the two rooks, the misplaced black king means that Garry has to make a great deal of repairs to his position. I had just put on the board the move Nd5 that comes with tempo. Then the idea was Bg3 Qe7-b4, trying to develop the queen, preparing the knight retreat Nb5-e7. He wants to make sure the knight when it comes to d5 stays there, and there will not be the move c2-c4. So what he did with this last move, b7-b5 is to try to secure the square d5 for his knight so that he won't have to worry about the move c2-c4. MAURICE ASHLEY: But, Yaz, can this position tolerate more pawn weaknesses? His king on D eight won't find a home for a very, very long time. Of course the point of the game is to checkmate the king and the best thing you can do is have pawns behind it. These pawns look suspicious as defenders of this king on d8. I don't even see how the king will ever get back to the king-side. This could be a long trek indeed. b7-b5 seems a little loose to be playing in this kind of situation. YASSER SEIRAWAN: Well, Garry's got problems, and -- (Audience laughter.) Garry does have problems in this position. The first thing he's got to do is address his development. He hopes that he's not going to get run over in the center and his king. For example, we could try to understand Garry's hopes after the move Ne5 by Deep Blue. In general it's a bad idea to trade attacking pieces for defending pieces, so the move Ne5 is an unlikely choice. It does, however, threaten Ne5xc6 checkmate. So that after Nxe5 again we would see an unlikely choice in dxe5. DB MOVE: 12 a4 MAURICE ASHLEY: Before that could even settle itself as a positional concept, Deep Blue has instantly moved, seeking to undermine the b5 pawn with the move a2-a4, and that pawn will have to make a decision sometime soon. GK MOVE: 12...Bb7 MAURICE ASHLEY: Instead Kasparov has left the pawn on that square and has moved quickly with Bc8-b7. And now Kasparov is moving a little bit quicker than one would suspect in such a dangerous situation, Yaz. Bc8-b7 was probably seen by Deep Blue in response to its a4 idea. I guess he's just settled down, I guess he's just figured, well, I made a mistake, got to live with it, let's play chess. YASSER SEIRAWAN: Exactly. In this case he's going to have to live with it, so he's already reconciled himself to that. The idea of a2-a4 is the struggle. White wants to knock the b5 pawn out of the way so that c2-c4 becomes a possibility. At the same time, the move a2-a4 will open up the rook on a1 into active service, and the rook will get drafted. MAURICE ASHLEY: It's interesting to me, though, Yaz, what's very interesting about this position is that, if we look first at Fritz, the way Fritz is looking at the position, Fritz says right now that black only has a .28 advantage, .28 of a pawn. Now, one point is one pawn, that's 1.00 is one pawn. Now, mathematically, what's been programmed into Deep Blue is that a knight is worth three pawns, 3.00 and a pawn is one point. Now simple subtraction, that means that right now, black has the advantage of two pawns. Now, we see the compensation immediately. The bad king, the developed pieces for white. We see that white has tremendous compensation and could work to try to win the position. But what if Deep Blue sees the material disadvantage and thinks, for example, "Maybe I can win the e6 pawn back and start to equalize material somewhat, and maybe just play like Qe2, gang up on the e-pawn, and if Kasparov gets Deep Blue to take this pawn, which is what Deep Blue might want to do considering it's down material, he might suddenly be winning the game. YASSER SEIRAWAN: How many of us think that? MAURICE ASHLEY: My point is this. What would prevent, Mike, maybe you can answer this question. What would prevent Deep Blue from seeing the e6 pawn and just taking it if Garry leaves it that so that it can get closetory redressing the material imbalance? After all this sacrifice it played was not played on its own, on its own volition, it was programmed in. Maybe by now Deep Blue is thinking when the new moves started on the board "Who sacked my knight?" (Audience laughter.) YASSER SEIRAWAN: Good point. DB MOVE: 13 Re1 MAURICE ASHLEY: Re1, eyeing a weakness. YASSER SEIRAWAN: Which is what with you just pointed out. Maybe he wants to play Qe2 to gang up on this -- MAURICE ASHLEY: In fact Qe2 in this position practically wins a pawn because it attacks the b5 pawn and the e6 pawn. That would actually show a flaw in Deep Blue. MIKE VALVO: I would like to address this. I think in game two we saw evidence of reasons why Deep Blue will not settle for just winning that pawn. Remember that it didn't play Qb6? YASSER SEIRAWAN: Correct. MIKE VALVO: It could have won two pawns but it didn't do so because its king would be exposed. Now, it's aware of the other guy's king being exposed, too, and it won't settle necessarily just to win a couple of pawns when the king's exposure is worth more to it. I don't think it'll settle for a couple of pawns. Most computers would. I don't think this machine will. MAURICE ASHLEY: Well, that would be disastrous indeed if that did occur and it would show a flaw in the computer's estimation and valuation of the position. MIKE VALVO: If it did. MAURICE ASHLEY: If it did. At the moment Deep Blue has played Re1. Kasparov is thinking about how to finish developing his pieces. It's a very tricky task indeed. We should say to our in-house audience that we love it for you to participate. We will be sending ushers around with microphones to ask several questions, and we will do our best to answer those questions. MIKE VALVO: Who is going to win? MAURICE ASHLEY: We would also just like before we do that to welcome some students who have been invited by IBM -- every day IBM has given tickets for students to come and watch the games, and we have people from everywhere. First, I would like to introduce the ridge way public school from White Plains, New York. We give them a plan. -- we give them a hand. Who's champ of Ridgeway? Who's the best player in ridge way? Two hands went up. Can we get a microphone over to the two youngsters? I see two hands, so I'm going to have to get two names. What's your name? AUDIENCE MEMBER: Oscar. MAURICE ASHLEY: And next to him is also the champion of ridge way? AUDIENCE MEMBER: And the nation. MAURICE ASHLEY: They are -- AUDIENCE MEMBER: National champions. MAURICE ASHLEY: They are national champions? (Audience applause.) Who do you think is going to win today? AUDIENCE MEMBER: Deep Blue. And Oscar, what do you think? AUDIENCE MEMBER: Kasparov. MAURICE ASHLEY: All right! Okay, we've got two sides here. All right, we would also like to welcome Port -- are they from Port Washington school district? Port Washington? We also have kids from all over the Port Washington area, not a particular school. We would like to welcome you for coming. And as a group, who thinks Kasparov is going to win? Let me see some hands. And who thinks Deep Blue is going to win? Just one lonely camera. Okay. And we also would like to welcome community elementary school 70 from the south Bronx Bronx. Are you here? One of program C.E.S. 70, they have a strong coach, David McNuety. They have won many, many titles. Who is the champ here today? YASSER SEIRAWAN: Two hands. MAURICE ASHLEY: No consensus today? All right, still strong players. Welcome to students. We love it when kids come to watch. After all, they are our greatest fans. So welcome. (Audience applause.) YASSER SEIRAWAN: By the way, maybe you can help me on my chess history. We talk about Garry Kasparov as being the 13th in line of world champs. We go back to the original world champ, or the original recognized world champ, Wilhelm Steinitz, and I believe he was in a banquet with Zukertort, and it was a closing banquet of a great tournament, and the master of ceremonies had said something to the effect "Now, please, an applause for the world's best chess play," and both Zukertort and Steinitz stood up. And here we have it again, the best school player. MAURICE ASHLEY: Well, we have Kasparov deliberating on the position trying to figure out exactly what he was do. It's not easy to make a decision here because so much is going on, Yaz. YASSER SEIRAWA