Commentary for black move 17:

YASSER SEIRAWAN: Let me just suggest from the point of view of a professional Grandmaster for a moment. First of all I think in terms of this event as a sporting contest, that is to say you've got a prize fund of 700,000 to the winner, $400,000 to the loser and I think of it as a -- DB MOVE: 17...a6

YASSER SEIRAWAN: Secondly, I think that Garry knows very well how computers play. And it also led me to believe that you're constantly experimenting with the program. Suppose Deep Blue indeed had been given to Garry, Garry spends a week with it, discovers that it could -- that he, Garry, could defeat the program in such-and-such an opening, the Deep Blue programmers in the week had improved the program, he comes to the game and he doesn't win the way he was supposed to, and he says, "Hey, you guys gave me a bum dash a bum program." So it's not working that way.

MURRAY CAMPBELL: There is an issue after we've given him the program, are we not allowed to improve it -- AUDIENCE MEMBER: I don't mean give the program, I mean show a test of one game that it would make, so that, if, for example, if I open a4, it will move -- not that it does it all the time, but it did in that case.

YASSER SEIRAWAN: Pardon me, Mike, we do have a move.

MIKE VALVO: a6.

YASSER SEIRAWAN: a7-a6.

MIKE VALVO: By the computer, and it looks like the computer is getting ready to maybe break with b5, or prevent any idea on white's part to play Nb5 at some point.

YASSER SEIRAWAN: Well, it's surprising because if the move Kg8-h8 had been a pass, the move a7-a6 was certainly not a pass. This move does weaken the b6 pawn so that if white were to playing, for example, Nc3-a4, or indeed Nc3-b5 at some future moment, there may be a weakness here (d5.) Also we see that white's rook on b2 is quite good because after an exchange, the b-file will be open.

MIKE VALVO: I think we've got to let Murray go.

YASSER SEIRAWAN: Murray Campbell. Murray, thank you very much. I know the gentleman was making his point, and I think you've made --

MIKE VALVO: I was thinking about his question. At what point do you stop letting Garry see everything? Garry could just go into the a room with the computer and they could play each other not -- all the time not a game but a series of variations, at some point you say let's stop and play a game. That's what we're doing today, we're stopping and replying a game.

YASSER SEIRAWAN: I think Garry is taking this match very seriously, has sat with his own computer experts and certainly the Deep Blue team has a lock on the computer secrets over the Internet and to the entire world -- so I am sure we are seeing a very well prepared Garry Kasparov in this match. I'd like to just talk about the game position for a moment because we do have some moves. It has been very much a cat and mouse game for some time. We saw this pass, this null move by Deep Blue. This was reflected also by Garry with his move Rb2. The move a6 is quite committal. It has weakened the queen-side a bit, but as Mike Valvo /KHR*EFRL pointed out it does make --

MIKE VALVO: /KHR*EFRL! Wow.

YASSER SEIRAWAN: /KHR*EFRL pointed out b5-b5 -- b6-b5, a blind man could still find a seed. I think Garry is going to have to take some concrete steps here to move the position forward. His problem is he would dealer love to play the move f3-f4 so that he could get this pawn so f5, but at the moment, of course, the bishop on h5 pins the pawn. If you were to play, for example, g4 here, bishop back to g6 and then play the move f4, then of course the problem is that after exf4, Bxf4, this pawn advance hasn't gotten Garry what he wanted, which is this pawn on f5. And that was why we saw earlier the move c5 is very important, reinforcing the knight. So what I had thought --

MIKE VALVO: What do you think? Do you think black is okay here?

YASSER SEIRAWAN: No, I like white. I like white.

MIKE VALVO: What does Fritz think here? Fritz just thinks that white's ahead a little tiny bit.

YASSER SEIRAWAN: A little tiny bit. But I actually do like white. This bishop on e7 would bug me. It's not a bishop that I like.

MIKE VALVO: Besides we know computers think like each other.

YASSER SEIRAWAN: Yes, exactly. They enjoy supporting each other. And I would have preferred that Garry take some active measures for a moment with the move h2-h3. One of the dreams of white is to jump his knight from c3 to d5 but that would be ineffective because if he were to do so immediately the computer would just exchange the knight and there would be no benefit from it, so I like the move h2-h3, with the intention of Nf2-g4. That would make the possible exchange of knights so that the knight on c3 could move to g5. After h2-h3, let's just make a random move if you don't mind for just a moment. Let's take back a move by --

MIKE VALVO: That looks pretty constructive, too.

YASSER SEIRAWAN: Taking back the move. Nf2-g4. This feels right because the knight has made Rd2-f2 possible. Again, forgive my bad play on behalf of Deep Blue for just a moment. Let him pass once again. Then we could see a benefit of this strategy by white, Nxf6 Bxf6 Nd5 and white has achieved a lovely goal. He's attacking the bishop on f6, he's attacking this weak pawn on b6. You might see just a continuation of a bad variation, Bd8, the queen could step up to d2 and now the possibility of f3-f4 begins to loom. So all this means is that I think it's time for Garry to take some concrete decisions, and I think h2-h3 is the best way to do it. Otherwise it's very awkward for white to unravel his pieces.

MIKE VALVO: That still don't seem so overwhelming.

YASSER SEIRAWAN: Oh, no.

MIKE VALVO: Nobody is winning any material, black is forced back a little bit, the black squares might become weak for white.

YASSER SEIRAWAN: There's a terrible myth in the public's imagination about chess, and that myth is, gosh, yeah, you're an International Grandmaster, and I just know chess a bit. You could probably beat me in two moves.

MIKE VALVO: That's not true?

YASSER SEIRAWAN: No. I can't. The game starts from a position of two armies that are in equal equilibrium, they're balanced, and it takes a long time to win a game of chess. We've got to accumulate small advantages. And Garry's not going to win this position with just one simple master stroke. Oh, no, no, no, no. It's going to be an accumulation of tiny advantages. And one tiny advantage is this bishop on e7 is a little bit passive. The square d5 can be occupied once we get rid of that knight on f6. So Garry has to work the position a bit. An advantage here, an advantage there. And then finally everything looks like a beautiful chain interlocking all the moves --

MIKE VALVO: I was just wondering about your question you asked Murray here, aren't you concerned about -- I wonder if Garry feels the same kind of way, he would not like to play this -- -- doesn't he want to go out there and let's nail my shoes to the floor and nail his shoes to the floor, or let's just box it out until one of us drops?

YASSER SEIRAWAN: Well, yes, you just described it beautifully why you can't do that. Let's take the analogy of boxing. We've got two opponents that are mismatched. One is a great boxer with wonderful, fluid ballet-like movements around the ring. The other is just this brute animal that will just whoom, whoom. Guess what she, you don't trade blows? You'll die.

MIKE VALVO: So you run around the ring.

YASSER SEIRAWAN: So you run around the ring. So even a great calculator, perhaps the best in the world, Garry Kasparov, it doesn't try to out-calculate the machine. At least that's the proper --

MIKE VALVO: Won't it be a great game?

YASSER SEIRAWAN: I don't know. It might be a wipe-out by one side of the other. It might be a quick game. You won't see any long decisions.

MIKE VALVO: If Garry keeps resigning like this, it will be really quick. (Audience laughter.)

YASSER SEIRAWAN: Let's take a couple questions from the audience with the gentleman, waiting for a mike for a moment. By the way, time for a poll, Mike?

MIKE VALVO: Sure, let's have a poll. How many think that Garry is going to win today?

MIKE VALVO: And how many think the computer is going to win?

YASSER SEIRAWAN: How about a draw? It's nice to say -- the gentleman with the microphone. AUDIENCE MEMBER: It's a factor at this point. Down by about 30 --

MIKE VALVO: Well, maybe in a second they'll show you what the --

YASSER SEIRAWAN: Garry has 41 minutes.

MIKE VALVO: It says on the block. 41 minutes, for 18 moves, quite frankly, yes, I would be concerned (Seirawan) as a human player. Normally speaking -- uh-oh, he's shaking his head. It's like he heard us. Now he's holding his breath.

MIKE VALVO: He's got 43 moves to make.

YASSER SEIRAWAN: No, he's get 23 --

MIKE VALVO: That shows how many he's made.

YASSER SEIRAWAN: I see. The point is as a professor Grandmaster, normally I'm playing my first ten to 15 moves quite quickly usually within ten to 15 minutes and my real thinks come after the 15 moves to say 30, and that's when I really tried to /STPEPBD a lot of time, and then I catch up a little at the end. So Garry is not a/TPORGD himself the standard method of playing, which is to take a lot of time in the middle game. In fact, he's going to have to accelerate his play considerably, and definitely that is going to be a factor. Thank you for pointing that out.

MIKE VALVO: Do you know what time it is?

Real-time text commentary is made possible by LiveNote, Inc. and Vincent Varallo Associates