Commentary Banner

---

Welcome to the commentary of the Kasparov - Deep Blue Challenge. What you're reading is exactly what Maurice and Yasser are saying at the match - thanks to the fine work of the stenographers at the scene. Please forgive them a typo or two and enjoy the match!!

Commentary by Maurice Ashley and Yasser Seirawan

---

Commentary and Analysis -- Game 3

Mr. Levy Good afternoon, ladies and gentleman. My name is David Levy, and I'm vice-president of the ICCA International Computer Chess Association, and on behalf of the ACM, I would like to welcome you all here this afternoon for the third game of the ACM Chess Challenge.

As I'm sure you all know, because there's been so much press and TV coverage, in the first game Garry was somewhat surprisely beaten very, very convincely by DEEP BLUE, and on the second game on Sunday Garry got his revenge in a very long game. It went 5 and three-quarter hours. So the match couldn't be more exciting. The score stands now at 1 point each. There are 4 games to go.

This afternoon DEEP BLUE is White, and the rate of play is 40 moves in the first 2 hours by each player. Then 20 moves in each subsequent hour and the rest of the moves in the last half hour. International Grandmaster Yasser Seirawan, three time international champion and international master Maurice Ashley, who has also helped with commentaries in the past all over the world including the most recent world championship in New York.

Mr. Ashley I'd like to welcome to the ACM Chess Challenge where world champion Garry Kasparov is really in the match of his life against DEEP BLUE, IBM's DEEP BLUE. So far a very entertaining match. 1 might say unlike the Kasparov match very slowly this 1 has just been all systems go right off the jump. So Kasparov losing the first game with the Black pieces. Now, after bouncing back, finding himself again with the Black pieces, and here in this game they have operated; that is, DEEP BLUE has operated to begin the same way. I mean if the serve worked the last time, why not play the same move. And the move here begins E4, while has come out with E4 and Garry Kasparov, I don't know why he has paused for so long. He has thought about the move for 2 minutes and played the response E -- C5 and DEEP BLUE has played C3, repeating the exact same moves.

Well, Yasser, how do you feel about this opening from the last game. The last time they played this opening was very successful for DEEP BLUE.

Mr. Seirawan Absolutely.

Mr. Ashley How do you feel about it in this game?

Mr. Seirawan Well, absolutely. I mean that's what chess is all about. I mean, if something worked, you keep at it until the opponent has come up with a neutralizing move. We have the same opening as in game 1. It was a Sicilian defense.

Deep Blue begins 1. e4 Kaparov responds 1. ... c5 2. c3 d5 3. exd5 Qxd5 4. d4 Nf6

DEEP BLUE began with E4. Kasparov replied with his favorite Sicilian move, C5, and then DEEP BLUE played the Alapin move, C2 C3, which is regarded in most circles as a conservative way of meeting the Sicilian. It doesn't go for all of those explosive opening positions that we've seen in the Sicilian, but it is a very good idea, I think, for the computer to get a few pawns off the board, to begin the game, because then it can put its great calculative skill to work, as we've seen. And, again, I had expected Garry Kasparov to try on his second move knight F6. Instead Garry Kasparov did in fact play D5. E takes D5, queen takes D5. D4 knight F6. And that's the position on the board as we see it now. Again the exact position as game 1. So clearly Garry Kasparov has worked on his opening. He's got an anecdote prepared and we'll be seeing it shortly.

Mr. Ashley Well, it's interesting that DEEP BLUE is hesitating on this particular move in this position. The last time we saw this in the first game of the match on Saturday DEEP BLUE wiped out the first few moves, no thought to the matter, just played all the moves. And now we're seeing this move here, this position here, the computer giving pause on this move after knight to F6, but we've seen it give pause in obvious situations before, haven't we?

Mr. Seirawan Yes.

Mr. Ashley We've seen it think about its moves in situations which we consider very, very obvious, and it does that from time to time.

Mr. Ashley I am just checking the move order here. We will be fielding suggestions and questions from the audience. We should let you know that This is going over the Internet, the World wide web, so if happen not to be here you can check it out, find our comments over the Internet.

Mr. Seirawan There are several web sites. The big web site of course is at www.ibm.com, something or other. And then of course there's the USA Today world wide web site that also features a lot of graphics. I spoke just in coming here with the IBM team that is putting together their world wide web site

Of course, chess is such a global game that a lot of people who can't be with us here in the Philadelphia -

Mr. Ashley Yasser, how do you explain the interest in this match, the buzz, the excitement? It's like it's a world championship match. And even the world championship -- well, the beginning of the match with Anand and Kasparov really saw a lot of press interest. But here we are really seeing what you won't expect at first blush. Kasparov against the computer more like an exhibition match, but this is really like you're seeing

Heavy weight boxers. People are really into this match.

Mr. Seirawan As we saw in the first game of the match, this is an exact repeating of the first game. And the first game, a few moves down the road from this point, Garry Kasparov introduced novelty bishop B4. And we've no other move by DEEP BLUE and no other move by Garry. DEEP BLUE did repeat as in game one bishop to E2, and Garry very quickly responded E6. And then they seemed to be having lots of conversations. And I notice that we have got a new programmer, a new operator, Murray Campbell replacing Feng. I think maybe Garry had asked for a relief pitcher.

Mr. Ashley Well, after E6 DEEP BLUE played castle.

Mr. Seirawan Very quickly.

Mr. Ashley And Kasparov, with knight c6 and now bishop to e3.

Mr. Seirawan And again a direct repeat of the first game. And here we saw what Jay's question was. That after the further move c takes d4, e takes d4, as in game one, which they haven't yet played, Garry came out with a move bishop b4. And Jay's question was from perspective of a grand master. This was an unusual move and novel idea. And what was my insight into the move? Frankly I liked the move. It worked out very well. I agree with C.J. that after a3 it looks he just gave white an extra tempo.

But what Garry was trying to do was keep the e7 square in reserve. And as we saw, we saw that in game one Garry eventually used the e7 square for his queen instead of his bishop. Garry was absolutely fine throughout most of game one, and that was when he decided to go on the attack, and it turned out to be unjustified, and in his own words was appropriately massacred. But if Garry were to repeat everything as in game one I feel he had a fine position through most of the game.

Mr. Ashley And for him it would be very, very pleasant I would think because he would save so much time on the clock if he really had the full confidence of the moves and decisions he made in that game. No doubt he has already taken a close look at that game with whatever trainers he has here.

Mr. Seirawan Absolutely. One of the things people have overlooked is the advent of the computer and especially in the last 2 decades we're seeing enormous, enormous number of computer programs.

You can run out to your local electronic boutique and pick up a software program and those software programs have been killing, just destroying their human competition. And I think the software programs that are readly available in stores today just for 50 odd dollars will beat 99.99 percent of all humans.

And so here what we're seeing is all of those people who bought all of those programs and been defeated by their computers are now saying, well, who's the last bastion? And this is what's happening is you have ten million, a hundred million people getting clobbered by their chess computers and they want to see the computer get clobbered by the human. Mr. Ashley Maybe they also want to see the human get clobbered. Make him seem a little normal. We should say here that we are using 1 of those excellent software readly available software programs to help us analyze, that is Fritz 4, one of the genuinely strongest in the market, a move has been made by the computer, the move knight to F3, after spending about 4 minutes, DEEP BLUE has played knight to F3. And now something has happened. You know we've seen the human operator prone to error in this competition. He's pulled back his knight, the knight back from F3 to G1. We saw it the other day the human operator excellently played a horrible move for the computer and it turned out that -- and Kasparov was -- and now he has indeed played knight F3, made the actual move knight F3.

I don't understand why that happened. But the human operator is now explaining to Kasparov why this occurred. It's funny that it's the human that makes the mistake most often. Deep Blue plays 5. Nf3

Mr. Seirawan Well, in this particular case I didn't see the monitor, but it seemed to me that doctor Feng who had put the move knight F3, he was probably just looking at the screen and somehow saw that DEEP BLUE in all of its calculation was ready for the move knight F3 and made the move. You owe. There's something going on on the screen and Garry has quizzed and we've got Mike Valvo, international master Mike Valvo the arbiter speaking with the 2 players. And the programmer seems to have been properly chided. Kasparov plays 5. ... Bg4

Mr. Ashley What I've noticed, though, with Garry is his -- you know, he's really been good about everything. Now, he's played the same move as in the other game, the move bishop to G4. He's really been good about any mistake by the operator. Hasn't blown up and made a stink about it. The last time for example when the computer played that horrible -- when the operator played a move that was actually not the computer's move and it was a really horrible choice Kasparov was very very happy and then suddenly the human operator really iced his owner or and informed Kasparov that they had to retract that move.

Now, some players around the world would have made a big stink over that. Kasparov was fine. He said no problem, you know, all right, let's go on. And he just -- he got himself back into the game and got a great position.

Mr. Seirawan I agree. I think that's very difficult. Just put yourself in Garry's shoes for just a second. What we saw was that the computer operator had made a move on the board that was wrong. That was not what DEEP BLUE had intended. And in fact it was such a good mistake that when Garry pounced he was really really lap I about it and then immediately he said I'm sorry we're going to have to take the last few moves back. And it really could throw you off your railway them. And you're absolutely right the manners of Garry have exceptional. He's done everything he can to cater to the IBM team. I know the gentleman has been raising his hand and I apologize I didn't get back to you. Yes, sir?

(Audience commentary.)

Mr. Ashley We will be fielding suggestions and questions from the audience here. We should let you know that this is going over the internet, the worldwide web. So if you happen not to be here, you can check it out, find our comments over the internet.

Mr. Seirawan There are several web sites. The big web site of course is IBM's web site, at IBM, Chess. And then of course there's the USA Today worldwide web site that also features a lot of the graphics and things that the IBM is showing. And then there are many chess sites on the worldwide web, including TCC Dot Net. That's GMT's web site. I spoke just in coming here with the IBM team that is putting together their worldwide web site, and they had explained to me they're getting so many hits, somewhere in the neighborhood of 6,000 hits a second. And they've been only able to accommodate a few thousand hits a second. So there's a lot of overflow. And IBM has told me that we're talking about maybe 1 million to 2 million hits a day on the worldwide web site. So now they're expanding their band window. They can accommodate more people because they underestimated the damage to follow the match. And, of course, chess is such a global game that a lot of people who can't be with us here in Philadelphia --

Mr. Ashley Yasser, how do you explain the interest in this match, the buzz, the excitement? It's like it's a world championship match. And even the world championship -- well, the beginning of the match with Anand and Garry Kasparov really saw a lot of press interest, but here we're really seeing what you wouldn't expect at first. At first blush Garry Kasparov against the computer more like an exhibition match, but this is really like you're seeing heavy weight boxes. People are really into this match. 6. Be2

Mr. Seirawan The judge said that as we saw in the first game of the match this is an exact repeating of the first game, and in that first game a few moves down the road from this point Garry catch introduced the novelty bishop B4. And we have another move by DEEP BLUE and another move by Garry. DEEP BLUE did repeat as in game 1 bishop to E2 and Garry very quickly responded E6. And again they seem to be having lots of conversations. And I notice that we've got a new programmer, a new operator, ACM BELLE replacing Feng. I think maybe Garry had asked for a relief pitcher. Kasparov plays 6. ... e6 Deep Blue responds 7. O-O

Mr. Ashley Well, after E6 DEEP BLUE played castles.

Mr. Seirawan Very quickly. Kasparov 7. ... Nc6 Deep Blue instantly plays 8. Be3

Mr. Ashley And Garry Kasparov with knight C6 and now bishop to E3.

Mr. Seirawan And again a direct repeat of the first game. And here we saw what the gentleman's question was, that after the further move C takes D4, E takes D4 as in game 1, which they haven't yet played Garry came out with a move bishop B4. And the gentleman's question was from the perspective of a grandmaster this was an unusual move and novel idea and what was my insight into the move. Frankly I like the move. It worked out very well. I agree with the gentleman that after A3 it looks like he just gave White an extra tempo. But what Garry was trying to do was keep the E7 square in reserve. And as we saw, we saw that in game 1, Garry eventually used the E7 square for his queen instead of his bishop. Garry was absolutely fine throughout most of game 1 and that was when he decided to go on the attack, and it turned out to be unjustified and in his own words was appropriately massacred. But if Garry were to repeat everything in game 1, I feel that he had a fine position through most of that game.

Mr. Ashley And for him it would be very pleasant I would think it would be very pleasant because he would savor so much time on the clock if he really had the full confidence of the moves and decisions he made in that game. No doubt he's already taken a close look at that game with whatever trainers he has here. We haven't been informed of any of the trainers who are with him, but normally he carries with him a very large entourage.

Mr. Seirawan Right.

Mr. Ashley And no doubt he's already looked at that game in minute detail, decided look if this position were to occur again I'd be satisfied because obviously here he's playing just those moves. He has to wonder for a moment if DEEP BLUE has anything else up its sleeve. But if he just plays the moves automatically, you know, with of course good time for reflection as any professional player would do, I would imagine he would save a lot of time on the clock and be able to make decisions later in the game at a better pace and with much more time for reflection.

Mr. Seirawan Indeed. But I'd just say this once again, and I've been repeating it throughout the whole match. Garry Kasparov is a phenomenal player, especially in the openings. He has brought more new ideas, novel ideas to chess than virtually any other world champion. And just to buttress that point, he's written so many books about openings. So the opening phase of the game has never been his problem. He's always quick in his openings. He knows very well the missile game ideas and especially the end game positions that result from the openings he plays.

He's very well prepared. So normally speaking we do see Garry playing very quickly in the opening and already DEEP BLUE has taken 9 minutes, I think; Garry 7.

Mr. Ashley That's correct

Mr. Ashley And the moves have passed. So a very basic pace.. Kasparov captures 8. ... cxd4 Deep Blue recaptures 9. cxd4

Mr. Seirawan And I'm sure Garry was expecting this position and he's repeated once again game 1. C takes D4, nice little capture. This -- and DEEP BLUE very quickly recaptures C takes D4.

Mr. Ashley And a big moment Yasser. Bishop to B4. Will we see this move bishop to B4 which we saw in the first game.

Mr. Seirawan Well, Garry has always been a great champion of his own ideas. And I don't think he's going to back down. He might choose something a little bit different just to put the computer off a bit.

But I suspect we're going to see bishop B4 again. And once again if we do see the game going all the way to that critical position as in game 1 where Kasparov made king -- where Kasparov made king H8 I think Garry will have a fine position and I know Hans Berliner 1 of our resident experts and I really want to thank him for being with us in the audience. I'll call upon him to give a comment.

Hans.

(Audience commentary.)

Mr. Ashley There is a slight difference. The bishop is on G4. H3 bishop H5 has actually not been played yet. So this specific position from the last game is not here. What are the differences? Well, you know a slight difference can be huge in a chess game. So that is a valid point, that just the fact that H3 and bishop to H5 has not been thrown in as yet could be extremely significant. Right now I don't see how --

Mr. Seirawan Well let's talk about that for a moment. Hans said speaking as a computer -- did I pick up on that right?

(Audience commentary.) Kasparov plays 9. ... Bb4

Mr. Seirawan Speaking as a computer scientist then. Okay. What he would like to say is that in that game 1 DEEP BLUE played the move H3. There it is. Kasparov in fact has played bishop B4. And if we see the inclusion of the moves H3 and bishop to H5 which we haven't seen. 10. a3 Ba5

Mr. Ashley A3 instantly by DEEP BLUE. Deep Blue 11. Nc3

Mr. Seirawan And bishop B5 by Garry also very quickly. I'm sure he didn't even write down the moves. He just played it very quickly. And again DEEP BLUE's very happy. Knight C3, attacking black's queen. And Garry also very quickly is preparing to drop his queen back 1 square, queen D6. Kasparov responds 11. ... Qd6

Mr. Ashley What's interesting, what's intriguing is that DEEP BLUE -- whoa.

Mr. Seirawan Now this is --

Mr. Ashley Well, let us give pause for a moment. 12. Ne5

Mr. Seirawan Knight to E5.

Mr. Ashley Because the new move apparently ostensibly just for a second, a sacrifice, knight to E5 has been played by DEEP BLUE, and I'm sure as we calculate for just a little while we realize that this is not a true sacrifice because usually the computer would not give away material. But this looks like if I may say, Yas, some special home brew, indeed. It seems like DEEP BLUE's been burning the midnight oil. But I'll go back just for a second because we talked about something very very important that could have -- as you said it's subtle, but it could have a major impact on the game. And that is DEEP BLUE's pawn on H2 remains on H2. He did include the move H3, bishop H5. The move H3 is what we would call making Luft, Luft being the German word for air. The move H3 creates the -- it gives White's king on G1 the H2 square. So the inclusion of the move H3 should have been to White's advantage. And I'm a little bit surprised in fact that the move H3 wasn't thrown in prior to the move knight E5. Now let's look at what has happened with the move knight E5. Whenever you see a move like this, the first thing you've got to calculate is captures. So let's give it a shot, Frederick. Let's try the move knight takes E5 I assume you would have to see D takes E5 and then we have a move like queen takes D1 and again I assume the move bishop takes D1 is virtually forced and then we'd have a move for example like -- I like the move bishop takes C3. And then we would probably see B Takes C3. And then bishop takes D1. And now E takes F6. And in this series of exchanges, this is going to work out into DEEP BLUE's favor because what has happened is black's bishop on D1 is under attack as is his pawn on G7. So Garry will not be able to do this particular sequence. Should we go back just for a bit. Let's throw the move -- now we're back in the game position and let's assume Garry plays bishop takes E2. How does Fritz like this? Much better. Fritz in its very quick calculations also works very, very hard and very fast, considers that black has an advantage after the move bishop takes E2.

Mr. Ashley But Yas, I mean, come on. How could DEEP BLUE in the space of 2 seconds -- there was no lag time when Kasparov played queen D6. It instantly throughout this move knight to E5, probably quicker than even the operator had time to put it on. I mean the operator came back so quickly with a move that the moment Kasparov played queen D6 DEEP BLUE seemed to have this response knight E5 ready prepared. I mean it's got to be something going on here that is deeper than we are envisioning.

Mr. Seirawan Well then Fritz has horizon. Let's assume that we can go a little bit further in the variation, knight takes E2. Let's try knight takes E2. And now if the move knight takes E5 -- we'll just put that on the board for a moment. D takes E5. We now have -- if you play for example queen takes D1 rook takes D1 White's going to have a slightly advantageous position because after queen takes rook takes white has an advantage in development. So it's only natural that white will -- that black would like to take the pawn. Queen takes E5. And now I think this is the sacrifice that DEEP BLUE has prepared. He's prepared to play queen A4 check, putting a little bit of annoying pressure against black's king. If black's king has to move then, black will have given up the opportunity to castle. What would Fritz have -- knight D7. A sensible move. And now this is a long term sacrifice it looks to me here Maurice. Now we're going to play rook D1, rook F1 D1 and keep black's king firmly in the center.

Mr. Ashley I got to tell you I like this kind of sacrifice. I would give this pawn up in a second if I knew I could keep the world champion's king in the middle of the board pin down his knight have a tremendous lead in development and again I don't even see how 1 would properly defend this position because it's so difficult to get the king out of the middle in this variation

Mr. Ashley This would be great home analysis.

Mr. Seirawan Yeah, this would be very nicely prepared. And also keep in mind that the Black queen is stuck to the defense of the bishop on A5 for another move or 2. We'd probably see rook D8. I don't think that there's another move in the position. Well, the immediate sacrifice rook D7 probably doesn't work, but even just a move like D4 -- pardon me rook D3, I've got to feel that Garry is not going to be a happy camper.

Mr. Ashley No, not at all. There are other moves in that position. Rook A1 is playable, even containing the A7 pawn just to get the pawn back. You know, you could do it. You wouldn't but there are just too many dynamic possibilities for white in that position. The tricks would just become too tough to handle in a situation like that. I don't think Garry Kasparov wants to gets -- I'm sure Garry Kasparov does not want to get in a position like this. So the question is how can you avoid this situation.

Mr. Seirawan Well, what we're analyzing here -- by the way, that's what we'll be doing throughout, is just analyzing all kinds of possibilities down the road. Let's back to the position at hand. And I would say this is the way I would treat the position as a grandmaster.

First of all, what always happens in grandmaster chess is when somebody comes up with a new idea like what we're seeing now with knight E5, the first thing I do is take a deep breath. I go, okay I've walked into it. Now let me get my bearings back. Let me feel the position. And I would take 5, ten, fifteen minutes just calming myself down and analyzing all the crucial variations. The next thing I'd start to look at is what are the key elements of the position. The key elements of the position can be summed up very simply White has a safe king. 1 of the crucial things in the opening that you want to come out of the opening with is a safe king. Black conversely hasn't castled. So I'll start thinking to myself how can I get myself into a situation where I can quickly castle. The second thing I have to say is the active pieces. White's pieces control more space. So I look at a way of eliminating them. 1 of the first things I'd look at is I'd look at the move I don't like my bishop on A5. So the first thing I'm going to do is get that piece into the game. Let me try the move bishop takes C3. This seems a little bit nicer to me. How does Fritz like this? He's still giving -- Fritz is still calculating an advantage for Black. Kasparov responds12. ... Bxe2. Deep Blue 13. Qxe2

Mr. Ashley Well, Garry Kasparov has quickly responded instead with the move bishop takes E2, which is actually not a bad exchange. It's a sort of preamble to your move except there is that 1 possibility of playing knight takes E2. Now that avoids the exchange of bishop for a knight on C3. Instead of playing knight takes E2, DEEP BLUE has instantly responded with queen takes E2, which is definitely a sacrifice.

Mr. Seirawan And a surprising 1 again.

Mr. Ashley Well, let us begin to analyze these variations, Yas. Knight D4 looks like a very bad move in this position. If Black were to play knight takes D4 in response.

Mr. Seirawan Let's explain why. After knight takes D4 again that bishop on A5 is left vulnerable after bishop takes D4, capturing the knight, queen takes D4, the old double attack queen B5 check would win the game for DEEP BLUE. Because Black's king would be under attack as would the bishop on A5. So snapping off that pawn in the middle of the board knight takes D4 is absolutely wrong.

Mr. Ashley Let's take a look at maybe knight takes E5. What do you think about this option?

Mr. Seirawan Knight takes E5 is nice but I'm still a little bit worried that I'll leave this B square open. Would I prefer to get rid of that knight on C3. Let's go back just a moment. I would prefer to play bishop takes C3. I think DEEP BLUE is going to have to retreat that knight with knight C4, but I don't know. If you played B takes C3 -- let's just take a look at B takes C3, then knight takes E5, D takes E5 -- oh, I see. There was this move bishop F4. I think a gentleman in the audience has suggested it before. Yes, you couldn't play D takes E5 because queen takes E5 just snaps a pawn for no compensations, but there is this move bishop F4. Let's take a look at that.

Now we're going to get into a very interesting end game where if you -- in this analysis if you look at White's pawns in the middle of the board, the C3 D4 pawns, well this structure is known as the hanging pawn structure. Hanging pawns are wonderful in middle game positions. But they're weak in ending positions. And after bishop F4 you'll probably see a move like knight F3 check. Okay. Fritz preferred knight D7, but knight F3 check. Queen takes F3, queen D5 and now we have a position that -- well, has mutual risk for both sides. Again, Black is a little bit underdeveloped. White's center is kind of nice, but that pawn on C3 has a potential of being vulnerable in an ending.

Mr. Ashley I don't mind having hanging pawns. I tell you I play isolated pawn positions. I love the space they give you. Hanging pawn positions also give you that potential for the pawns to rush down the board and maybe sometimes even break in and get a nice attack going. I love the attack. And I tell you I love the attack so much that I like having a bunch of pieces to follow behind the horse.

Mr. Seirawan That's right.

Mr. Ashley In this position so many pieces are off that what is White hoping for? All I see here are weaknesses. I see the C3 weakness. I see the A3 weakness. I see the C4 square. And I got to tell you if all those pieces can come off the board I wouldn't trust White to win this game. I don't care who it is.

Mr. Seirawan Exactly. Well, I'll just buttress what Maurice was saying. Maurice is a very dynamic player and he enjoys having that initiative as does Garry Kasparov. I myself, I enjoy -- I'm more of a counter puncher

Don't mind giving my opponent a little rope so he can hang himself in this position that rook is going to get together because what's going to happen is in this type of ending Garry will just clobber the computer. So I think we're going to have to go back. We're not going to see this variation. And I know a gentleman has his hand raised. Yes, sir?

(Audience commentary.)

Mr. Seirawan What we're looking at right now in the middle of the screen is a software from Germany called Fritz, and this is their very latest. I don't even think it's commercially available yet. Is that right, Frederick?

Frederick Freidel, who works with Fritz, informs that in another month's time it will be commercially available. I think we're getting the beta test version, the top of the line model.

So again we have the position at the board, and what White will do after bishop takes C3. If I were in Garry's shoes again, what I would say is the bishop on A5 is a little bit out of play, and I'm happy to eliminate the knight. Furthermore, just what Maurice was saying, because of the pawn structure on the board what Black's ultimatum should be is to try to trade off as many pieces as possible.

I thought we would see this move knight C4 queen B5. B takes C3 castles. After this sequence of moves, again I like Black's positions because those hanging pawns, those pawns on C3 and D4 are nice when there's a lot of pieces behind them, but they can be easily weakened when you get into an end game. It was better to have that knight on the board and better not to exchange off the E5 for White.

Mr. Ashley In a position like this it's really White with the weaknesses and no coordination. Black's pieces would be extremely well centralized.

Mr. Seirawan Exactly. And I'd like to emphasize that to the audience here, if you take a look at the position that we have on the middle screen, look at those beautiful squares that Black's queen controls. The Black queen on D5 controls the E4 F3 G2 square. It