(1) Paul Hoffman (1950) - Garry Kasparov (2830) [B43]
ZMD Simul Game, 25m + 10s Simuls with titled players, 04.04.2003

Comments by Paul Hoffman except where noted.

Two days earlier, I had watched Garry play a 24-board simul at the New York Stock Exchange. He disposed of everyone in under two hours, and he did it by playing solid, classical chess. He tried to build up an e4, d4 pawn center. He never played risky lines. He avoided theoretical opening disputes. Indeed he avoided most mainlines, preferring to deviate early and throw his opponents on their own.

My first goal was not to embarrass myself: I did want to play worse than I normally play. More ambitiously, I hoped the game would be sharp because even though he is a great tactician I thought that if my simul-mates could keep him occupied a bit on their boards, he might not have time to work everything out. At the National Open in Las Vegas I had drawn Yasser Seirawan in a sharp game in a simul. I was the last man standing against Seirawan and I would be the last man here, but the result would be different.

Because my chutzpah Elo is much higher then my USCF rating, to prepare for the game I e-mailed the few elite GMs who had fared well against Garry and asked for their advice. Vladimir Kramnik did not respond, presumably because he didn't want to reveal any trade secrets before a possible rematch with Kasparov (or else because he didn't understand my sense of humor).

Boris Alterman, the Israeli GM behind Deep Junior, was more forthcoming. He noted that in the recent six-game match with the silicon monster Kasparov had punishingly advanced his g-pawn in two of the encounters. He advised me to eliminate Garry's lethal g-pawn as soon as possible, even at the cost of significant material. He assumed that I would play Black and suggested an offbeat variation of the Scandinavian: 1. e4 d5 2. exd5 Qxd5 3. Nc3 Qxg2!! I studied the variation but feared that Garry might slyly vary with 3. Nf3, blocking the assault on his coveted g-pawn.

But my opening preparation was for naught when I found myself playing White. I had also asked other strong players who knew me personally for their advice. GM Alex Baburin, editor in chief of Chess Today, told me to find the best moves and not think about who my opponent was. Good advice -- particularly because I would not see be able to see him online -- but easier said than done. It's hard to "forget" that you're playing the world number 1! Other friends were very supportive: "Ha ha!" IM Greg Shahade told me, "You will lose!"

1...c5 2.Nf3 e6 3.d4 cxd4 4.Nxd4 a6 5.Nc3
[ I knew that 5.Bd3 is a more flexible move, leaving open the possibility of c4. I remember that Fischer played 5.Bd3 against Petrosian in their candidate match (That's how old I am). But I've had no experience with 5.Bd3 and didn't want my first time to be against the strongest player in history.]

5...Qc7 6.Be2 b5
[ I was hoping that Garry would play 6...Nc6 which I know a little about. Indeed, I reached this position a few days before in a G/30 tournament at my local chess club (Middletown, NY) 7.Be3 Nf6 8.0-0 Bb4 9.Na4 Bd6 10.g3 b5 11.Nb6 Rb8 12.Nxc8 Rxc8 13.a4 Nxd4 14.Bxd4 e5 15.Be3 Bc5 16.Bxc5 Qxc5 17.axb5 axb5 18.Ra5 0-0 ( 18...Rb8 1-0 Shirov,A-De la Riva Aguado,O/Andorra 2001) 19.c4 Rb8 20.Rxb5 Rxb5 21.cxb5 Rc8 22.Bd3 Qb4 23.Qe2 d6 24.Rd1 g6 25.Qd2 Qxd2 26.Rxd2 Kf8 27.Rc2 Rxc2 28.Bxc2 Nd7 29.Kf1 Ke7 30.Ke2 Kd8 31.Kd3 ( stronger is 31.Bb3 f6 32.Bg8 h6 33.Bf7 g5 34.Kf3 which enables me to play on both sides of the board) 31...Kc7 32.Kc4 Nb6+ 33.Kb4 f6 34.Bb3 g5 35.f3 h6 36.Bf7 Nd7 37.Ka5 Nb6 38.b3 Nc8? ( stauncher resistance is offered by 38...Kb7 blockading the pawn and keeping my king out of a6. I can probably still make progress by sacking the advanced b-pawn at the right moment.) 39.Bd5 Nb6 40.Bc6 Nc8 41.Bd5 Nb6 42.Be6 Na8? ( again 42...Kb7 ) 43.Ka6 Nb6 44.Ka7 h5 45.h3 d5 46.exd5 f5 47.d6+ Kxd6 48.Kxb6 Kxe6 49.Kc7 g4 50.hxg4 f4 51.gxf4 h4 52.b6 h3 53.b7 h2 54.b8Q h1Q 55.Qe8+ Kf6 56.Qxe5+ 1-0 Hoffman,P 1927-Almeida,S 2145/Middletown G/30 2003]

7.0-0 Bb7 8.Bf3
I wasn't crazy about blocking my f-pawn with my bishop but I was nervous about ...b4, kicking the knight from defense of the e-pawn. The alternative 8. a3 also seemed passive. Later I learned from Chessbase that strong GMs (Khalifman, Svidler, S, Polgar and Kasparov himself) who have reached this position prefer [ 8.Re1 evidently not fearing 8...b4 because White apparently gets a strong attack after 9.Nd5 exd5 10.exd5 ]

At this point I started worrying about mouse-slips. The night before, when I was testing the playchess software for the first time, I was always slipping up, moving my king to f8 when I meant to castle. I feared that I was completely spastic when it came to the computer, particularly compared to the tykes who would be playing next to me. I thought that it would be tragic if on my one chance to play a world champion I made a losing mouse-slip before I even got out of the opening. But I soon discovered that the mouse I was using was defective, and I switched to another one. I could not get out of mind, however, the possibility of losing not because he outhought me but because he outfingered me.

9.Nxc6 dxc6 10.Qe2 e5 11.Be3
Mig: We now enter a very instructive sequence in the opening. Once you get beyond the basic concept of "develop your pieces" many players start to wander. This is why you read critical remarks about "routine developing moves" in GM commentaries. Sometimes that means there was a sharper, less obvious opportunity. But Master opening play is all about move order and not all developing moves are created equal. Move order is critical for two main reasons: interfering with your opponent's development and allowing maximum flexibility for your forces. It's all about keeping your options open, narrowing your opponent's options, and not wasting time.

You can't really criticize 11.Be3 as a move. It develops a piece to an active position and takes the c5 square away from the black bishop. It's the concept behind the move and the follow-up that show White is drifting and playing routine moves. Let's look at the position before 11.Be3 using the concept of "play the most obvious moves first" and keeping in mind that we want to keep our options as open as possible.

White has one major problem with his development, the knight on c3. It looks like a perfectly reasonable piece on a perfectly reasonable square, yet it has no future at all on c3 as long as there is a black pawn on c6. You might expect Black to want to play c5 to open up the bishop, but you'll see that Kasparov studiously avoids that move and instead relocates the bishop via c8-e6 just to keep the clamp on the knight. So the lousy knight is at least discouraging ..c5 and thus creates a lousy bishop on b7. Meanwhile we have the bishop on f3, which is headed to g2 after g3. Since this is clearly the only possible plan for the f3 bishop, why not play g3 and Bg2 immediately? The point is that the c1 bishop might be better placed somewhere else down the line depending on what Black plays. (This all goes back to why the old Tarrasch saying "knights before bishops" makes sense. The knights have more obvious homes so you develop them first.) Since there is no other logical plan for the f3 bishop than g2 and Bg2, White should play that before touching the c1 bishop, which could go to g5 after ..Nf6 or capture on f4 after f4 exf4 in some lines, which would save a whole tempo if the bishop hasn't moved yet.

Even before that, another move White should always consider in similar positions is a4, attacking the strong black queenside pawn formation and opening a line for the a1 rook. Any open line will accentuate White's lead in development. (Another reason why g3, Bg2, f4 should be played as quickly as possible.) White isn't afraid of ..b4 because the knight is useless on c3 anyway and will be much better off after relocating via b1 or d1 (e3 is open!). And it's not as if the a1 rook has a day job to get to somewhere else. Even with all that, the slow development by White with the long bishop maneuver doesn't promise more than equality. Kasparov's opening play in this game is typical simul play in that he is not ambitious either. [ 11.a4 ; 11.g3 ]

11...Nf6 12.Rfd1
[ I thought about 12.a3 but was worried it might make it easier for Black to push pawns and open up the queenside to his advantage. Maybe the other rook belongs on d1 so that a rook on f1 could assist in a kingside action if I ever manage to relocate the white-squared bishop and play f4.

Mig: This move is fine, but Paul should be ashamed of even thinking about a3 and yet not thinking about a4!; 12.Rad1 ]

[ I was hoping for 12...Bb4 13.Rd3 ]

I want to relocate the bishop and play f4

13...0-0 14.Bg2 a5 15.f3
? Mig. Ugh. Burying the bishop on g2 alive. Where is White going to find play now? Black has the advantage on the queenside and there are no entry points on the easily disputed d-file. It's the kingside or bust for White. This loss of time dooms White to a completely passive existence for the rest of the game. [ Now I chickened out from 15.f4 because I was afraid it might be too loosening . I played the passive 15. f3. My idea was to put the queen on f2, where it is doing some work on the g1-a7 diagonal. But of course I've now buried the light-squared bishop. ]

Mig: I like this move a lot. The pawn on c6 is killing the c3 knight (a signature of Kasparov's play throughout his career) so the bishop finds a better home. This move also prevents f4 (in case f3? was a mouse-slip!), which would now meet with ..Ng4, winning the bishop pair and leaving the white king open on the dark squares.

Somewhere around this move the toddler playing at the screen next to me started squealing that when Garry moved one of his rooks it had turned into a knight! His father and his handler from the Long Island Chess Nuts club gathered around the screen looking concerned, and the organizer of the simul placed a frantic call to Dresden to see what Kasparov was seeing on his screen. Everything turned out to be in order in Germany. The toddler gamely played on, pretending that the knight was really a rook. At this point I started day-dreaming about which of the pieces on my board I'd like to magically turn into a knight.

Mig: Distractions are a major part of simul play. It's not just the simul giver who plays below his normal level. For proof we have the famous example of a British player who lost to Korchnoi in a simul before a tournament only to beat him in the tournament proper! It must be said that Paul didn't get much help from his teammates, most of whom were rated around 700 points lower. As Garry told him after the game, "You didn't have a chance, but because of your teammates!" Kasparov could tell after only dozen moves or so that this was the only game that might cause him trouble, although he continued to play conservatively. During the game I thought he might try something like ..h5 (with the rook still on h8) just to ruffle Paul. [ 16.f4? Ng4 ]

16...Be6 17.Bf1 Nd7 18.Bd3
I started thinking that I've horribly misplayed the opening because it took my bishop five moves to reach d3, the place it could have reached in one stroke on move five. So I was annoyed with myself because I usually don't play so passively. But then, as Garry atypically consumed some time on the clock, I started thinking that despite the bishop's long journey the position might be even--or if I stood worse, it was not by much. Not what you generally want with White but not so bad considering that Black outrated me by close to 900 points.

Mig: White's position is solid, but I don't think he's equal at this point. Black has a queenside space advantage and better potential breaks with moves like ..c5 and ..f5. Certainly nothing decisive, just better long-term prospects, especially if White continues to play aimlessly.

18...Rab8 19.Kg2
The problem is I didn't have a plan, so with the clock ticking, I shuffled pieces.

19...Rfd8 20.Rd2
Another purposeless move.

Mig: ? Now Black can force the doubling of the white pawns, another positional advantage in the bag. It wasn't too late to try for counterplay with f4 now that the knight can't go to g4. The chance of Kasparov making a mistake in a technical grind with positional advantages is not good so you may as well take a shot and enjoy yourself a little. Playing the entire game without crossing the fourth rank is a little depressing when you have white. [ 20.a3 ; 20.Ne2 ]

20...Bb4 21.Rdd1
More dumb shuffling but if I double rooks, the a-pawn falls after he takes my knight.

Mig: At this point I realized, as did Garry, that Paul was never ever going to play f4!

21...Nb6 22.a3
[ 22.Ne2 ]

22...Bxc3 23.Bxb6 Qxb6
Mig: Kasparov is confident that he can win this endgame. I doubt he would have captured with the queen against Anand. (Sorry Paul!) Keeping the queens on the board would accentuate Black's space advantage and make it harder for White to protect all his weaknesses. With just rooks and one set of minors on the board it's easier to patch the holes, at least in theory.

24.Qxb6 Rxb6 25.bxc3
And of course White's position is already tough because Black's king can travel up the f8-c5 diagonal

As we moved into the endgame, I started wondering about the pawn-promotion setting on the software. Would it give a choice of pieces or just automatically promote the pawn into a queen? And now I had an idea for how I could outfox him. Maybe I could adjust the settings so that it would promote my pawns into queens and his pawns into bishops. Imagine the expression on his face when, after a pawn race in which he thinks he's going to queen first, he instead gets a bishop on the eighth rank! Then self-doubt overwhelmed again. Was I sure I could win a queen vs. two or three bishops endgame? I never studied that chapter in Fine.

Mig: If that's the sort of thing you are thinking about, you are in trouble in more places than on the board! The position is very close to being technically lost, although there is a lot of work for Black to do. Paul is perfectly right about the biggest problem being the Black king's future home on c5.

Mig: f4! Augh!

26...Rdb8 27.Kf2 Ke7 28.Ke3 f6 29.f4 g5
Mig: Kasparov wants to lock things up so he can focus on winning on the queenside.

[ I think that it might be better to try to close the kingside with 30.f5 but I figured that if he was allowing me to try to close the kingside, it couldn't be the right thing to do.

Mig: Interesting logic, and not far from wrong. The big question here is this: Will it help or hurt White's defensive chances to open another area of potential activity for the rooks? Should White try to limit the action to the queenside defense or try for counterplay on the kingside? Counterplay is usually the best way to go, but of course each case is different. You have to evaluate whether or not Black can force a win on just the queenside. White's structure isn't pretty, but it's not going to be easy for Black to just make the pieces disappear and win with better pawns. 30...Bf7 31.Be2 h5 32.h3

Mig: This is one possible plan, preparing to seal the kingside like the mummy's tomb if Black advances. The downside is that this further space concession creates one more weakness. 32...h4 33.g4 c5 34.Rd2 b4 35.cxb4 cxb4 36.axb4 Rxb4

Mig: This endgame is very hard to hold for White. There are pawn weaknesses on c2, e4, and h3. The black a-pawn is a killer if the rooks come off because the f7 bishop controls the critical a2 square. The distant passer also makes king and pawn endgames good for Black. That's a lot of trumps to hold. White's practical defensive chances are slim, just as in the game.]

30...gxf4+ 31.gxf4 Rg8 32.Rg1 Rbb8 33.Rxg8?
I made it easier for him by swapping rooks. But at the time I thought it would be useful to keep a rook on b1 and try to play a2-a4.

Mig: No reason to hand the file over without a fight. This is looking very close to a draw! White wants to play c4 and set up a blockade. All of the sample lines below are better than trying the same thing but with Black in control of the g-file. It would be very hard for Black to make progress if White just leaves his rooks on b1 and g1 and plays c4. [ 33.f5 Rxg1 ( 33...Bf7 34.Rxg8 Rxg8 35.c4 bxc4 36.Rb7+ Ke8 37.Rb8+ Ke7 38.Rb7+ Kf8?? 39.Rxf7+ Kxf7 40.Bxc4+ Kf8 41.Bxg8 Kxg8 42.Kd3 Kg7 43.Kc4 Kh6 44.Kc5 Kh5 45.Kxc6 Kxh4 46.c4 Kg4 47.c5 h5 48.Kd5 h4 49.c6 h3 50.c7 h2 51.c8Q h1Q ; 33...Bd7 34.c4 ) 34.Rxg1 Rg8 35.Rxg8 Bxg8 36.c4!?

Mig: Giving up a pawn to break up the black pawns and set up a blockade. 36...bxc4 37.Be2 Kd6 38.Kd2 Kc5 39.c3 Setting up a blockade that will be complete after a4. Black has no way in. ( 39.Kc3 Kb5 ( 39...Bf7 40.a4 ) 40.Bh5 Ka4 41.Be8 Kxa3 42.Bxc6 Bf7 43.Bd5 Bh5 44.Bxc4 ( 44.Kxc4 a4 45.Kc5 Kb2 46.Kd6 Bf3 47.Ke6 Bxe4! ) 44...a4 ( 44...Be8 45.Bg8 h6 46.Kc4 Ba4 47.Kc3 Be8 48.Kc4 Bc6 49.Bd5 Ba4 50.Kc5 Bxc2 51.Kd6 a4 52.Ke6 Kb4 53.Kxf6 a3 54.Kg6 Bxe4 55.Bxe4 a2 56.f6 a1Q 57.f7 Qa6+ 58.Kg7 Qa7 59.Kg8 Qg1+ 60.Kh7 Qc5 61.Kg7 Qe7 62.Kg8 Qe6 63.Kg7 Qe7 64.Kg8 Qe6 65.Kg7 ) 45.Bb5 ) 39...Kb5 ( 39...Bf7 40.a4 Kd6 41.Ke3 Ke7 42.Bf3 Kf8 43.Bg4 Kg7 44.Bf3 Kh6 ) 40.Bd1 Bf7 41.a4+ Kc5 42.Ke3 ]

33...Rxg8 34.f5 Bd7
I think I'm just lost now. Suddenly it dawned on me that the other simul players had all been defeated and it was just me and him. I collapsed as quickly as the Republican Guard.

Mig: If this move is necessary White is lost for sure. Now the black king makes it to c5 before White can try to break with c4 or a4. c4 still needed to be tried. There were still some chances to set up a blockade. [ 35.c4 bxc4 36.Bxc4 ( 36.Be2!? Rg3+ 37.Bf3 Rh3 38.Rh1 Rxh1 39.Bxh1 Kd6 40.Bf3 Kc5 41.c3 Kb5 42.Bd1 Be8 43.a4+ Kb6 44.h5 h6 45.Kd2 c5 46.Ke3 Blockade.) 36...Rg3+ 37.Kd2 Rg4 38.Rb7 Kd6 39.Bd3 Rxh4 40.Ra7 a4! 41.c4 ( 41.Rxa4 Bxf5 ) 41...c5 42.Bc2 h5 43.Ra6+ Ke7 44.Ra7 Kd8 45.Ra8+ Bc8 ]

[ 35...Rg4 ]

36.Kf3 Kc5 37.a4?
Mig: Further accelerating defeat. Why play this if you can never capture on b5 because it would give Black an outside passer of an a-pawn? Better was sitting tight and waiting for Black to commit. There is no obvious plan for Black that does not allow White some counterplay. As long as Black cannot force the exchange of rooks it will be very tough to win. Paul was in time trouble for the rest of the game and things fall apart very quickly. [ 37.Be2 Be8 38.Kf2 Rg7 39.Rh1 ( 39.Rd1 Rd7 40.Rd3 Rd6 41.Ke3 ) 39...Rd7 40.Ke3 Bf7 41.Rg1 Bc4 42.Bd3 ]

37...Rb8 38.Kg4 Be8 39.Ra1 b4 40.Rb1 Bf7 41.cxb4+
[ 41.Kf3 b3 42.cxb3 Rxb3 ]

41...Rxb4 42.Ra1 Kd4 43.Ra3 Rb1 44.Kg3 c5 45.Kf2 c4
I should have resigned here, sparing my dropping a piece with just four seconds left on my clock. "And this guy writes about chess?" I could hear him thinking.


It was a lot of fun to play Garry but it was not one of my best efforts. When I told Greg Shahade the result, he responded: "Think of it this way, whenever you play in tournaments again, you're not going to face anyone stronger." That's what I'll keep reminding myself when I get into thorny positions at Foxwoods. 0-1