The tournament may have drawn players from all over the world, but the best game so far of the Berkeley International was the round two collision between two housemates, David Pruess and Jesse Kraai. Here’s how the action went down:
David Pruess — Jesse Kraai
French Defense
1. e4 e6 2. d4 d5 3. Nc3 Nf6 4. Bg5 Be7 5. e5 Nfd7 6. Bxe7 Qxe7 7. f4 a6 8. Nf3 c5 9. Qd2 Nc6 10. dc Qxc5 11. O-O-O b5 12. Bd3 b4 13. Ne2 a5
So far this is all book, I think. This position has arisen in 21 games on ChessBase, and White has a pretty good record of +12 -6 = 3. In 18 games White played the prophylactic move 14. Kb1. In the other three games, White concluded (as David did here) that prophylaxis was not necessary.
14. Rhe1 O-O?
Sometimes, even senseis sleep! Rather surprisingly, Jesse errs immediately after the game leaves “book.”
Okay, I have an imaginary story to tell. As I wrote in my last entry, the games are being played in a classroom at the Berkeley Chess School. There is a blackboard in the room with 10 pieces of advice for opening play and 10 pieces of advice for middlegame play. The advice is all traditional stuff — “Don’t move the same piece twice in the opening,” “Don’t make too many pawn moves,” and … “Castle early in the game.”
So I can imagine Jesse sitting at the chess board, facing the blackboard, and reading the pointers meant for 8-year-olds. “Ah!” he nods. “Castle early in the game. Very sound advice.” And proceeds to do so.
Good alternatives were 14. … Ba6 (the simplest, taking out one of White’s attackers and trading off the French bishop), 14. … a4 (the most ambitious, which the computer prefers and gives Black a small edge), or even 14. … b3 immediately.
White to play and give Black a really cold shower.
The move that Jesse overlooked, and David pounced on, was 15. Bxh7+!
Actually, I doubt that Jesse actually overlooked this move, because it’s very thematic in the French Defense, but I think that what lulled him into a false sense of security was the fact that this is not your ordinary, standard bishop sacrifice on h7. In particular, White’s queen is on d2 rather than d1, and so it has to take a slightly different route into Black’s kingside, via d3 and maybe h3 rather than via h5. Jesse may have felt that White’s pieces were not working together quite well enough for the attack to succeed. What is the knight on e2 doing? The rook on e1?
Well, we’re going to see! Kudos to David, by the way, for playing a brave sac that really could not at this point be calculated all the way to the end.
16. Ng5+ Kg8 17. Qd3 Re8
17. … g6 doesn’t keep the queen out because it can go to h3.
18. Qh7+ Kf8
Now another really nice move by David.
19. Qh5! …
I think that most people would have played 19. Qh8+ here, in order to start regaining the sacrificed material after 19. … Ke7 20. Qxg7. But it’s not at all clear that White is winning after 20. … Rf8 21. Nh7 Ba6 22. Nxf8 Rxf8. Material is more or less even — Black has two pieces against a rook and two pawns. Black has weathered the kingside attack, at least for the time being, and now White is going to have to worry about defending his king.
Instead, David plays a move that induces another weakness in Jesse’s position. Winning the material back is not so important. Checkmating the king is.
19. … g6
The alternative is 19. … Nd8. I won’t bore you with a lot of computer analysis, but suffice it to say that the problems with this move are: (1) it blocks the escape route of Black’s king to the queenside, and (2) it abandons the center to White. White simply plays 20. Nd4 and has all sorts of threats, like Qh8+ followed by Qxg7 and f5. Meanwhile, Black has zero counterplay on the queenside.
20. Qh6+ Ke7 21. Qh4! …
Another nice move, threatening to win the queen with a discovered check. Black’s king has to go back into the rabbit hole.
21. … Kf8 22. Rd3 …
Threatening Rh3, Nh7+, etc. Here Jesse comes up with a very good defensive idea.
Black to play and throw a monkey wrench into White’s plans.
To save the game, Black has to do something to shift the action to the queenside, and in general to wrest the initiative away from White. This Jesse accomplishes with the feisty move
22. … b3!
threatening checkmate. David, of course, played 23. Rxb3 Nb4, and now White faces some new problems.
White to play and snatch the initative back away from Black.
Again, Black is threatening mate on c2, and 24. Rc3? doesn’t work because of 24. … Nxa2+. Jesse must have been expecting (or at least hoping for) 24. Nc3?, a plausible looking move that causes White’s pieces to start tripping over themselves. After, say, 24. Nc3? d4 25. Nge4 Qc6 26. Nd6 Nc5 we have a very complicated and unclear position where, again, Black has succeeded in drawing White’s attention away from the kingside.
Instead, David found the exquisite and decisive move,
24. Nd4!! …
No price is too high to pay when you can get your opponent in a mating net! This move certainly answers the question of what the knight on e2 was doing. Now, from its outpost on d4, it defends White’s king and (if Black doesn’t take on d4) it also serves the very important attacking function of putting pressure on the weak square e6. For instance, on 24. … Ba6 25. Qh7 the f-pawn would fall, and the e-pawn next, and the queen next.
So Jesse has no choice but to take the knight, but Black’s king is now caught in a net.
24. … Qxd4 25. Nh7+ Kg8 26. Rh3 Nxe5
Jesse continues to put up stiff resistance.
27. Nf6+ …
Diagram 5
It’s interesting that the computer recommends 27. … Kg7 with a 2-pawn deficit for Black, rather than 27. … Kf8 with a 7-pawn deficit (!). The point of 27. … Kg7 is that even though Black is giving up a rook with check after 28. Nxe8+ Kf8, there are still quite a few things for White to worry about. Black has ideas like … Nd3+, or … Rb8 followed by … Nxa2+ and mate on b2, or simply … Kxe8. So it’s not as routine as you might think for White. The best line, according to the computer, is 29. Nc7! Ned3+ 30. cd Qc5+ 31. Kb1 Qc2+ 32. Ka1 Qxc7 33. Qf6! (diagram) Â
Position after 33. Qf6 (computer analysis)
Once again, any amount of material is worth giving up for a mating net. If 33. … Nc2+ 34. Kb1 Nxe1 35. Rh8 is checkmate. I also want you to notice, from the diagrammed position, that Black’s attempt to set up a draw by repetition with 33. … Nc2+ 34. Kb1 Na3+ is foiled by 35. ba Rb8+ 36. Ka1, and Black cannot deliver checkmate on c3 because White’s queen is guarding the long diagonal.
I am reminded again of something that GM David Bronstein once wrote. When Black gets checkmated on the kingside, you usually find that the reason is that the queenside pieces did not participate in the defense. The most common offenders are the queen bishop and the queen rook. That is certainly the case here!
Anyway, this is all fantasy. The way the game really ended, starting with Diagram 5 above, was 27. … Kf8 28. Qh8+ Ke7 29. Qxe8+ Kxf6 30. fe+ (Remember when I asked what the rook on e1 is doing? It’s defending the e5 pawn, and cutting off the escape of Black’s king — that’s what.) 30. … Kg5 31. Rg3+ resigns.
A beautiful game by David! For Jesse, it was probably akin to waterboarding. I wish I had been a fly on the wall of the “GM house” last night!
As for how the tournament as a whole is going, the surprise leader after two rounds is Marc Esserman, who has beaten two grandmasters in a row — Friedel and Sharavdorj! However, it’s still very, very early. The standings are:
1. Esserman — 2-0
2-7. Izoria, Krush, Kacheishvili, Rensch, Pruess, Haessel — 1.5-0.5.
The round three games should be getting started right about now.
{ 5 comments… read them below or add one }
Wow preuss beat jesse, I have an attacking book and one of its main example is the french defence, is this a common sac?
Two comments:
1. I bet David Pruess saw more than Dana thinks. According to ICC, Pruess spent about 25 minutes before playing the sacrifice. After 15.Bxh7+, you can get to move 18 right away because the action is forcing. The only real question now is whether black does Re8 or Rd8. Thus the calculations truly begin with move 19. Finding Qh5 was a finesse, but one that appears in books. Then moves 19-22 were all natural, plus black’s replies are very limited. I can easily see a strong player calculating to move 22 or 23 within a few minutes because so many of the moves are forced. However, I’ll be curious to know whether Pruess saw the brilliant 24.Nd4 before he sacrificed.
2. You analyzed 19.Qh8+ Ke7 20.Qxg7 Rf8 and then suggested 21.Nh7. How about the spectacular 21.Nxe6 Kxe6 22.f5+ where now Kxf5 is suicide (Qh6 and Ng3+) but the alternative Ke7 runs into Nf4 with ideas such as Nxd5, Rxd5 and e6. I’ve analyzed quite a few lines with my students and we can’t find much for black, although admittedly black has quite a few defenses.
Bravo David! A nice brute force game of calculation skill.
Thadeus — bishop sacrifices at h7 are thematic in the French and some other openings (e.g., the Colle System) whenever White is able to chase Black’s knight away from f6 with the move e5. However, saying that they are thematic is not the same as saying they are common. Most Black players are aware of this motif and will take steps to prevent it. What’s surprising about this game is that Jesse castled right into it when he didn’t have to. As I said in my entry, my hunch is that he didn’t quite believe that White’s pieces were well enough coordinated to make it work. The other possibility, as Michael suggests for David, is that Jesse also calculated all the way to move 23 but thought David had to play 24. Nc3.
We may find out eventually. I e-mailed Jesse and asked if he would do a ChessLecture on this game, and he said probably. Because he lives with David, I’m sure that they will talk about the game, and so maybe Jesse can tell us what both of them saw and didn’t see. I’m looking forward to that!
Looks like Jesse it totally out of form. I hope he’s gonna play decent chess from round three and stop looking at those beginnersadvise slogans. 🙂
before sacing on h7 i calculated up to 22.Rd3. i did not figure out what to do about b3 in advance; i was confident white would be winning easily in that position. but seeing qh5 and qh4 was very important. 19…Nd8 required a bunch of my time as well.
{ 2 trackbacks }