French Defense - a wild piece sacrifice

Shirov's Bracket After his match with Alexei Shirov from Round 4 of the FIDE World Championship Tournament, Nigel Short commented that some of the positions in which he found himself were "absolutely ridiculous". In fact, Shirov's strategy for the entire match-tournament was to play as agressively as possible. Thus, while Shirov ruled the lower bracket of the tournament, the other clear favorite in the event, Vladimir Kramnik, was making steady progress in the top bracket with a methodical "hold the draw with Black, grind them down with White" type of strategy. But Shirov preferred to rely on his great creativity, combined with fantastic tactical ability, to get to his destination, and against Short he pitched a knight in the early middlegame to get a sustained attack. To keep the fire burning against Short's spirited defense, more wood had to be thrown in, until Shirov eventually found himself down a queen for a rook. But when Short's king was finally chased to the center of the board, it became clear that Shirov had everything well under control.

As it turned out, in the QuarterFinals of the event, both Kramnik and Shirov were upset by players using their own styles against them - as Kramnik was turned away by the super-solid play of Michael Adams, and Shirov became too carried away in a wild game against Liviu-Dieter Nisipeanu. [Event "World Chess Championship"] [Site "Las Vegas USA"] [Date "1999.08.09"] [Round "10"] [White "A. Shirov"] [Black "N. Short"] [Result "1-0"] 1. e4 e6 Short has played the French Defense for many years; he must have noticed that Shirov has been playing it as well, particularly in the match against Gilberto Milos 2. d4 d5 3. Nc3 The main line, allowing the sharp Winawer 3 ... Bb4 and the difficult 3 ... Nf6 Nf6 4. Bg5 dxe4 This is the line which Shirov used against Milos, losing in regulation with it but later coming back to score with it in the playoff 5. Nxe4 Be7 6. Bxf6 gxf6 It is a strange phenomenon that to find a gap in one's opponent's repertiore, a line should be played which is similar, but branches off, to a line the opponent plays. Here Short has imitated Shirov's use of the 3 ... Nf6 French, but using 6 ... gxf6 rather than 6 ... Bxf6. Moreover, Short has experience with these kinds of positions from his younger days. Overall, an impressive choice 7. Nf3 Nd7 8. Bc4 The bishop is good here, trying to intimidate Black with ideas such as 8 ... b6 9 d5 or 8 ... c5 9 d5, because 8 ... Nb6 9 Bb3 simply misplaces the Black knight. The knight on d7 is either headed to f6 (after ... f6-f5) or will support an eventual central break with ... c7-c5 c5 Provoking an immediate crisis, as White can play 9 d5. If White refrains from that advance, then the exchange of c-pawn for d-pawn in general favors Black who would then have a central preponderance 9. O-O It appears that Shirov spent a long time deciding on this move.

The problem with 9 d5 was the very perplexing 9 ... Nb6 10 Bb5+ Bd7 11 Bxd7+ (on 11 Qe2, Tseshkovsky - Bronstein, Moscow 1981 continued 11...Bxb5 12.Qxb5+ Qd7 13.Qxd7+ Kxd7 14.dxe6+ fxe6 15.0-0-0+ Kc7 with no problems) Qxd7 12 d6 Qc6! 13 Qe2 f5! Therefore it seems that 9 d5 has not proved convincing and in any event Shirov was not interested in running into Short's home preparation in this line and plays the most practical move instead O-O Now Black has put a marker on White's center as 10 d5 does not work at all due to 10 ... Nb6. Meanwhile, the Black king is not all that badly placed,as Black should be able to bring pieces over to guard him at the right moments 10. Re1 Playing calmly. On 10 Qd2 Kh8 (I think this is better than 10 ... f5 11 Ng3 Bf6 12 Nh5 and Black has committed his defensive formation too early) and if 11 Qh6 Rg8 looks ok. But Speelman has pointed out that 10 c3 could be a logical improvement as then 10 ... Nb6 can be answered by 11 Bd3 as d4 is sufficiently protected. But on 10 c3, Short would have had no reason to move the knight to b6 at all. This is part of Shirov's brilliance - his ability to embroil the opponent in complications. Nb6 11. Bf1 The bishop can always come back out later. 11 Bd3? cxd4 was not sound cxd4 12. Nxd4 A route has opened for the queen. White's knights are impressively placed, but the Black pawn cluster is very compact. White has a small edge based on greater piece mobility Kh8 13. c3 e5 Black advances in the center, precipitating a crisis. If he can get the knight on d4 to leave down, the light-squared bishop will find a nice post on e6 or f5. On the other hand, this move allows White tactical possibilities which aim ultimately for control of those same light squares. DIAGRAM 14. Qh5 Shirov has provided for this possibility. After 13 c3, I wondered whether 13 ... e5 would really make the knight on d4 go away (with, for example, 14 Nb5, which does not give any real prospects for the advantage) and cursorily considered 14 Qh5 without seeing the real point - that 14 ... exd4?? flat out loses to 15 Nxf6!! Bxf6 16 Bd3 and mate happens on h7. Meanwhile, White is trying to win the square f5 for his knight Nd5 A multi-tasking defensive move. By covering f6 with the knight, sacrifices on that square are eliminated. And Short denies 15 Nf5? due to 15 ... Nf4 15. Rad1!? This kind of move is a logic-driven sacrifice. If the White knight retreats, and realistically the only such move is 15 Ne2 to stop 15 ... Nf4, then Black has a reasonable game. On the other hand, if after the text, Black cannot capture White's knight, then White is doing very well. But these considerations pale compared to finding the actual justification of the sacrifice after 15 ... exd4 16 Rxd4 f5!

With hindsight we can trace back to this and conclude that Black is doing quite well after he takes the knight. Should therefore the move 15 Ne2 be given a double exclam for analyzing the sac 15 Rad1 and rejecting it? exd4 Black had a choice about this move. He could have played 15 ... Nf4 16 Qh6 Qb6! with an ok situation. 16. Rxd4 f5 This move is a lot more efficient than 16 ... Be6 - it puts a bishop on guard against Rd4-h4 and puts the f-pawn on that horrifying b1-h7 diagonal, as well as defending the knight on d5 through blocking the lateral action of the White queen. Aside from this, the problem with 16 ... Be6 is that it puts the e6 bishop in play and after 17 Bc4 White wins back the piece after 17 ... f5 18 Bxd5 (hitting e6) Bxd5 19 Qxf5. 17. Ng3 A very deep conception by Shirov. He figures that he will be able to plant the knight on f5 without fear of its exchange as the Black bishop on c8 is duty-bound to head for e6 to shore up the pinned knight. Certainly, in this position, Short will be interested in getting his dark-squared bishop into the action. Bc5! Ouch! The bishop hits not from the defensive f6 square but from c5! On 17 ... Bf6, Black would have a lot of problems after 18 Rd2 (threat: 19 c4) Qa5 19 Nxf5 (threat: 20 Qh6). But now, after 17 ... Bc5, if 18 Rd2 then Black can get out of the pin with 18 ... Qb6 although White then has 19 b4. And on 18 Rh4 Nf6 defends, although after 19 Qh6 White may be threatening 20 Nh5, but in this line f2 looks very soft (Black may have the bishop sacrifice ... Bc5xf2+ in some lines) and Black has other defenses (such as 19 ... Rg8). 18. Rd2 Qb6! DIAGRAM Black liberates himself from the pin. I did not think the issue would be clear though because of the miracle move 19 b4. Then 19 ... Nxb4 20 axb4 Bxb4 21 Rb1 is not convincing; but the key point is that after 19 b4, Black has 19 ... Bxf2+!! 20 Rxf2 Nf6 21 Qh6 also does not work for White because of 21 ... Qxf2+! 22 Kh1 Ng8!, while 21 Qg5 runs into 21 ... Ng4 22 Nh5 Qg6!! (not 22 ... Qxf2+ 23 Kh1 Rg8 24 Qxg8+! Kxg8 25 Re8 mate!) 19. Qe2 A very surprising move. White is trying to dominate Black's minor pieces right in the middle of the board! But as we saw above, the planned-for 19 b4 does not work, and Speelman points out the vicious line 19 Qf3 Nf6 20 b4 Ng4! with the idea 21 bxc5 Qh6! After the text, 19 ... Be6 does not work due to 20 Qe5+ Kg8 21 Nh5. So the Black knight must move. Nf6 The most natural, although it also looked feasible to go to c7 or e7. When one is a piece up and trying to keep it, everything has to be looked at! These moves may be well worth looking at; the main point is that the kinght is safe and Black can still play ... f6 to cover his king. To me, neither 19 ... Nc7 20 Qe5+ f6 21 Qf4, nor 19 ... Ne7 20 b4 Bd6 21 Rxd6 Qxd6 22 Qxe7 (going into an endgame with a pawn for the Exchange) is clear. 20. Qe5 The White queen is amazingly stable on e5. Now 20 ... Re8?? would lose to 21 Qxf6+! Qxf6 22 Rxe8+ followed by Ng3-h5+. Kg8 Black is looking forward very much to the endgame in the case of 21 b4 Bxf2+ 22 Rxf2 Qxf2+ 23 Kxf2 Ng4+. 21. h3 Getting rid of the ... Ng4 motifs, Shirov is now finally ready to play b2-b4! But Black is still a piece up and it is his move! Is all of this possible? DIAGRAM Re8 This should win. Black has a bishop that can drop back to f8 in variations such as 22 Qxf6 Rxe1. But Speelman recommends the less forceful 21 ... Bd7 when Black is clearly better after 22 Nxf5 Bxf5 23 Qxf5. 22. Qxe8+ There is nothing better. White still has a little pressure as Black's back rank is undeveloped, but White will only have a rook for the queen Nxe8 23. Rxe8+ Kg7 Not 23 ... Bf8 24 Rdd8 24. b4 DIAGRAM Qc6!? Against the brilliant Shirov, Short obviously had to be very careful, and this looked like progress. The idea is 25 Rdd8 Bb6 26 Rg8+ and then most likely 26 ... Kf6 (or possibly 26 ... Kh6!? - but see below). As an example of the terribly complicated variations Short had to choose from, there was 24 ... Bd6, which only leads to equality after 25 Rxd6! Qxd6 26 Rxc8! (idea: Ng3xf5+) Qxg3!! 27 Rxa8 Qxc3; and another move which he probably did not notice - 24 ... Be7!! (as Speelman points out) 25 Rxe7 Be6, trapping the e7 rook. 25. Rdd8 Bb6 26. Rg8+ Kf6? This looks natural but the bizarre move 26 ... Kh6! was better. Then 27 Rxc8 Qxc8!! (27 ... Rxc8?? 28 Nxf5+) 28 Rxc8 Rxc8 and Black should still be winning). Aside from this, although the king looks terrible on h6, I do not see any way for White to exploit this, and meanwhile Black will work on getting counterplay against f2. 27. Nh5+ Ke5 Another horrible spot for the king. It is possible that 27 ... Ke7 is a better move, although it involves giving back the entire material advantage after 28 b5 Bxd8 (he cannot move the queen away due to 29 Rge8 mate) 29 bxc6, as I suspect that Black is better in this endgame with his two bishops after 29 ... bxc6. But it is tremendously difficult to simply give back the queen like this. 28. Rd3 DIARAM By defending c3, White prevents the queen from breaking in to his position. Meanwhile, the Black king is in a very weird position, the c8 bishop is still pinned, and the queen must guard against a check on e8. What is going on here?? Qh6 Malcolm Pein of The Week in Chess recommended 28 ... a5, although apparently 29 b5 Qc5 30 Rd2 is still good for White, as the threat of Rg8-e8+ is renewed. Short is prepared to give back a rook in any event to the maneuver Rg8-e8+ forcing ... Bc8-e6, but with the text, for now he is attacking the White knight. 29. Be2 Qc1+ This is the move that officially gives back a rook, but it was probably too late for defensive moves like 29 ... Qe6 30 Kf1 (threatening f2-f4+) or 29 ... Qc6 30 Bf3. 30. Rd1 Qxc3 31. Bf3 With the threat of 32 Rd5+ Ke6 33 Re8 mate. Be6 32. Rxa8 DIAGRAM Shirov restores material equality, but with a big advantage based on the horrible situation of Black's king. Qxb4? It seems to me that 32 ... Qb2 would be more usefully obscure, with a hit on f2, although White could proceed with 33 Re8 Qxf2+ 34 Kh1 Bd4 35 Rd8. 33. Re8 Threatening 34 Rd5 mate! Bd4 34. Rd8 Bb6 35. a3! Now the Black queen has to keep an eye on e1. Qa5 36. R8d2 With White's rooks finally connected, the Black king is in big trouble on e5. Either rook is poised to make the fatal check on the e-file. f4 This is the only move. It both opens the f5 square (but only temporarily) and provides for an e-file block with the move ... Bb6-e3. 37. Bg4 Shutting the door on any escape to the f-file. Bc4 Doing his best to interfere with the coordination of White's rooks. 38. Re1+ Be3 39. Rdd1 White is winning a piece. Bb3 40. Rb1 Bc2 41. Rb4 Because 41 ... Qxa3 42 Rxf4 would end all resistance. f5 42. Bf3 Be4 43. fxe3 Bxf3 44. gxf3 Qxa3 45. Nxf4 Defending the rook indirectly throught the tactic 45 ... Qxb4? 46 Nd3+. Now the two rooks plus knight can easily weave new mating threats against the Black king. b6 46. Rd4

1-0