Commentary: Dortmund 2015, Round 6 (Kramnik - Nisipeanu)

I've been working on finishing off my 2015 collection of commentary games, with only one remaining after this one.  The below game nicely complements Nakamura - Harika by allowing us to examine a similar opening structure for White.  Black here chooses a somewhat different approach and we end up with a position similar to a Nimzo-Indian in reverse, in which White has comfortable play against Black's isolated queen pawn (IQP).  This game is something of a classic IQP strategic fight, in which Kramnik (as White) squeezes Black relentlessly, but in the end he is unable to convert the rook endgame.  ("All rook endgames are drawn" is a technically incorrect, yet wittily insightful remark attributed to Tarrasch.)  Although Nisipeanu made a few possibly dubious strategic decisions in the middlegame, including the strategic no-no of exchanging minor pieces unnecessarily while possessing an IQP, he had the mental toughness to resist the resulting pressure and found a saving tactical resource (60...f2+!) in the endgame to reach a technically drawn position.

[Event "43rd GM 2015"] [Site "Dortmund GER"] [Date "2015.07.04"] [Round "6"] [White "Kramnik, Vladimir"] [Black "Nisipeanu, Liviu Dieter"] [Result "1/2-1/2"] [ECO "A13"] [WhiteElo "2783"] [BlackElo "2654"] [Annotator "ChessAdmin/Komodo 8"] [PlyCount "165"] [EventDate "2015.06.26"] [SourceDate "2015.02.07"] 1. Nf3 d5 2. e3 {an unusual approach, but not a bad one. Should Black proceed in a classical fashion, for example, White can play a Nimzo-Indian in reverse and with an extra tempo.} Nf6 3. c4 e6 4. b3 {now the game is in English Opening territory.} c5 5. Bb2 Nc6 6. cxd5 {otherwise Black is strong in the center and can even think about ...d4.} exd5 7. Bb5 {White now has a similar structure to a Nimzo-Indian (reversed). One point of the move is to restrain .. .d4 by pinning the Nc6.} Bd6 {the most active square for the bishop. Black does not need to worry about the d5 pawn at the moment; the only drawback of the bishop's presence on d6 is blocking the Qd8's defense of the pawn.} 8. d4 { an unusual decision at this point, as White normally castles and may not necessarily follow up with d4.} cxd4 9. Nxd4 {White now has a classic (reverse) IQP position.} O-O {the obvious move for Black here would be to protect the Nc6, but if White takes it and the pawn after double captures on c6, the resulting positions would be unfavorable.} 10. O-O (10. Nxc6 bxc6 11. Bxc6 Bg4 {this is the problem with White's choice in this line. Black now develops his pieces with tempo and forces White into an awkward situation.} 12. f3 { compromising the pawn shield around White's king.} Rc8 13. O-O {is ugly but may be best.} (13. Bb7 Bb4+ 14. Nd2 Rc7 $17) (13. Bxd5 $2 Bb4+ $19) 13... Rxc6 14. fxg4 Qb8 $15) (10. Bxc6 bxc6 11. Nxc6 Qc7 {threatening the Nc6 and forming a battery against h2. Now White will not be able to castle and Black has a lead in development, which add up to more than a pawn's worth of compensation. For example} 12. Bxf6 Qxc6 13. Bd4 Bb4+ 14. Kf1 Bf5 $15) 10... Qc7 11. h3 { now we are out of the database. Kramink may have used this prophylatic move as an improvement to the following game:} (11. Nf3 Bg4 12. Bxf6 gxf6 13. Nc3 Bxh2+ 14. Kh1 Be5 15. Nxd5 Qd6 16. Nf4 Bxf4 17. exf4 Qxf4 18. Bxc6 bxc6 19. Qd4 Qh6+ 20. Nh2 Be6 21. Kg1 Rfd8 22. Qc3 Qg6 23. Rfd1 Rxd1+ 24. Rxd1 Bd5 25. f3 Re8 26. Ng4 Kg7 27. Kf2 h5 28. Ne3 h4 29. Rh1 Qg3+ 30. Kf1 Re5 31. Rh3 Qg5 32. Kf2 Be6 33. g4 Rc5 34. Qd2 Rd5 35. Qc3 Rc5 36. Qd2 Rd5 {1/2-1/2 (36) Agzamov,G (2485) -Geller,E (2545) Yerevan 1982}) 11... Bh2+ 12. Kh1 Be5 {this sequence involving forcing the king to h1, instead of simply moving ...Be5 to pin the knight, results in a position where white's king has fewer potential flight squares.} 13. Qc2 {protecting the Bb2 and therefore unpinning the knight, while exerting pressure down the c-file.} Bd7 14. Nf3 {this quasi-forces the exchange of bishops, as Black does not want to give White the long diagonal for free.} Bxb2 15. Qxb2 Rac8 16. Rc1 {this move is rather committal, as White's heavy pieces are now all on the queenside.} Qd6 {unpinning the Nc6.} 17. Nc3 {White finally gets all of his pieces developed; the queen's knight has been rather neglected until now.} Ne5 $6 {the normal strategic rule in playing IQP positions is for the player NOT possessing the IQP to exchange pieces; this leads to fewer complications and causing the pawn's weakness to become more pronounced. It's not clear why Black chose to ignore this rule, since the piece exchanges that follow are entirely voluntary.} ({something like } 17... a6 18. Bd3 Rfe8 {looks simple and balanced for Black.}) 18. Be2 $5 (18. Nxe5 {is the more obvious approach to the position.} Qxe5 19. Bxd7 Nxd7 20. Rc2 $14 {with the idea of following up with Rd1 and Ne2 to pressure the IQP and control the space in front of it.}) 18... Nxf3 {again unforced, but a logical follow-up to the previous knight move.} 19. Bxf3 {the engine rates the position as equal, although it's clearly easier for White to play against the IQP than it is for Black to come up with any real counterplay.} Qe5 20. Qd2 { unpinning the Nc3.} Be6 {the bishop being used as a "big pawn" is ugly but effective in supporting d5.} 21. Nb5 Bd7 {Black appears content to shuffle pieces at this stage, since he doesn't have anything better than a draw as a prospect.} 22. Qd4 {continuing to exchange pieces.} (22. Nxa7 $2 Ra8 {and the knight is out of squares.}) 22... Qxd4 23. Nxd4 a5 $6 {this seems to excessively loosen Black's position and grant White some initiative. The pawn becomes a significant weakness later in the endgame.} (23... a6) (23... Ne4) 24. g4 {This is the type of move that is sometimes difficult for Class players to envision, since it weakens the pawn shield in front of the king. However, Kramnik reacts appropriately to the situation and judges that the game has now reached the stage, with the queens off the board, of an endgame where his king needs to be activated. The advanced g-pawn also controls additional space.} h6 25. Kg2 $14 Rxc1 26. Rxc1 Rc8 27. Rb1 {an interesting choice. Basically, if White exchanges the rooks, he technically has a slight edge, but it's very doubtful he'll be able to do anything significant to compromise Black's position with only the two minor pieces remaining. With the text move, White retains the rook and Kramink evidently still hopes to squeeze out a win if Black falters.} Ra8 {perhaps this is intended as a threat to advance the a-pawn, although it doesn't seem very significant.} 28. Ne2 {White switches plans and repositions the knight to directly pressure the IQP.} g5 {I'm not sure what this was supposed to accomplish, as White's g-pawn was already restrained and the Ne2 can simply go to c3 instead of f4. It was probably better to use the tempo to get the rook back in the game.} (28... Rc8) 29. Nc3 Be6 {back to static defense of the pawn.} 30. Rd1 Rd8 {play is predictably revolving around the IQP. With the Rd8 unprotected, however, White can use this tactical situation to force the issue with his next move.} 31. e4 d4 { otherwise the pawn is lost.} 32. Kg3 {Kramnik does not rush to take the pawn, first removing the king from a potential tactic involving a check from a future ...Nf4.} ({For example} 32. Nb5 $6 d3 33. e5 Nd5 $11 {and the d3 pawn is tactically protected by the knight fork on f4.}) (32. Rd3 $5 {is also possible, blocking the pawn from advancing. The knight still cannot be taken due to the pin against the unprotected Rd8.}) 32... Rc8 33. Nb5 Nd7 {as the d-pawn is lost, black repositions the knight to a better square in the center, also blocking the e-pawn.} 34. Nxd4 Ne5 {although White is now a pawn up, Black has some positional compensation in the form of his better-placed pieces. In particular, the Rc8 threatens to penetrate White's back ranks.} 35. Be2 (35. Nxe6 $6 fxe6 36. Be2 Rc2 $11) 35... Rc3+ 36. f3 Kg7 37. Rd2 {relieving the Nd4 of the burden of guarding the c2 square against Black's rook.} Kf6 {Black is looking to get his king into the fight.} 38. Nf5 Bxf5 {a good decision by Black to exchange. The knight would be powerful on the 5th or 6th ranks (f5 or d6) and White's pawns are doubled as a result, making it a little harder for him to make progress.} 39. exf5 Ke7 $16 40. Rd5 {correctly and aggressively striking at the centralized knight and the undefended, advanced a-pawn.} Nc6 ( 40... Kf6 $5 {is possible, for example} 41. Rxa5 Rc2 {and Black will regain the pawn. One sample continuation:} 42. Kf2 Nc6 43. Rd5 {defending against ... Nd4, winning the pinned Be2.} Rxa2 $16) 41. Bb5 b6 42. h4 {getting the kingside majority moving.} f6 {shoring up g5 and creating a potential knight outpost on e5.} 43. Bxc6 {done in order to eliminate the knight, which otherwise would become strongly centralized on the dark square e5 and help restrain White's kingside advance.} Rxc6 44. hxg5 hxg5 {"all rook endings are drawn" is a saying that is not technically true, but is often the case because they are so hard to win. White's extra pawn is doubled and Black's rook and king are active, so it's not simply "a matter of technique" to win from this point.} 45. a4 Rc3 46. Rb5 Rc6 47. f4 {Kramnik decides to force the issue on the kingside immediately.} Rc3+ {in order to drive the king away from supporting f4. Of course, the drawback is abandoning the b6 pawn. As typical in rook endgames, one cannot get something for nothing. The engine assesses that conducting a straight exchange on f4 and continuing to protect the b-pawn would be better for Black.} 48. Kf2 gxf4 49. Rxb6 $18 {Komodo 8 shows White at over +4 at this point.} Kf7 {the king heads back to the kingside, with the evident idea of trying to make it to g5.} 50. b4 {this was an unnecessary concession.} (50. Rb5 Kg7 51. Rxa5 Kh6 (51... Rxb3 52. Ra8 $18 {and the a-pawn cannot be stopped by Black.}) 52. Ra8 $18) 50... Rc2+ 51. Kf3 Rc3+ 52. Kf2 { this seems to be where Kramnik passes up a good winning chance.} (52. Kxf4 Rc4+ 53. Kf3 axb4 54. a5 $18) 52... Rc2+ 53. Ke1 Rc1+ 54. Kd2 f3 {compared with the variation above, Black is doing much better. The Rc1 of course cannot be taken, since the f-pawn would then queen.} 55. Ke3 Rc3+ 56. Kf2 axb4 57. Rxb4 Kg7 { Black's king slowly gets closer to its goal.} 58. Kg3 (58. Rb8 $5 {is the best chance for White, according to the engine.} Rc4 59. Ra8 Rxg4 60. Kxf3) 58... Ra3 59. Rb7+ Kh6 60. Ra7 f2+ $1 {the key move for Black, which allows him to draw.} 61. Kxf2 Kg5 {the combination of the Ra3 cutting off the White king and the blockade on f6-g5 mean that White can no longer make progress. If White simplifies to having a passed pawn, Black's king and rook position mean that it will not be able to queen, lacking the White king's support.} 62. Ke2 { White nevertheless continues the game in the hopes of his opponent committing an error.} (62. Rg7+ Kf4 63. Rg6 Rxa4 64. Rxf6 Kxg4 $11) 62... Kxg4 63. Ra5 Kf4 64. Kd2 Ke4 65. Kc2 Kd4 66. Ra6 Rc3+ 67. Kb2 Rc4 68. Ra5 Rb4+ 69. Ka3 Kc4 70. Ra6 Rb3+ 71. Ka2 Rb4 72. Ra8 Kc5 73. Ka3 Rf4 74. Rb8 Rf1 75. Rb5+ Kc4 76. Kb2 Rf2+ 77. Kb1 Kc3 78. a5 Kc4 79. Rb7 Rxf5 80. a6 Ra5 81. a7 Kc5 82. Rf7 Kb6 83. Kc2 1/2-1/2