Annotated Game #168: A classic Classical Caro-Kann

This next tournament game features a significantly higher level of play than the previous two, although I get myself in trouble with a small but significant inaccuracy in the early middlegame.  The opening phase is a fine illustration of the Classical Caro-Kann (with 4...Bf5), which my opponent as White played knowledgeably.  I was quite comfortable after the typical pawn break 16...c5, but failed to assert my counterplay as necessary on move 19, handing the initiative over to my opponent.

This defense for some reason has a rather staid reputation, but as occurred in this game it often features opposite-side castling situations where both White and Black need to be careful and also relatively sophisticated with their play.  Here my opponent failed to go for the critical continuation on move 22, allowing me to regain near-equality and think of some possible counterplay of my own, eventually equalizing again.  After another significant error on move 36 on my part, in part due to tiredness' effect on my calculating skill, I weathered some significant pressure (with a bit of luck) and forced a transition to a position where I was down material but could hold the draw, despite my opponent's hopes of me making an error.

[Event "?"] [Site "?"] [Date "????.??.??"] [Round "?"] [White "Class B"] [Black "ChessAdmin"] [Result "1/2-1/2"] [ECO "B19"] [Annotator "ChessAdmin/Komodo 10"] [PlyCount "120"] {B19: Classical Caro-Kann: 4...Bf5 main line} 1. e4 c6 2. d4 d5 3. Nc3 dxe4 4. Nxe4 Bf5 5. Ng3 Bg6 6. h4 h6 7. Nf3 Nf6 8. h5 Bh7 9. Bd3 Bxd3 10. Qxd3 e6 11. Bd2 Nbd7 12. O-O-O Be7 13. Ne4 {this is played about 50 percent of the time, according to the database. White prefers to exchange off the somewhat passive Ng3.} Nxe4 14. Qxe4 Nf6 {there is no reason not to play this move, which develops the knight to a better square with tempo.} 15. Qd3 O-O {Black has postponed castling for long enough. Here I also contemplated playing ...Qd5, the other major option, but decided on safety first. This is the modern way to play the Caro-Kann, with opposite sides castling and more dynamic strategic tension. Castling queenside is also a legitimate option in many lines, however. } 16. Ne5 {taking advantage of the absence of the knight on d7 and occupying the central outpost. Black is fine, but needs to take care with White's sacrificial options, which now include sacrificing the knight on f7 or g6, along with the bishop sacrifice on h6.} c5 {generating counterplay is needed, otherwise White can simply proceed with organizing his kingside attack. This is a typical (and necessary) pawn break in the Classical Caro-Kann.} 17. dxc5 Qc7 {my first significant think, where I decide against the immediate recapture. The queen clears d8 for a rook and develops to a better square for its activity, while simultaneously pressuring e5 and c5.} 18. Qe2 Qxc5 { I had another significant think here, although not as long. I eventually decided that leading with the queen on c5 would allow better prospects against White's king position; most of the database games continue this way as well. However, it leaves the Be7 in a somewhat passive role.} (18... Bxc5 {is preferred by the engines.} 19. g4 Rac8 20. c3 Nd5 21. Kb1 Bd6 22. f4 Rfe8 23. Rdf1 b5 24. Bc1 f6 25. Nd3 e5 26. fxe5 Bxe5 27. Qf3 Qc4 28. Nxe5 Rxe5 29. Re1 b4 30. Rxe5 fxe5 31. Rd1 Rd8 32. cxb4 e4 33. Qf1 Qxb4 34. Ka1 Rf8 35. Qa6 Nb6 36. Qe2 Qc4 37. Qxc4+ Nxc4 38. Rd4 Rf1 39. Rxc4 {0-1 (38) Fedorchuk,S (2503) -Khenkin,I (2609) Ohrid 2001 CBM 084 [Lukacs]}) 19. g4 {my opponent plays aggressively and immediately advances a pawn for the attack, which is consistent with the needs of the position.} Rfd8 $6 {a significant inaccuracy. Black needs to assert his counterplay immediately as well, which is best on on the c-file.} (19... Rac8 20. c3 Bd6 21. f4 Qd5 $11) 20. g5 $14 Rac8 {now that the Nf6 is hanging, this move to the c-file has less impact, since Black will have to respond to the threat to the knight after White's obvious next move.} ( 20... Nd7 21. Nxd7 Rxd7 22. gxh6 Bg5 $14) 21. c3 hxg5 {another significant think here. I concluded that getting rid of the g-pawn would do the most to reduce White's attacking chances. Komodo doesn't fully agree with me.} (21... Nd7 22. Nxd7 Rxd7 23. gxh6 Bg5 $14) 22. Bxg5 {not the critical continuation.} ( 22. h6 g6 $14 {so White is better here, but has to find the not so obvious idea of moving a rook to e1 to continue the attack.} 23. Rde1 (23. Bxg5 Rxd1+ 24. Rxd1 Nd5 25. Bd2 Bf6 {and Black is OK.}) 23... Qd5 (23... Nh7 $2 24. Nxf7 { a thematic sacrifice} Kxf7 25. Qxe6+ Kf8 26. Qxg6 $16)) 22... Rxd1+ {this seemed the easiest way to further reduce White's attacking chances, getting a rook off the h-file and additional material off the board.} (22... Rd5 { is suggested by the engine and was a possibility I considered for a while. Eventually I didn't see enough utility in the move after White's response f4.} 23. f4 Rcd8 {and now Black can also liquidate material effectively, for example } 24. h6 Rxd1+ 25. Rxd1 Rxd1+ 26. Qxd1 gxh6 27. Bxh6 Qe3+ 28. Kb1 $11) 23. Rxd1 Nd5 {played with the idea of getting counterplay going. White's c3 pawn is currently pinned and can be a good sacrifice target, while the knight can also go to b4 as a result.} (23... Rd8 {would continue of the original idea of reducing material.}) 24. Bd2 {Essentially a forced concession by White, as the alternative of swapping bishops on e7 would leave him with even fewer prospects.} Bf6 {I thought for a while here and was unsure if the bishop move would be the most effective, but in the end decided that the added pressure on the long diagonal was a good thing. I considered that my opponent's most likely response (which he did play) would lead to an equal game.} (24... Nb4 { I also considered, but did not see much of a point after} 25. Kb1 {The engine offers} Nc6 {as a possibility, however, which is a good way of getting rid of the Ne5 without ending up with a knight vs. bishop situation as in the game continuation.}) 25. Nd7 (25. h6 $5 {is of course more testing.} Qc7 26. f4 Bxe5 (26... g6 $5) 27. fxe5 Qc4 28. Qg2 g6 29. a3 Kh7 30. Rf1 Rf8 31. Bg5 Qc7 { with a slight edge to White.}) 25... Qa5 $11 {the correct response, threatening the a-pawn and maintaining the latent pressure along the 5th rank against the h5 pawn. An example of dynamic play.} (25... Qc7 {would of course be perfectly adequate as well.}) 26. Nxf6+ Nxf6 27. Kb1 {the obvious "safe" choice, jettisoning the h-pawn, but not necessary.} (27. h6 {is an aggressive try.} Qxa2 28. Qe5 Qa1+ 29. Kc2 Qa4+ 30. Kc1 Qe4 $11) (27. Qf3 Qxa2 28. Qxb7 Qa1+ 29. Kc2 Qa4+ 30. Kc1 {and Black can head for a draw by repetition.}) 27... Qxh5 {here I thought about the intermediate check on f5, but did not see how it would be of good use. The engine considers the alternative to be superior, however, likely due to the fact that the king on a1 is further from the action in the endgame, while also being more vulnerable to back-rank threats. If the king goes to c1, then the c-pawn is again pinned and the a-pawn again unprotected.} (27... Qf5+ 28. Ka1 Qxh5 29. Qxh5 Nxh5 {White's bishop is better than the knight and the rook is better positioned on d1 as well, but a pawn is a pawn.} 30. Be3 a6 31. Rd7 Rb8 $15) 28. Qxh5 {essentially forced, otherwise Black's queen is much better positioned for action.} Nxh5 {now we have the same position as in the above variation, only with White's king one square closer to the center, making the value of the intermediate check (a "tempo move") more clear.} 29. Be3 b6 {blocking the Be3.} (29... a6 {was the other possibility I considered. Both it and the game continuation are OK for Black, although it might have been a bit easier to play this variation's positions.} 30. Rd7 b5 31. Rd6 Ra8 $11) 30. Rd7 Ra8 31. Kc2 (31. Bd4 Nf6 32. Bxf6 gxf6 $11) 31... Nf6 32. Rb7 Kf8 {played on the general principle of bringing the king closer to the center, but also with an eye towards possibly trapping White's rook, if he makes an error.} 33. Bf4 Ke8 34. c4 {taking the d5 square away from the knight.} Nd7 {White is still fine, but the Rb7 is uncomfortable with so few squares.} 35. b4 (35. Bd6 Rc8 36. Bc7 a5 $11) 35... Kd8 {around here I started to get tired, which was reflected in the quality of my calculations, although I'm still able to see key ideas. The immediate ...e5 would be better, kicking the bishop first.} (35... e5 36. Be3 Kd8 37. c5 {otherwise the rook is trapped} bxc5 38. bxc5 Rc8 $11) 36. Bd6 e5 $6 {as I've often noted in my analysis, another example of the right idea but played a tempo too late, thereby creating problems. Here I was thinking about trapping the bishop, too, but it doesn't work out.} (36... Kc8 37. Rc7+ Kd8 $11 {with the idea of ...a5 to follow up.}) 37. Kd3 $14 {simple yet effective. Now the c-pawn is protected, reducing my potential threats.} f5 $2 {here I needed to play the key ...a5 idea. The text move looks good, keeping the king from penetrating via e4, but now White's queenside pawns can effectively mobilize.} (37... a5 38. c5 { now this advance is opposed} Rc8 39. bxa5 bxc5 40. Ra7 $14) 38. f3 $6 $11 { simply marches past the door to victory, comments Komodo via the Fritz interface.} (38. c5 $1 {wins:} bxc5 39. bxc5 Kc8 (39... Rc8 40. Kc4 a6 41. Kd5 $18) 40. c6 $1 Nb8 41. Bxe5 $18) 38... g5 $6 {still not adequately seeing the danger on the c-file.} (38... Kc8 39. Rc7+ Kd8 40. c5 bxc5 41. bxc5 Rc8 42. Rxa7 Nxc5+ $11) 39. c5 $16 {now my opponent puts on the pressure, but this time he's the one who plays the correct move a tempo late, to less effect.} Rc8 40. Rxa7 $6 {my opponent's choice of going for the material allows me to equalize again, now that my rook can become active.} (40. Kc4 $5 bxc5 41. bxc5 $16 {maintains the pressure.}) 40... bxc5 $11 41. bxc5 Nxc5+ 42. Ke3 f4+ { a tough decision and a relatively long think here, but a correct one. White's king is pushed back by the combination of knight and pawns.} 43. Ke2 e4 { played after another relatively long, difficult think. I still harbored some hopes of being able to do something with the kingside majority, but this is not possible in the end.} (43... Rc6 {would have been an easier path to a draw. } 44. Be7+ (44. Bxc5 Rxc5 45. a4 $11) 44... Ke8 45. Bxg5 Ne6) 44. Be7+ { I had seen this far...} Ke8 45. Bxg5 {but now I realized that my planned ... Ne6 would fail to Re7+, so I have no choice but to liquidate and go for a drawn endgame.} exf3+ 46. Kxf3 Rc6 {I was pleased to find this idea, in fact the best, which is in keeping with the principle of emphasizing rook activity. After the f-pawn goes, I force a rook exchange and White cannot win with his remaining material.} 47. Bxf4 Ra6 48. Rxa6 Nxa6 {we could have stopped playing here, but my opponent wanted to see if I would commit an error. As long as the knight avoids domination or exchange by the bishop and Black gets his king over in front of the pawn, it's a draw.} 49. Ke4 Kd7 50. Kd4 Kc6 51. Kc4 Nc5 52. Be3 Na6 53. a4 Kb7 {an illustration of how ineffective a lone bishop can be in the endgame; in this case, Black dominates the light squares and therefore draws.} 54. Kb5 Nc7+ 55. Ka5 Nd5 56. Bd4 Nc7 57. Kb4 Ka6 58. Be5 Nd5+ 59. Kb3 Nb6 {the knight now can fulfill his destiny and end the game.} 60. Kb4 Nxa4 1/2-1/2