Annotated Game #110: Failed opening experiment

The following game from an ongoing Slow Chess League tournament features a failed opening experiment, in this case on move 9 for Black.  I decided to avoid the main line and venture off into an "easier" sideline to remember, involving an offer to exchange queens and simplify, which would work in Black's favor.  My opponent correctly rejected the offer (after some thought) and went on to win the game.

It is worth underlining the fact that the idea itself did not immediately lose, but it put Black in a less desirable position developmentally versus the main line, essentially a tempo down on developing the kingside, which gives White some additional tactical possibilities. What did lose more or less immediately - but not obviously so - was my decision to castle queenside a few moves after White's aggressive response with 11. c4.  From previous analysis of similar games I knew that this was a relatively risky decision, but as far as I could see, White could not directly exploit it with good defense by Black.  Unfortunately, I was wrong and the decision doomed me strategically.

The game is unusual in that respect, as normally a single error is either an immediate blunder (obviously non-recoverable) or can be recovered from later on with better play.  Here Black is put on an inexorable path of doom, which only materializes a number of moves later.  My opponent deserves full credit for taking the time to work out how to do this, although he missed a chance on move 19 to more quickly put me away.

In this case, although the opening experiment was a failure, it's helped give me more insight into the opening and middlegame dynamics for future use in the main line with an immediate ...e6 (and more confidence in that being the best way to proceed).

[Event "DHLC Slow 1-2 Pairing #4"] [Site "Chess.com"] [Date "2013.12.10"] [Round "2"] [White "nate23"] [Black "ChessAdmin_01"] [Result "1-0"] [ECO "B19"] [WhiteElo "1707"] [BlackElo "1432"] [Annotator "ChessAdmin/Houdini"] [PlyCount "49"] [EventDate "2013.??.??"] [TimeControl "45"] {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. Bd3 {White decides to skip the more usual h4-h5 push.} Bxd3 9. Qxd3 Qa5+ {entering the sideline} 10. Bd2 Qa6 11. c4 $146 {although a novelty in this position, in the analagous line with the h4-h5 push thrown in, it scores 90 percent!} (11. Qxa6 {was what I was hoping for} Nxa6 12. O-O {1/2-1/2 (12) Golubovic,B (2450)-Lalic,B (2545) Pula 1998}) ( 11. Ne5 e6 12. O-O Qxd3 13. Nxd3 Bd6 14. Bf4 Bxf4 15. Nxf4 O-O 16. Rfe1 Rd8 17. c3 Nbd7 18. Ne4 Nxe4 19. Rxe4 c5 20. dxc5 Nxc5 21. Rc4 b6 22. Kf1 Rd2 23. b4 Na4 24. Rc7 Rad8 25. Rxa7 Nxc3 {Kirjuskin,A-Svetlov,D (2104) Sochi 2007 0-1 (39)}) 11... e6 {Prevents intrusion on f5, notes Houdini via the Fritz interface. Other than being the obvious move to release the Bf8 and help control d5, this was in fact a consideration of mine.} 12. O-O Nbd7 {by this point Black is behind in development on his kingside, thanks to the queen moves.} 13. Rfe1 O-O-O $2 (13... Be7 $5 $14 {is still the best move in the position, although now White has some tactical threats along the e-file that Black needs to worry about. Although objectively they are not decisive, they worried me to the extent that I decided to avoid them by castling queenside. Normally Black would already be castled by now, the fact he is behind schedule makes him more vulnerable.} 14. Nf5 {for example was one major tactical idea of concern, taking advantage of the Be7 only being able to be defended by the king, although the knight sacrifice cannot be immediately played; in my original calculations I missed the idea of interposing the knight to defend e7. } (14. a4 O-O 15. Nf5 exf5 16. Rxe7 Rae8 17. Rae1 Rxe7 18. Rxe7 Qxa4 19. Qxf5 Rd8 $11) 14... exf5 15. Qe2 Ne4 $1 {is winning for Black.}) 14. b4 $16 { this is evaluated by Houdini as the best choice, with White immediately taking the initiative against Black's king position and threatening b4-b5, trying to crack open the c-file.} b5 $2 {this seemed to stop the attack, but in fact leads inexorably to a White victory. Houdini considers White the equivalent of two pawns up already.} (14... Qb6 $5 $16) 15. c5 $18 Qb7 (15... Qa4 $18 { is what the engine recommends and what I considered as the only alternative, as Black needs to get his queen back in the game somehow. After some thought, however, I rejected it thinking that the queen would be too far offside and could potentially come under threat of being trapped.}) 16. a4 a6 17. axb5 { at this point I realized simply capturing with the a-pawn, which stops the pawn advance, would in fact give White a win via penetrating on the open a-file.} Qxb5 18. Qxb5 axb5 {this loses as well, but I did not see the immediate rejoinder with Ra8+. I rejected the also losing recapture with the c-pawn because I thought White's path to victory there would be more obvious.} (18... cxb5 19. Rxa6 $18) 19. Ra7 $6 {my opponent also misses the better line, which I had seen only after making my previous move (i.e. too late). Black is still badly off, but not completely knocked out.} (19. Ra8+ {finishes off the opponent, says Houdini.} Nb8 20. Ne5 $18 {wins, for example} Rxd4 21. Rxb8+ Kxb8 22. Nxc6+) 19... Nd5 {the best defense. At this point I thought I could salvage the game.} 20. Rea1 Nb8 $2 {I fail to realize the power of White's next move, which targets the f7 vulnerability, having concentrated instead on my king.} (20... f6 {is the best chance to fight on.}) 21. Ne5 $1 {now it is essentially over for Black, as White's forces run rampant over his position, although I gamely play on for a few more moves.} Rg8 22. Ra8 Nc7 23. Rxb8+ $1 { I missed this sacrifice as well, although it didn't make a difference by this point.} Kxb8 24. Nxc6+ Kb7 25. Nxd8+ 1-0