Annotated Game #299: What the Caro-Kann is good for (part 2)

The below third-round tournament game is thematically similar to Annotated Game #294, in which White first presses Black - this time in a Panov variation of the Caro-Kann - then missteps, not sensing the danger. This is typically what Black is going for in the Caro-Kann overall, being patient and then seizing counterplay when it becomes available. Here I deliberately chose the most dynamic middlegame plan on move 13, after a long think, which is what opens up the opportunity later on for White to stumble.

[Event "?"] [White "Class C"] [Black "ChessAdmin"] [Site "?"] [Round "?"] [Annotator "ChessAdmin/Dragon 3.2"] [Result "0-1"] [Date "????.??.??"] [ECO "B14"] [PlyCount "56"] [GameId "2108266128867374"] {[%evp 0,56,25,16,59,71,77,47,47,24,10,3,32,20,39,35,25,13,4,-10,15,-17,-20,-17,-10,-14,-5,-2,0,0,14,12,3,3,3,0,0,-64,-12,-87,-96,-88,-94,-77,-73,-68,-26,-149,-213,-231,-231,-231,-238,-264,-204,-693,-669,-29993,-29994]} 1. e4 c6 2. d4 d5 3. exd5 cxd5 4. c4 {Panov variation.} Nf6 5. Nc3 e6 {this seemed to be the end of my opponent's opening familiarity.} 6. Nf3 Be7 7. Bf4 {a non-standard move here.} O-O {no reason not to castle here, although I had to check to see if the Nb5-c7 idea would work for White.} 8. Rc1 (8. Nb5 {and Black can simply ignore the supposed threat to play Nc7, as the Bf4 is vulnerable. For example} Nc6 9. Nc7 $2 Nh5 10. Nxa8 Nxf4 $19 {and White's knight is lost.}) 8... Nc6 (8... b6 {is the engine preference, preparing to develop the bishop on the long diagonal and restraining c4-c5.}) 9. c5 {I was not concerned about this advance at the time, since in analagous positions it's not a real threat. However, here it is worth preventing, since it creates a potential outpost on d6 which both the Bf4 and a future Nb5 could use.} Nh5 {not a bad idea to chase the bishop, but it leaves me a little underdeveloped.} (9... Ne4 $5 {and Black would welcome a trade on e4, which would expose the suddenly weak d4 pawn.} 10. Nxe4 dxe4 11. Ne5 Nxd4 $19) (9... a6) 10. Bg3 Nxg3 {this was perhaps a little too hasty, giving White the half-open h-file while removing a key defender on the kingside. It is still theoretically balanced, but it causes stress for me later.} (10... a6 {again would restrict the Nc3 from reaching b5.}) (10... f5 $5 {the engine considers White vulnerable to a kingside pawn advance, the immediate threat being f5-f4.} 11. Be5 g5 $11) 11. hxg3 Bf6 12. Bd3 g6 {I felt for defensive purposes it was better to blunt the bishop's diagonal, although ...h6 would be good as well.} 13. Qd2 Nxd4 {long think here. I can in fact get away with the pawn capture, which is the most dynamic choice and the engine agrees is best (slightly).} (13... Bg7 $11 {is the simpler way to play.}) 14. Qh6 Nxf3+ 15. gxf3 Re8 16. Qxh7+ Kf8 17. Kf1 {I had correctly calculated/visualized to this point. Black has several options here, I went with the one I could best visualize as limiting White's counterplay.} a6 {I decided prophylaxis was most important here, taking away the b5 square from my opponent and not letting the knight get into the fight that way.} 18. Ne2 $6 {looking to redeploy, but this gives me time to gain some piece activity. This type of error was why I chose this middlegame approach in the first place, since it gave my opponent more difficult problems to solve, and along with that more opportunities to go wrong.} (18. Kg2 $11) 18... e5 $17 {while a nice way to help restrict the Ne2, it more importantly also frees the Bc8 on the diagonal.} (18... Bg7 $17 {I also considered, with the point being that it opens up the diagonal for the Qd8's use. The engine considers it best; during the game, I was unable to confirm its safety during calculation, so chose the text move.}) 19. Nc3 Be6 {I had envisioned this defensive idea from move 17. Black has now consolidated the extra pawn.} 20. Kg2 {connecting the rooks and thereby renewing plans for the h-file attack, but with a loss of time.} Ke7 {finally getting out of the corner, although the king is not safe yet, with tactics having to be calculated.} 21. Qh6 {the queen of course was in danger of being trapped, with the path to h8 cleared for a rook.} (21. Bxg6 $2 Rh8 $19) 21... Rh8 $6 {this allows the queen to escape without consequence.} (21... g5 $1 $17 {keep the pressure on.}) 22. Qd2 $11 Qd7 {...Bh3+ looks dangerous, but I cannot actually make progess with it - until my opponent attempts to block it with} 23. g4 $2 {a decision made based on the fact that the g-pawn visually looks safe, being protected by another pawn. However, this is tactically incorrect.} (23. Bc2 Bh3+ 24. Kg1 {and the position is certainly awkward-looking for White's king, but there is no breakthrough.}) 23... Bxg4 $1 $19 {now the White king's shield is broken and Black is winning.} 24. Be2 (24. fxg4 $4 Qxg4+ 25. Kf1 Rxh1#) 24... Bh3+ 25. Kg1 {otherwise it's mate} Be6 {this was the simple approach for my still-stressed brain, at the time.} (25... Rad8 {is preferred by the engine, bringing the rook into play and protecting the d-pawn at the same time.}) 26. Kg2 (26. Rxh8 Rxh8 27. Bf1 $19 {would have forced Black to maneuver some more to get at the White king.}) 26... Rh5 (26... Bh3+ 27. Kg1 Rad8 {would free up the queen by reinforcing d5.}) 27. b4 $2 (27. Rxh5 gxh5 28. Rg1 Rg8+ 29. Kh1 {and there's no immediate win, although White will still lose once Black consolidates the center.}) 27... Bh3+ {now the bishop returns with deadly force.} 28. Kh2 Bf1+ $1 {and the queen will be able to come to h3 and then mate.} 0-1