Annotated Game #197: Play the long game when needing a win

Having lost in the previous two rounds, including rather shamefully in round 3, I very much needed a turnaround win in this tournament.  "Needing" a win can, however, be a dangerous state of mind, like when gamblers keep making larger and riskier bets to try to catch back up to where they think they should be; it rarely ends up well.  Here I will give myself credit for having enough patience to "play the long game" and recognize the need to patiently maneuver, rather than try to break through prematurely, although my play was not necessarily optimal along the way.

There are a couple of key strategic moments that lead to the win.  The first comes at move 26, where I correctly realized that pawn breaks on the queenside, where both my opponent and I had castled, would favor me (Black).  About 20 moves later in a double rook endgame, I find the final breakthrough idea, involving a temporary rook sacrifice with a deflection tactic (which the engine awards a '!!' in its analysis).

That said, this game's analysis is perhaps even more valuable for me in the long term for the missed ideas, for both myself and my opponent, which will help me refine my understanding of the middlegame structures in the Classical Caro-Kann.

[Event "?"] [Site "?"] [Date "????.??.??"] [Round "?"] [White "Class C"] [Black "ChessAdmin"] [Result "0-1"] [ECO "B18"] [Annotator "ChessAdmin/Komodo 11.2"] [PlyCount "122"] {[%mdl 8256] B18: Classical Caro-Kann: 4...Bf5 sidelines} 1. e4 c6 2. d4 d5 3. Nc3 dxe4 4. Nxe4 Bf5 5. Ng3 Bg6 6. Nf3 Nf6 7. Bd3 {a solid but unambitious continuation for White.} e6 {I judged it better to make a developing move (releasing the Bf8) rather than exchanging on d3. Having the bishop exchanged on g6 can sometimes weaken Black's king position, but here it's not yet a concern. Primarily Black has to worry about sacrifices on g6 that undermine the e6 pawn, and/or play up the h-file once the king is castled.} 8. Bg5 Be7 9. Bxg6 hxg6 10. Qd3 Nbd7 {it's standard to develop the queen's knight before castling, in part to provide the option of castling queenside.} 11. O-O-O { consistent with the idea of exchanging on g6 and hoping for active play on the kingside.} Nd5 {my plan here is to clarify the situation on the kingside by encouraging the trade of the Bg5, then castle queenside, which I felt was more solid than castling kingside. Black should be careful about bringing a knight to d5 in the Classical Caro-Kann, however, when it can be chased off by the c-pawn.} (11... Qc7 $5) 12. Bxe7 {my opponent goes for the obvious response, exchanging on e7.} (12. h4 {would be a more challenging response, putting the onus back on Black. Exchanging on g5 would not be good, as the h-file could then be opened to White's benefit.} Bxg5+ $2 (12... b5 {is the engine's choice, starting immediate counterplay on the queenside}) 13. hxg5 Rxh1 14. Rxh1 $16) 12... Qxe7 13. Qd2 O-O-O 14. Ne2 (14. c4 $5 Nc7 $14) 14... N7f6 (14... e5 { instead would be a thematic pawn break. Black is well positioned to play in the center.} 15. Nc3 Nxc3 16. Qxc3 e4 $11 {the pawn can be reinforced by ...f5 and Black has a comfortable, if no more than equal, game.}) 15. Kb1 {keeping an eye on the weak a-pawn and clearing the c1 square.} Ne4 {the original intent behind the previous knight move, taking an active central position.} ( 15... Ng4 $6 {hitting the f2 square looks tempting, but White can protect everything and effectively re-deploy his Ne2 at the same time.} 16. Nc1 $14 { and there are no good follow-ups to the previous one-move threat.}) 16. Qe1 { forced} Ndf6 {here I was trying to anticipate a c4 push and proactively re-deploy the knight.} (16... Qb4 {Komodo prefers this more assertive approach, activating the queen and preventing c4.} 17. c3 (17. Qxb4 Nxb4 18. Rhf1 g5 19. h3 f6 $11) 17... Qb5 18. Ka1 $11) 17. Nd2 {it's difficult here for White to come up with a useful plan, although the position is equal.} (17. h3 Kb8 $11) 17... Nxd2+ {the correct decision, improving the relative value of my minor pieces.} 18. Rxd2 Ne4 {obvious, but unimaginative.} (18... e5 $5 {would be a bit more challenging.}) 19. Rd1 Qf6 {the right general idea, of activating the queen, here with the intention of pressuring both f2 and d4. However, g5 may have been a better square for the queen, pressuring the g-file and the d2 square.} 20. f3 {the obvious reaction.} Nd6 {the position here is quite balanced now. It will require patient maneuvering.} 21. Ng3 Nb5 {Increases the pressure on d4, but again this is easily solved by White.} 22. c3 Rd7 { continuing with the single-minded idea of building up pressure on the d-file.} (22... Qf4 {would at least move the queen to a better square.}) 23. Ka1 (23. Ne4 {is an idea that my opponent seemed to miss. Although it's not enough for a real advantage, initiative shifts to White and Black has to be careful about things like covering the c5 square.} Qf5 24. Qe3 b6 $11) 23... Rhd8 {the problem with this is that the rooks now both have less space to work with, and the Ne4 idea gets better as a result. Luckily my opponent fails to find it.} ( 23... Nd6) 24. a4 {White makes the decision to weaken his kingside shield, apparently optimistic about the pawn push.} ({Instead} 24. Ne4 Qe7 25. Nc5 Rd6 $14 {is rather awkward for Black.}) 24... Nd6 25. Rd3 Qe7 {redeploying now with an eye toward the weakened queenside.} 26. b3 $6 {although this covers c4, it makes the next move more effective in punching holes in White's pawn shield. } (26. Qe2 Nf5 $11) 26... b5 {this break favors Black, who is better positioned with both the heavy pieces and his knight to exploit the resulting holes on the queenside.} 27. axb5 $6 {this simply plays into my plan. White instead should move the queen onto a better defensive square, for example e2 (covering the 2nd rank) or b1.} Nxb5 $17 28. Qc1 c5 {the best follow-up. Now the rooks on the d-file can make their pressure felt.} (28... e5 {is not as effective due to} 29. Re3 $15 {pinning the e-pawn and getting the rook away from the d-file threat.}) 29. Ne2 e5 {with the added pressure on d4, now this move is effective.} 30. d5 (30. Re3 f6 31. f4 e4 $17) 30... e4 $2 {this looks aggressive but would allow White to stablize the center.} (30... Rxd5 $5 { is simple and breaks through immediately.} 31. Rxd5 Rxd5 $17 32. Rd1 Rxd1 33. Qxd1 Qd6 {heading for a pawn-up endgame.}) 31. fxe4 $6 (31. Re3 {holds things together.} Qf6 (31... f5 $6 32. c4 Nd4 33. fxe4 $14) 32. fxe4 $11) 31... Qxe4 { Black forks: d3, g2+e2} 32. Qe3 {now White forks: c5+e4} (32. Re3 $5 Qxg2 33. c4 Nd6 $15) 32... Qxg2 {after some thought, I mis-evaluated the possible continuations, although the text move is still fine for Black, and perhaps represents the best practical chances for an advantage.} (32... Qxe3 33. Rxe3 Rxd5 $15 {and White has some compensation for the pawn, although the engine doesn't think it's enough to offset Black's advantage. I was worried about} 34. c4 {but} Rd1+ 35. Rxd1 Rxd1+ 36. Kb2 Nd6 $15 {should be fine, because after} 37. Re7 {White's rook can't take advantage of the 7th rank due to the rook fork on d2.}) 33. Qf3 $2 (33. Rhd1 {is the only good defensive move here, but White I'm sure didn't want to abandon the h-pawn.} c4 (33... Qxh2) 34. R3d2 Rxd5 35. Nd4 $11) (33. Qxc5+ {doesn't work, although it's a rougher ride for Black:} Nc7 (33... Rc7 $15 {is perhaps the easier route to go}) 34. Rhd1 $5 Qxe2 35. d6 Rh8 36. R1d2 (36. dxc7 $2 Rxd3 37. Rxd3 Qxd3 38. Ka2) 36... Qe4 37. Rd4 Qh1+ 38. Rd1 Qb7 39. dxc7 Qxc7 $17) 33... Qxf3 {now I make the correct evalution and exchange queens.} 34. Rxf3 f6 {here I choose safety over activity, which is not usually the way to go in rook endings. It's still enough to maintain the advantage, though.} (34... Rxd5 $1 35. c4 Rd1+ {we saw this idea in a previous variation} 36. Rxd1 Rxd1+ 37. Kb2 Rd2+ 38. Kc1 Rxe2 39. cxb5 Re7 $19 {and now Black can consolidate the two-pawn advantage without much trouble.}) 35. c4 Nd6 {this looked like the obvious move to me, but the engine disagrees. It also again shows how rook activity should be maximized.} ( 35... Re8 $5 36. Rf2 Rde7 $17) 36. Nc3 {this is a much less effective square for the White knight. Evidently my opponent's idea was to cover the e4 square.} (36. Nf4 $5 $14 {goes after the weak kingside g-pawn, en route to an excellent post at e6.}) 36... Re8 {now I start activating the rooks.} 37. Rhf1 (37. Na4 Rc7 $15) 37... Rde7 38. R1f2 $6 {this doesn't make a lot of sense, as the knight is currently covering the e2 square, so penetration on the 2nd rank isn't an immediate concern.} Ne4 (38... g5 $5 {looks like a good preliminary move, protecting the g-pawn and threatening ...g4 at some point, as White has nothing constructive to do in the meantime.}) 39. Nxe4 Rxe4 {here I felt confident that although White has the passed d-pawn, my rooks were better and could do more damage with White's knight out of the way. It's a somewhat premature simplification, though, and could allow White to more easily equalize.} 40. Kb1 (40. Kb2 {would be better, protecting the b-pawn and getting closer to the action.}) 40... Re1+ 41. Kc2 Kd7 (41... Rh8 $5) 42. Kd2 ( 42. h4 {is the key idea for White, fixing the g-pawn on g6 and allowing White to pressure on the g-file, for example} R1e4 43. Rg2 Rxh4 44. Rxg6 Re7 $11) 42... a5 {not a bad move, but both I and my opponent continue to ignore the ideas around g5 for Black and h4 for White.} 43. Rg3 Ra1 {the idea being to switch focus and break through on the queenside.} 44. Rgg2 $2 (44. Kc3 { and White hangs on} g5 45. h4 $11) 44... a4 45. bxa4 {it's better to take than to allow Black to create a passed a-pawn, but White is still in a great deal of difficulty.} Rxa4 46. Kd3 g5 {ironically, this is no longer Black's best move, although it is still good.} (46... Ra3+ 47. Kc2 g5 $19) 47. Rc2 $2 (47. Ra2 $5 {this is the defensive idea for White that the rook check on a3 would have prevented.} Rxa2 48. Rxa2 $17) 47... Ra3+ {now I find the idea.} 48. Rc3 { this would be an equally good defense, except for} Re3+ $3 {Komodo gave the exclamation points via the Fritz interface, so I've left them in as coming from an objective source. This is an aesthetically pleasing deflection tactic that forces a breakthrough on the queenside.} 49. Kxe3 Rxc3+ 50. Ke4 Rxc4+ 51. Kf5 Rd4 52. Kg6 Rxd5 53. Kxg7 Rd6 {not the quickest route to victory, but I was playing conservatively to keep the win in hand.} 54. Kg6 c4 {passed pawns must be pushed!} 55. Rc2 Rc6 56. Kf5 Rc5+ {although this gives back a pawn, it permanently bars White's king from the fight to prevent the pawn from queening. } 57. Kxf6 Kc6 {now Black wins with a simple king march.} 58. Kg6 Kb5 59. h3 c3 60. Rxc3 Rxc3 61. Kxg5 Rxh3 0-1