Annotated Game #217: Recovering in the middlegame

After my first-round victory against an Expert (Annotated Game #216), I had a short draw against an equally-rated player. My rule these days is not to take draws unless the position is in fact drawn with no real play left. Although it was still technically the middlegame, I would say it met the criteria and the engine assessment corroborates that, so I don't feel bad about the result. It also let me conserve energy for this next game, which was again against someone 300+ rating points higher.

The story of the opening into the early middlegame is unfortunately a familiar one. In an English vs. Queen's Gambit Declined (QGD) structure, I do fine for the first 10 moves of "book" and then flounder in an unfamiliar position. One of the insights I've had is that this is in fact a completely normal phenomenon. The point being, learn the ideas of the position through game analysis afterwards, so the next time it will be familiar. I believe this is one of the most powerful ways of improving your game on a practical level. It also means that frequent tournament/serious chess and post-game analysis is necessary.

Unlike in a number of previous games, I manage to recover after the rather silly 10. Na4?! and break the trend giving Black the initiative. It's interesting to see how early trends in a game often take psychological hold and a small but real advantage for one side just keeps getting (unnecessarily) larger. Here, Black has pressure in a complex position through move 13, then allows me in the next several moves to simplify and improve the relative position of my pieces. By move 17 I feel much better about the position and by move 22 things are completely level.

However, that doesn't mean that there were no more opportunities to go wrong. My opponent kept trying to create chances for me to degrade my position, for example on moves 37 and 41. The latter one is instructive, as one typical Class player error is to always take an even material exchange, in the belief that it will lead to a quicker draw. That's not what the position demanded, though, and the draw was sealed soon afterwards.

One GM comment I recall reading a while back noted that against lower-rated players, masters can often get significantly behind coming out of the opening, but then draw or win in the end after their opponents make a key error, typically close to or in the endgame. This seemed to be my opponent's thinking as well, which I can't fault him for. One of the differences in this game is that after recovering in the middlegame, I did not let up in my focus and calculation, assuming that it would be drawn; games don't magically end themselves (or if they do, it's likely because of a blunder on your part). So although my level of play in the first part was sub-par, it was good to see that I had what it took to go the full distance in the game.

[Event "?"] [Site "?"] [Date "????.??.??"] [Round "?"] [White "ChessAdmin"] [Black "Expert"] [Result "1/2-1/2"] [ECO "A13"] [Annotator "ChessAdmin/Komodo 11.2"] [PlyCount "83"] {A13: English Opening: 1...e6} 1. c4 e6 2. Nf3 d5 {going for a QGD setup} 3. b3 Nf6 4. Bb2 Be7 5. e3 O-O 6. Be2 c5 7. O-O Nc6 8. cxd5 Nxd5 9. Nc3 {the most common move in the position, according to the database, but not particularly well scoring (44 percent).} (9. d4 {stakes more of a claim in the center.}) (9. a3 $5 {goes for more of a Hedgehog type structure and scores 55 percent in the database.}) 9... Bf6 10. Na4 $6 {this is just a fancy way of misplacing the knight. The idea was to reposition it on c4 after a bishop exchange on b2.} ( 10. Qc1 {scores the best in the database, at 56 percent.} Nxc3 11. Bxc3 Bxc3 12. Qxc3 Qe7 13. Rfc1 b6 14. d4 Bb7 15. dxc5 Rfc8 16. b4 bxc5 17. bxc5 a5 18. a3 Rc7 19. Rab1 Rac8 20. Rxb7 Rxb7 21. Ba6 Rcb8 22. Bxb7 Qxb7 23. Nd4 a4 24. h3 g6 25. Qc4 Nxd4 26. exd4 Qc6 27. Rd1 Rd8 28. f3 h5 29. Rb1 Kh7 30. Rb4 Ra8 31. Qe2 Rd8 32. Qe4 Qc7 33. Rb7 Qg3 34. Rxf7+ Kg8 35. Qxe6 {1-0 (35) Lagarde,M (2586)-Skuhala,L (2059) chess.com INT 2018}) (10. Rc1 {seems natural as well, here's a high-level example:} b6 11. Qc2 Bb7 12. Rfd1 Rc8 13. Qb1 Nxc3 14. dxc3 Qc7 15. Nd2 Ne5 16. f4 Qc6 17. e4 Ng6 18. g3 c4 19. Nxc4 Rfd8 20. Bf3 b5 21. Ne3 Qb6 22. Kf2 e5 23. Rxd8+ Rxd8 24. Rd1 exf4 25. Rxd8+ Qxd8 26. Nd5 Be5 27. Qd3 fxg3+ 28. hxg3 Qg5 29. Be2 h5 30. Qe3 Qxe3+ 31. Kxe3 Bxd5 32. exd5 b4 33. Bxh5 bxc3 34. Bc1 Bxg3 35. Kd3 Nf4+ 36. Bxf4 Bxf4 37. Kxc3 g6 38. Be2 Kf8 39. b4 Ke7 40. a4 Kd6 41. Kc4 f5 42. Bd3 Be3 43. a5 Ke5 44. Bc2 Bf2 45. Bd3 Be3 46. Bc2 g5 47. d6 g4 48. d7 Bg5 49. b5 g3 50. b6 axb6 51. a6 g2 52. a7 g1=Q 53. a8=Q Qc5+ 54. Kb3 Qb5+ 55. Ka2 Qxd7 56. Qf3 Qa7+ 57. Kb3 Qf7+ 58. Kb2 Qe6 59. Qc3+ Kd6 60. Qd3+ Ke5 61. Qc3+ Kf4 62. Qd2+ Kg4 63. Qd1+ Kh4 64. Qh1+ Kg3 65. Qg1+ Kf4 66. Qc1+ Kg4 67. Qg1+ Kf4 68. Qc1+ {1/2-1/2 (68) Svidler,P (2754) -Naiditsch,A (2689) Moscow 2009}) 10... b6 {simply reinforcing the c5 pawn and allowing development of the bishop on b7.} (10... Bxb2 11. Nxb2 b6 12. d4 Nc3 13. Qd2 Nxe2+ 14. Qxe2 cxd4 15. Nxd4 Nxd4 16. exd4 Bb7 17. Qe3 Qd5 18. f3 Rfd8 19. Rfd1 Rac8 20. Rac1 Qa5 21. Nc4 Qb4 22. a3 Qe7 23. b4 Rc7 24. Ne5 Rdc8 25. Qd2 {Brozhik,V (2189)-Karnaukh,A (2253) Kiev 2003 0-1 (51)}) (10... Bxb2 11. Nxb2 b6 12. Na4 $11) 11. d4 (11. Nc3 {simply retreating the knight is probably easiest in response, although it's harder for humans to re-evaluate and simply admit a mistaken idea over the board.}) (11. Qb1 {is a related idea.}) 11... cxd4 12. Nxd4 Bb7 $15 {by this point Black's pieces are working well together, he has centralized knights and both bishops on the long diagonals. Mine are not as effective, with the misplaced Na4 and a potentially under-protected Bb2 needing to be watched carefully.} 13. Rc1 {a straightforward move, which Black could have responded to more effectively.} (13. Bf3 Ne5 {didn't appeal to me.}) 13... Rc8 {this was a relief, as I'm now able to reorganize my pieces and make exchanges.} (13... Bxd4 14. exd4 {saddles White with an isolated queen pawn, with Black already dominating the square in front of it.} Qg5 (14... Rc8 { is also good for an edge}) 15. g3 $15) 14. Nxc6 $11 Bxc6 15. Nc3 {finally getting the knight back into play.} (15. Ba6 $5) 15... Bxc3 16. Bxc3 Nxc3 17. Rxc3 Qg5 {I think my opponent was counting on this to sustain his advantage, but the threat is easily blocked. With the reduced material on the board and symmetric pawn structure, I have no problems.} 18. g3 Qf6 19. Qc1 {the obvious way to protect the rook, although Komodo is a bit more creative.} (19. Qa1 $5 { with a latent positional threat of exchanging on f6, also allowing the Rf1 to come to c1.}) 19... Bb7 20. Rxc8 Rxc8 21. Qd2 Qe5 {Black may still have hopes of taking advantage of the light-square weaknesses around my king, so I shut that down.} (21... Rd8 22. Qc2 $11) 22. f3 {now Black is blunted on the light squares.} Qc5 {immediately pressuring the weakened e3 square, but Black has no other way of increasing it.} 23. Kg2 {I played the king here rather than f2 in anticipation of Black pressure on the h-file with the queen.} g5 {the only way to try and make progress for Black is to advance the kingside pawns.} 24. g4 h6 25. Rf2 {my concern was Black playing ...Qc2 and after an exchange, getting his rook on the second rank.} Kf8 26. Qd7 {penetrating into my opponent's side of the board for the first time. There are no real threats, but there is an annoyance factor.} Qd5 {I'm happy to exchange queens. My opponent apparently thought that he could outplay me in the endgame, despite the equal position.} 27. Qxd5 Bxd5 28. e4 {further blunting Black's bishop.} Bb7 29. Bc4 {with my bishop finally in play and the situation on the board simplified, I felt confident that I would hold the draw.} Rd8 30. Rc2 {my strategy now is simply a preventative one, to prevent any Black breakthroughs.} Ke7 31. Kf2 Rc8 32. Rd2 a6 {preparing b5. This would certainly drive my bishop back, but its Black counterpart on b7 would also have nowhere to go.} 33. Ke3 {centralizing the king and protecting the Rd2.} e5 {although it doesn't change the engine evaluation, this move by my opponent sealed the draw for me, I felt. Black no longer has any hope of engineering a breakthrough on the kingside.} 34. Bd3 Rc5 35. Rc2 Kd6 36. Rxc5 {choosing to transition to a drawn bishop ending.} Kxc5 37. a3 {perhaps Black was hoping I would let his king penetrate on the queenside.} a5 38. Bc4 f6 39. Kd3 {keeping the Black king under watch and also protecting the Bc4, allowing the b-pawn to move if needed.} b5 40. Be6 Bc6 41. Ke3 {as long as the entry squares on the fourth rank are covered, Black cannot make progress.} b4 {hoping for a pawn exchange.} 42. a4 {now the board is completely sealed and my opponent offered a draw.} (42. axb4+ $2 Kxb4 $17) 1/2-1/2

On improving to the master level

An honest answer to the question:

If you happened to be at around 2000ish for a while, what do you think helped you improve past that to 2200? After being stuck for a while at 2000 I took a break for a year and I am contemplating whether to start playing in tournaments again.

Can be found here, from a newly minted master: 

https://www.reddit.com/r/chess/comments/cj444b/holy_shit_guys_im_officially_a_national_master_now/

The short answer: play frequent tournament games, and pay close attention to which structures and strategies work best for you. Also, don't rush when you have an advantage.

Training quote of the day #25: Artur Yusupov

GM Artur Yusupov, in Training for the Tournament Player:

What enables a chessplayer to be successful? In response to this question two essential factors are usually singled out: talent and hard work. But it is not sufficient just to be talented and hard-working. Physical condition, competitive character and the ability to concentrate during play are also very important. No less important is the ability to choose correctly the direction that such work should take and to be able to reach the required standard. Needless to say, this task is far from easy...

Of course, in order to be able to choose a direction leading to self-improvement it is necessary to have a critical understanding of one's game. The authors are totally convinced that the serious study of one's own games is an essential requirement for any chessplayer who wishes to improve. Therefore the theme 'analysing one's own games' occupies a central place. 

Video completed: "Why You Should Never Give Up in Chess" by Tatev Abrahamyan

"Why You Should Never Give Up in Chess" is the fourth video in the new Chess.com series by Tatev Abrahamyan. Like the others, it is around 15 minutes and presents its main theme using some narrated game examples. She states in the intro that the point is not to play on when mate is inevitably coming (which seems to be the fashion these days in OTB tournaments at the Class level) or if you are losing all your pieces. Instead, it's to take advantage of the fact that winning a supposedly won game is not automatic. She reminds the viewer of all the times that you may have thrown away a win after you thought it was a done deal, which I think is something relatable for many of us, and an excellent motivator on the flip side for playing on.

The first of the games is Frank Marshall - Georg Marco, in which the American chess legend manages to work some endgame wizardry versus his opponent, who had an advanced, unstoppable passed pawn. Marshall plays actively to complicate the position and obtains chances when his opponent loses focus on his winning idea, forcing through the passed pawn. Although the pawn does queen, Marshall now has a brilliant tactical resource available, coordinating his rook and knight to win it. While the position is probably drawn afterwards, the psychological blow is too much for Black and he goes on to lose. This illustrates how continuing to fight back, even if objectively losing, can still give you the opportunity for real chances and change your opponent's mindset from winning to losing.

The second game, IM Anita Gara - GM Irina Krush, highlights the tricky nature of rook endgames, which means you should not give up in them. The video picks up the action on move 103 (!) with White in a winning position, having an advanced (7th rank) rook pawn that is passed, with Black's king and rook on the other side of the board. This turns into a RvP endgame that Black manages to hold, once White makes a crucial error. Exhaustion is naturally a factor, as is Krush's knowledge of tricky endgame principles.

Abrahamyan mentions in passing GM Sam Shankland - GM Anish Giri, which is infamous for Shankland resigning in a drawn position, then moves on to the final game, GM Alexander Beliavsky - GM Larry Christiansen. White is up a pawn with an excellent position, so Black has to (in Abrahamyan's words) resort to desperate measures. As in the previous game, a stalemate motif is used by the losing side, except in this case Christiansen finds it from an unexpected middlegame position.

This is another good entry in the video series, both for the psychological and the technical ideas behind using every resource on the board you may have, in order to give yourself chances to save a losing game. The flip side of that coin is the idea is of giving your opponent the most chances to go wrong.

Annotated Game #216: Starting the breakthrough

This first-round tournament game, a win as Black against an Expert, started my breakthrough in performance from Class B to Class A (see "The Long Journey to Class A"). I had to get up on a Saturday morning and force myself to drive to the tournament, which afterwards I was glad I did. I'd been playing pretty regularly for the previous four months (one tournament / month) but had rather meh results. I was not looking forward to another mediocre tournament. However, diligence appeared to pay off and my game was elevated enough to produce better results over the board.

We reach an interesting and unbalanced position in the main line Slav (with the 5...Na6!? Lasker variation) by move 10, with my opponent deliberately inviting doubled f-pawns in exchange for potential play down the g-file. At 300+ rating points above me, I could tell he was clearly looking to create winning chances in an imbalanced position. The next several moves were critical and both of us missed chances to improve on the game score. A key idea was fighting against White's idea of f4-f5 to crack open my position, which at first didn't work. After my dubious 16th move then it did, as analysis shows, however neither I nor my opponent saw this.

By move 19 I've sufficiently protected the critical f5 square and the strategic nature of the game shifts, as White runs out of ideas and I take over the initiative. I exercise the simplest (and most effective) plan of building pressure down the d-file, which was largely risk-free, although there are some interesting possibilities in the variations. I was particularly pleased to see my two knights clearly better than my opponents' two bishops, which is a rarity in the Slav.

The winning blow comes as my opponent, under pressure, tries to cover his weak f-pawn, but fails to see a naked knight sacrifice that delivers check, picking up the exchange and a dominating position for Black. After that the win was just technical, although White held out until mate; as I mentioned in an earlier post, this seems to be much more the norm these days. It was slightly ridiculous, although there was perhaps a glimmer of hope on his part that I'd overlook a mate.

While it was not a clean game, I felt reasonably good about it afterwards, both in (finally) finding a way to stymie my opponent's pressure, and then in seeing the winning tactic (25...Nc3!!) - the double exclamation points being awarded by Komodo via the Fritz interface. It's nice to see the engine give positive feedback, now and again.

[Event "?"] [Site "?"] [Date "????.??.??"] [Round "?"] [White "Expert"] [Black "ChessAdmin"] [Result "0-1"] [ECO "D16"] [Annotator "ChessAdmin/Komodo 11.2"] [PlyCount "82"] {[%mdl 8256] D16: Slav Defence: 5 a4: Lines with 5...Bg4 and 5...Na6} 1. d4 d5 2. c4 c6 3. Nf3 Nf6 4. Nc3 dxc4 5. a4 Na6 {the Lasker Variation of the main line Slav. Rare to see in tournament practice, but easy to learn.} 6. e4 { the most aggressive response.} Bg4 {along with the queen knight's placement on a6, the bishop being played to g4 rather than f5 is the major alternative feature to the usual Black setup in the Slav. Here it's of necessity, given the pawn on e4.} 7. Bxc4 e6 8. Bg5 {there are only two games in my database with this move. The obvious threat is e4-e5 now that the Nf6 is now pinned.} Be7 $146 {the natural move, to unpin the knight, but Komodo prefers a more active approach.} (8... Bxf3 9. Qxf3 (9. gxf3 Nb4 $5 10. e5 {and now} h6 { threatens the bishop and allows Black to break the pin one way or another.}) 9... Qxd4 $11) (8... Qa5 $5 {was played in the only master-level game.} 9. e5 Ne4 10. Bxa6 bxa6 11. Qd3 Nxg5 12. Nxg5 Bf5 13. Qf3 Rc8 14. Nxf7 Kxf7 15. g4 g6 16. gxf5 gxf5 17. Rg1 Ke7 18. Kf1 Qb4 19. Rg3 Kd7 20. d5 Kc7 21. dxe6 Qc4+ 22. Kg1 Qxe6 23. Ne2 Qxe5 24. Rd1 Qe4 25. Nd4 Qxf3 26. Rxf3 Rg8+ 27. Kf1 Rg6 28. Nxf5 Rb8 29. b3 Rb4 30. Ne3 Bd6 31. Rf7+ Kb8 32. Nc4 Bc7 33. Rdd7 {1-0 (33) Rodewis,T (2268)-Mudelsee,M (2295) Germany 2002}) (8... Nb4 {is also possible, a standard theme in this variation, since per above e5 can be met by ...h6.}) 9. Qd2 {reinforcing the bishop's position and clearing the way for a possible queenside castling. This is also a clear invitation to take on f3, which I decided to do after some thought, as there is nothing better.} Bxf3 10. gxf3 { now White has the pair of bishops and a half-open g-file to use, in compensation for his ruined kingside pawn structure. My opponent was evidently interested in taking an aggressive posture.} Nb4 {getting the knight on the rim into play. The whole point of its development to a6 is for it to hop into b4 when a good opportunity presents itself.} 11. Be3 {removing the bishop from a square where it is underprotected, as well as shoring up d4.} a5 {the engines tend not to like this move in the Lasker Variation, although it seems logical to me to reinforce the b4 outpost in a more permanent way. One drawback is that ...Qa5 is no longer possible. Another is that supporting ... b5 with the a-pawn is also ruled out.} (11... O-O $5 {Komodo has no fear of castling into an attack. I did not want to do it at this point, though, believing that it would give White a too obvious and easy plan to follow.} 12. O-O-O Kh8 13. Rhg1 a6 {with the idea of ...b5 to follow is an alternate way to play.}) 12. Rg1 {the obvious place for the rook.} g6 {after some thought, I decided that the dark-square holes created would be offset by the peace of mind of not having to constantly worry about defending down the g-file.} (12... Nh5 {I also considered playing immediately here, since the knight can't be effectively challenged by White and it helps cover f4.}) 13. f4 $6 $11 { this is not helpful for White's attacking potential, since it blocks the c1-h6 diagonal and renders my dark-square weakness less accessible to my opponent.} Nh5 {I thought for a while here and was focused primarily on blunting White's pressure. Now if f4-f5, for example, I was considering ...Ng7 in response. However, the knight is generally better on f6 and its placement on h5 causes potential problems later.} (13... Qc7 {is a more useful developing move, clearing d8 for the rook and getting on a more productive diagonal.}) (13... O-O $5 {makes more sense now, even if White has some temporary pressure. The f-pawn advance does not prove truly threatening.} 14. f5 $6 exf5 15. exf5 Nfd5 {blocking the a2-g8 diagonal. Black will now be happy with any of the exchanges White could make.}) 14. O-O-O {not a surprise, since there's no better place for his king to go, and it brings the other rook into play.} (14. f5 $6 {is in fact premature.} exf5 (14... Ng7 $2 {at this point is bad for Black after} 15. fxe6 fxe6 $16 {since White's pawn structure is significantly better and my knight is in an awkward spot. White possesses a significant space advantage, while Black is cramped and the pieces are not cooperating well.}) 15. exf5 Ng7 {now the knight move is good again.} 16. fxg6 hxg6 (16... fxg6 $2 {leaves Black's king position too airy.}) 17. O-O-O Kf8 $11) 14... Qc7 {getting my queen away from the d-file pressure and preparing queenside castling.} 15. Kb1 {an normal precautionary move. At the time, I felt that it helped ease the pressure on me.} (15. f5 $5 {I was still worried about.} exf5 { the only good move} 16. exf5 Bf6 $14) 15... O-O-O {at this point I felt like the immediate danger had passed, and I could now start looking for counterplay. } 16. Qe2 (16. Be2 Nf6 $11) 16... Bd6 $6 {an example of where move sequence matters. I should have gotten the king to a less exposed square first.} (16... Kb8 $5 $11 {would be the simplest way to equalize.}) (16... Nxf4 {snatches a pawn but gives White good play for it.} 17. Qf3 g5 18. e5 f6 $11) 17. e5 $16 { I had expected this after calculating the previous move. My thinking was that the pawn advance should result in closing off White's dynamic prospects with his central pawns.} Be7 (17... Bf8 {I had also considered, but in the end I thought that being able to move along the d8-h4 diagonal was more valuable. The problem with the text move, which I did recognize at the time, was that the bishop on e7 screens the f7 pawn from the Qc7's protection. This fact could have been used much more effectively by my opponent.} 18. f5 Ng7 19. fxe6 fxe6 $16) 18. Rc1 $6 {missing the chance to break through.} (18. f5 $1 { superficially this looks like it just loses a pawn, which is probably why neither my opponent nor I seriously considered it.} exf5 {forced if I want to capture, otherwise the Nh5 hangs. I had not bothered to look at this line, just assuming that gxf5 would be possible. Now the f7 pawn is hanging and White is looking dangerous in the center, especially with what will be a protected passed pawn on e5.} 19. f4 ({or} 19. Bxf7 $16) 19... f6 20. Qf2 fxe5 21. dxe5 c5 $16) 18... Kb8 {this would be a good and standard sidestep for the king, except that the f5 idea is still possible.} (18... Ng7 {is now in fact best, as it adds a defender to f5.}) 19. Rgd1 (19. f5 $5 {again would give White an advantage.}) 19... Ng7 $11 {now I am truly out of trouble. Strategically, White has run out of ideas and the center is locked, so Black has a more comfortable game going forward. The backwards d-pawn, for example, is now an obvious weakness.} 20. f3 (20. Na2 $5 {would seek to exchange knights, a net benefit to White due to the relatively better activity of the Nb4.}) 20... Rd7 $15 {with the idea of doubling rooks on the d-file. My opponent now struggles to come up with useful moves.} 21. Ne4 {a better place than c3 for the knight.} Nf5 {targeting the backward pawn on d4, from a nice outpost on f5.} (21... Rhd8 $5 {I considered, but it would have allowed White to remove the d4 pawn weakness, which I preferred to keep on the board.} 22. Nc5 Bxc5 23. dxc5 Nd5 $15) 22. Bf2 $6 (22. Nc5 $5 {would disrupt my plans along the d-file, one way or another, although I still have an edge.}) 22... Rhd8 $17 23. Nc5 {this comes too late now, although I miss the best tactical continuation.} Bxc5 (23... Rxd4 $1 {successfully snatches the pawn, something I didn't consider at the board.} 24. Ne4 (24. Bxd4 $2 Nxd4 {and now the Nc5 is hanging, which is the main idea of the tactical sequence. White can give back the exchange or do a desperado move with the knight, but Black has a major advantage both ways.} 25. Rxd4 (25. Nxe6 Nxe2 $19) 25... Rxd4 26. Ne4 Nd5 $19) 24... R4d7 $17) 24. dxc5 Nd5 {threatening the f4 pawn. Strategically, it's also interesting to see the two Black knights versus the two White bishops. The knights can be exchanged, but not challenged by pawns, which makes life more difficult for White.} (24... Rxd1 {Komodo prefers exchanging off a pair of rooks first, to reduce White's ability to fight down the d-file. However, the subsequent winning tactic in the game would then not have been possible.} 25. Rxd1 Nd5 $17) 25. Qe4 $2 {protecting the pawn, but now the Rd1 is underprotected. I realized this and was able to take advantage of it immediately.} (25. Qe1 $11 {is the engine line, which for a human would be unlikely, seemingly just abandoning the f-pawn.} Nxf4 26. Rd6 {this is the idea, an unnatural-looking exchange sacrifice.} Nxd6 $2 {now leads to mate.} ( 26... Nd5 $11) 27. cxd6 Qc8 28. Bb6) (25. Bg3 $17 {is what I had expected.}) 25... Nc3+ $3 $19 {Komodo (via the Fritz interface) actually awarded this the two exclams, so I'll leave it in there. It's unusual to see a naked piece sacrifice like this, meaning that no initial capture is involved. This makes it harder to see as a candidate move during calculation.} 26. bxc3 Rxd1 $19 { now I'm an exchange up with no compensation for White. The open d-file means that my rook is quite valuable, as well.} 27. Rxd1 Rxd1+ 28. Kc2 Qd8 29. Bd4 { I saw this idea during calculation, but the rook retains mobility on the first rank and White's weak kingside pawns come back to haunt him.} Rh1 30. Qd3 Nxd4+ {I chose to reduce material and eliminate the two bishops factor, which I assessed would be better for my rook.} 31. cxd4 Rxh2+ 32. Kb3 Qh4 {now the queen can take advantage of the first and second rank weaknesses as well.} 33. Ba6 {hoping in desperation that I'll take the bishop and White can steal a perpetual check.} Qf2 (33... Qe1 {is the quicker win.} 34. Qe2 Rxe2 35. Bxe2 Qxe2 36. Kc3 h5 37. d5 cxd5 38. f5 gxf5 39. f4 Qa2 40. Kd3 d4 41. Kxd4 Qb3 42. c6 h4 43. cxb7 Kxb7 44. Kc5 Qb4#) (33... bxa6 34. Qxa6 Qd8 $19 {is still winning for Black, but would allow White to be more annoying with his queen.}) 34. Qb1 {hoping I'll miss a mate threat on b7.} Qxf3+ {at this point I was in "safe win" mode and did not feel like burning a lot of brain cells to try and calculate all the way to mate.} (34... Qxd4 35. Bc4 Qd2 36. Be2 Qb4+ 37. Kc2 Rxe2+ 38. Kd3 Rd2+ 39. Ke3 Qd4#) 35. Kc4 Qe2+ {picking up the bishop and squelching any chance of White getting a swindle.} 36. Kc3 Qxa6 37. Qb6 Rh3+ { at this point I saw the inevitable mate, so didn't exchange into the (totally won) endgame.} 38. Kd2 Qd3+ 39. Ke1 Re3+ 40. Kf2 Re2+ 41. Kf1 Qd1# 0-1