29 August 2019

Avoiding and overcoming negative game trends

One thing that happens when you keep analyzing your own games is that you start seeing meta-patterns across games. One of the earlier examples for me was the realization that I did not have a real thinking process in place, and as a result was doing especially poorly at discovering the best candidate moves for both myself and my opponent.

During the last set of annotated games (#216-218), the idea that stood out to me - not for the first time - is the idea of negative game trends and how to overcome them. It has been an all-too-common pattern to see: making a series of sub-optimal moves in the opening and early middlegame, based on an incorrect evaluation or shallow understanding of position's requirements. Overconfidence often can play a part, by encouraging you to not "switch on" and apply a full thinking process until too late in the game, typically resulting in assumptions about your position's safety that are quickly proven wrong.

It's interesting to see how this phenomenon works in practice. Analysis shows how an initial less-than-optimal decision may only result in a slightly less positive evaluation of a position. However, from there it becomes a slippery slope, since it is hard strategically to recover from implementing the 'wrong' idea on the board of what the position truly demands from you. Mathematically, this manifests itself in a reduction in the range of different moves that will keep you in the game (equal or better). This means that the necessary good moves inevitably become harder and sometimes nearly impossible to find for a human. As with other pitfalls of computer analysis, it's misleading to think that an unchanged top-line engine evaluation means that you are in fact playing optimally.

I think one of the practical elements of chess strength is an ability to persevere long enough to break a game's negative trend, including finding the optimal moment for starting counterplay. GM Alex Yermolinsky's The Road to Chess Improvement has a whole section on trend-breaking tools, which shows that this is a challenge at all levels. Personally, in analysis it's been evident time and again how I've missed chances to strike back and equalize, or even win, after being under pressure for a number of moves. One of the elements of my breakthrough in this particular tournament was the ability to find that chance and seize it, most cleanly in Annotated Game #216. As with anything else in chess, however, it's a work in progress.

21 August 2019

Annotated Game #218: The saving backwards move

This game falls into the category of a cringeworthy win. In the last round of the tournament, I faced someone of comparable strength, after having performed well in the previous three rounds. The course of the game I believe is explained more by psychological and "visual" factors than anything else. The one concrete takeaway from the analysis was the reason behind my error on move 11, which is a valuable teaching point in this variation of the Classical Caro-Kann.

In any event, after this White quickly assumes a dominant-looking position on the kingside, which I compound by an unwillingness to violate standard principles and give up the right to castle, which would have yielded an acceptable game. White makes an "obvious move" (the perils of which I've posted about before) on move 16 and I should have been able to effectively respond and turn the tables on him. However, I find only the second-best move and then fail to identify the right strategy, which would be to sacrifice the f- and g-pawns in return for real counterplay for my rooks and against White's airy king position. Instead, I unnecessarily sacrifice a bishop in return for some temporary, if awkward-looking, threats. My opponent, instead of pressing his advantage immediately, starts focusing on my psuedo-threats, however. This mistake leads to the opportunity for a backwards bishop move that forks king and queen and immediately wins. These types of backwards moves have a higher chance of being overlooked in calculation, since they appear to be considered less natural to the brain's board vision, unless you enforce a disciplined thinking process about considering candidate moves.

Although it wasn't really skill that decided this game, I will give some credit at least to the positive value of not giving up and the idea of seeking to pose problems for your opponent, who is thereby given the chance to go astray.

[Event "?"] [Site "?"] [Date "????.??.??"] [Round "?"] [White "Class B"] [Black "ChessAdmin"] [Result "0-1"] [ECO "B19"] [Annotator "ChessAdmin/Komodo 11.2"] [PlyCount "52"] {[%mdl 8192] B19: Classical Caro-Kann: 4...Bf5 main line} 1. e4 c6 2. d4 d5 3. Nd2 dxe4 4. Nxe4 Bf5 5. Ng3 Bg6 6. h4 h6 7. Nf3 Nf6 8. h5 Bh7 9. Bc4 e6 { as well as providing an outlet for the bishop, this also anticipates Ne5 and a threat to f7.} 10. Qe2 Be7 11. Ne5 $146 {although a novelty, it's the logical continuation for White after Bc4 and Qe2. Typically there are sacrificial ideas on f7/e6.} Nd5 {in an analagous position in the 8. Ne5 line, before Black has played ...Be7, this is the main defensive move, and is necessary to block a sacrifice on e6. However, here it's not needed, as the e-file is already blocked and there are no tactics for White.} (11... O-O $11 {Komodo prefers the plan of immediate castling, followed by ...Nbd7, to exchange off the Ne5.}) 12. Qg4 {this is the problem with the knight move, although Black is still OK.} Bf6 $6 {I thought this was a logical move to protect g7 and also get the bishop to a slightly better square.} (12... Rg8 $5 {giving up the right to castle kingside seems unnatural, but Black defends easily here. We will see later how the weak g-pawn becomes a real problem.} 13. Qf3 Rf8 { and it is slightly awkward for Black, but there are no weak points.}) 13. O-O ( 13. Bd3 {is a better try for an advantage, as the bishop is not doing anything on c4.} Bxd3 14. Nxd3 $14) 13... Nd7 {developing and challenging e5.} (13... Bxc2 {is possible here, but I was wary of pawn snatching in the opening while behind in development.}) 14. Re1 Nxe5 (14... Bxe5 $5 15. dxe5 g5 $14 {this solves the g-pawn problem by force, although leaving some holes behind.}) 15. dxe5 $14 Bg5 {now the bishop is boxed in and the g-pawn is a critical weakness. } 16. f4 $2 {this was the obvious next move to both my opponent and myself, forcing the bishop back. It visually looks solid, but in fact it leaves behind a tactical weakness.} (16. Bxd5 {preserves an advantage for White.} cxd5 17. Bxg5 hxg5 (17... Qxg5 18. Qa4+ $16) 18. c4 dxc4 19. Rad1 Qe7 20. Qxc4 O-O $14) 16... Qa5 {this is a good response, although I missed the following tactic.} ( 16... Nxf4 $1 17. Kh2 (17. Bxf4 $2 Qd4+ {picking up the now under-defended bishop.}) 17... Nd5 $17) 17. Kf1 O-O-O $2 {this was done out of desperation for counterplay, as I thought my position was collapsing anyway. The bishop sacrifice is unnecessary, however.} (17... Be7 {Komodo cold-bloodedly accepts the loss of the g-pawn.} 18. Qxg7 O-O-O 19. Qxf7 {White is now two pawns up, but at the same time his king is far more exposed, with the f- and g-files now open for Black's rooks to operate and the White queen having limited squares.} Kd7 $17) 18. fxg5 $18 Nb4 {this was my idea, which at least poses an awkward threat of Nxc2. White could easily just ignore it, however.} 19. Bb3 {this doesn't in fact help White, which gave me some psychological momentum, in that his own string of threats had been broken.} (19. g6 {seems to be the most forcing continuation.} Nxc2 20. Bf4 $18 {now that White has protected the Rf1 sufficiently, and with the Bh7 hanging, Black has no good choices.}) 19... Bxc2 20. Qc4 $4 {missing the backwards bishop move. White evidently wanted to protect the Bb3, given that the a-pawn is currently pinned against the rook.} ( 20. Bf4 {again would have protected the rooks sufficiently.}) 20... Bd3+ $19 { after this the win is technical.} 21. Re2 Bxc4 22. Bxc4 Rd1+ 23. Kf2 Nd3+ 24. Kf3 Nxc1 25. Rc2 Qxe5 26. Ne2 Qf5+ {continuing to gobble up material, although there was a short mate available. In any case, my opponent resigned.} (26... Rf1+ 27. Kg4 Qxg5+ 28. Kh3 Rh1#) 0-1

04 August 2019

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