Practical endgames

A long game doesn't always mean an interesting game. My interclub-game against Tom Piceu took 53 moves but was in fact decided already after 27 moves. It is surely for some part nonsense to state that the number of moves is linked with activity as I insinuated in my previous article. On the other hand I did encounter last season some complex endgames on the board which I believe are interesting to show the reader. 

Of course I start with the interrupted queen-endgame against Bart. Initially I find the right moves but with the draw at reach, it still goes wrong.

[Event "Interclub Deurne - KGSRL"] [Date "2014"] [White "Brabo"] [Black "Michiels, B."] [Result "0-1"] [ECO "C11"] [WhiteElo "2336"] [BlackElo "2510"] [SetUp "1"] [FEN "1Q6/p5p1/2p2k1p/2Pq1p2/8/8/PPP3PK/8 w - - 0 42"] [PlyCount "26"] 42. Qxa7 Qe5 43. Kg1 Qxb2 $5 {(Somewhat more accurate is first giving the intermediary check at d4 before capturing the pawn.)} 44. Qb6 $5 {(I used an enormous amount of time for this move as I had serious difficulties to concentrate.)} (44. Qa4 $5 {(A computer-move obviously which profits from blacks last inaccuracy.)} Qa1 (44... Qc1 45. Kh2 $11) 45. Kf2 $11 {(In both lines white wins the c-pawn after which the free pawn easily forces the draw. Black could avoid this with Qb7 but then he loses the coordination.)}) 44... Qc1 45. Kh2 Qf4 46. Kh1 Qh4 47. Kg1 Qe4 48. c3 $5 {(Criticized after the game by Bart but it all stays within the boundaries of a draw.)} (48. Qd8 $5 Kg6 49. Qd6 Kg5 50. Qd3 $5 {(Unnecessary but I do find the pawn-endgame sufficiently interesting to investigate.)} Qxd3 51. cxd3 Kf6 52. a4 Ke6 53. d4 Kd7 54. a5 h5 55. a6 Kc7 56. Kf2 h4 57. Ke3 g5 58. d5 cxd5 59. Kd4 g4 60. a7 Kb7 61. Kxd5 Kxa7 62. Kd6 g3 (62... h3 63. gxh3 $11) 63. c6 h3 64. gxh3 $11) 48... Ke6 49. c4 $5 {(Initially I wanted to play Qb4 but as it costs a pawn, I chose for an alternative.)} (49. Qb4 $5 {(After the game I indicated that probably Qb4 is also playable.)} Kd5 50. a4 Qe3 51. Kh2 Qxc5 52. Qb3 Qc4 53. Qa3 $11 {(The a-pawn is a sufficient counter-weight.)}) 49... g5 50. a4 g4 51. Kh2 $4 {(I had seen Qd8 but suddenly I detected Kh2 which also wards off g3. Without any further reflection I played Kh2 as my remaining time was shrinking but the refutation popped up quickly as a cold shower.)} (51. Qd8 Qe3 (51... g3 {(Now this move would even lose.)} 52. Qe8 Kf6 53. Qxe4 fxe4 54. a5 $18 ) 52. Kh1 Qe1 53. Kh2 Qe5 54. Kh1 Qxc5 55. a5 $11 {(The perpetual check and the a-pawn are sufficient for the draw.)}) (51. a5 $4 {(To ignore the thread is not an option.)} g3 52. Kf1 Qxc4 53. Ke1 Qc1 54. Ke2 Qc2 55. Ke3 Qe4 56. Kd2 Qxg2 57. Kc3 Qf3 $19) 51... f4 52. Qd8 {(The only way to avoid a direct check-mate. I already figured out that it loses but I still continue to verify if Bart also has found the win.)} (52. a5 g3 53. Kh3 (53. Kg1 Qe1#) 53... Qf5 54. Kh4 Qg5 55. Kh3 Qh5# {(I missed this check-mate over the h-file when playing my 51st move. Even with strong reduced material tactics remain important.)}) 52... g3 53. Kh3 Qf5 54. Kh4 Qg5 {(Our FM Pieter Tolk asked me why I already resigned but I considered the resulting pawn-endgame too easily won to continue.)} (54... Qg5 55. Qxg5 hxg5 56. Kg4 {(White can not capture the pawns due to f3 while blacks king can quietly via a small detour eat my pawns on the queenside.)} Kd7 57. Kf3 Kc8 58. Kg4 Kb7 59. Kf3 Ka6 60. Kg4 Ka5 61. Kf3 Kxa4 $19) 0-1

With this loss I had exactly a break-even for my fide-rating and with the relegation to 2nd division which was already confirmed for some time (see the wild west) you could argue that it didn't matter much. However honestly I have to admit that this loss came as a big blow. Indeed Bart did present in a clever way some problems but without my amateurism it should never given any success.

It was not the only queen-endgame which I encoutered on the board. In the Open of TSM I played 4 games of which 2 for rating. One game for rating I lost in a dramatic way against Steven which I earlier discussed on my blog, see the sadistic exam but also the second game didn't go smoothly. I wasn't able to cash an extra pawn against Jan Gooris.

[Event "TSM Tornooi"] [Date "2013"] [White "Gooris, J."] [Black "Brabo"] [Result "1/2-1/2"] [ECO "C56"] [WhiteElo "2130"] [BlackElo "2347"] [SetUp "1"] [FEN "8/Q4pk1/6p1/2p4p/P1q5/5P2/6KP/8 w - - 0 43"] [PlyCount "73"] 43. a5 {(Just before move 40 white blundered a pawn. White still has drawing chances but can not afford any further mistake.)} Qe2 44. Kg1 Qe1 45. Kg2 Qd2 46. Kg3 c4 $5 {(The program Stockfish, a specialist in the endgame recommends h4 but can neither find a decisive advantage.)} (46... h4 $5 47. Kh3 Qd5 48. Kxh4 Qe5 49. a6 Qf4 50. Kh3 Qxf3 51. Kh4 c4 52. Qc5 c3 53. a7 c2 54. a8=Q Qxa8 55. Qxc2 {(The finalgen-tool tell us this is a draw.)}) 47. Qc5 h4 $6 {(Immediately going for promotions with c3 is likely a bit more flexible and favorable but even in that scenario I can not find any win.)} 48. Kh3 c3 $5 {(A lovely alternative is Qe2 without changing the verdict of the position.)} ( 48... Qe2 $5 49. a6 Qxf3 50. Kxh4 Qf6 51. Kg3 Qxa6 52. Qd4 Qf6 53. Qxc4 { (Again the finalgen-tool tells us that is a draw.)}) 49. a6 c2 50. a7 c1=Q 51. Qxc1 Qxc1 52. a8=Q Qf4 53. Qa1 Kh7 54. Qc3 g5 55. Qc6 Qf5 56. Kg2 g4 $6 {(Hereby we liquidate to a theoretical drawn endgame. With Qe6 black can further prolong but progress I can not find against a correct defense.)} 57. fxg4 Qxg4 58. Kf2 Qf5 59. Ke3 Qh3 60. Qf3 Qe6 61. Kf4 Qg6 {(The best practical chance was surely the pawn-endgame.)} (61... Qf6 62. Kg4 Qxf3 63. Kxf3 f5 {(This is a draw but still is a bit tricky with little time left. E.g. h3 here loses.)}) 62. Qg4 Qd6 63. Kg5 Qh6 64. Kf5 Qe6 65. Kg5 Qe7 66. Kf5 Qc5 67. Kf6 Qf2 68. Qf5 {(I believe Jan here proposed a draw but I was so disappointed that I continued.)} Qxf5 69. Kxf5 Kh6 70. h3 Kh5 71. Kf6 Kh6 72. Kf5 Kg7 73. Kg5 f6 74. Kxh4 Kg6 75. Kg4 f5 76. Kf4 Kf6 77. h4 Kg6 78. h5 Kxh5 79. Kxf5 {(In the post-mortem at 1.30 in the night I kept searching for a missed win but now I just have to admit that it probably never existed.)} 1/2-1/2

I was unhappy after the game with the result but if I look at the endgame now then I just have to admit that a win was never there. Maybe with a slower tempo (after move 40 I only got 15 minutes extra so we played mainly with 30 seconds increment) I would've been able to create more problems but even then it is very unclear if this would force a mistake from Jan. Anyway I found it a pity that the competition already ended in december so in the following months I only played a few games interclub. The ghost of inactivity is difficult to repel.

This season was surely not all misery in the endgame. In my articles a moral victory and universal systems I already showed how I was able to hold slightly inferior positions. However those examples are light beer compared with the endgame below which I was able to draw in an incredible way against the French FM Ludovic Carmeille.

[Event "Interclub Deurne - Anderlecht"] [Date "2013"] [White "Brabo"] [Black "Carmeille, L."] [Result "1/2-1/2"] [ECO "B92"] [WhiteElo "2347"] [BlackElo "2320"] [SetUp "1"] [FEN "8/6k1/4pnp1/7p/5P2/3Bp1P1/7K/8 w - - 0 40"] [PlyCount "34"] 40. Kg2 {(Black has 2 extra pawns but the strange structure gives white still drawing-chances.)} Ng4 $6 {(Nd5 is announced as winning by the engines but I can not find a winning sequence although the defense for white is surely also not easy.)} (40... Nd5 $1 41. Kf3 Kf6 42. Bc4 g5 43. fxg5 Kxg5 $17 {(Black has 2 extra pawns but if the h and g-pawns disappear then a tablebase draw remains. I do not see how black can avoid the exchange if progress must be made.)}) 41. Be2 $2 {(The concentration fades after accomplishing move 40 or just a simple miscalculation.)} (41. Kf3 e5 $5 {(I feared this move.)} 42. Be2 $1 (42. fxe5 Nxe5 43. Kxe3 Nxd3 44. Kxd3 $19 {(I evaluated this lost pawn-endgame correctly.)}) 42... Kf6 43. Ke4 {(I missed this idea in the game.)} exf4 44. Kxf4 $11 { (Black loses the e-pawn after which the finalgen-tool tells us this is a draw.)}) 41... Nh6 42. Kf3 Nf5 43. Bd3 Kf6 {(E2 is winning which I already discovered during the game. Also with Kf6 black wins but it becomes much more difficult.)} 44. g4 e2 $2 {(However now e2 does not function anymore due to my next move. Hxg4 was still sufficient for a win as the black king can force the decision via e7-d6. It is remarkable that Stockfish instantly evaluates the position correctly while Houdini needs time.)} (44... hxg4 45. Kxg4 Ke7 46. Kf3 Kd6 47. Bxf5 gxf5 {(Exf5 wins too.)} 48. Kxe3 Kc5 $1 {(The only move which wins thanks to the opposition.)}) 45. g5 Ke7 46. Kxe2 Kd6 47. Kf3 Kd5 48. Bb1 Kd4 49. Ba2 Kd3 50. Bb1 Kc3 51. Ba2 Kd2 $6 {(With Kc2 black could still further prolong the game as now I can force the draw.)} 52. Kf2 Kd3 53. Bxe6 Ke4 54. Bf7 Kxf4 55. Bxg6 Kxg5 56. Bxh5 Kxh5 1/2-1/2

Such half points taste very sweet. An other spectacular escape happened in my game against the Bulgarian grandmaster Dejan Bojkov (extracts of the game were already discussed in camouflage and einstellung effect). The win was surely not trivial for black.

[Event "Interclub Deurne - Amay"] [Date "2014"] [White "Brabo"] [Black "Bojkov, D."] [Result "1/2-1/2"] [ECO "C99"] [WhiteElo "2336"] [BlackElo "2500"] [SetUp "1"] [FEN "RR2n1rk/7p/6p1/4pP2/1P4P1/4Bq1P/5P2/6K1 b - - 0 41"] [PlyCount "112"] 41... Qxh3 $5 {(I was lucky not having to resign immediately as after Nf6 I still have Bg5. An interesting alternative is gxf5 although I neither could find a clear win for black.)} (41... gxf5 $5 42. Rxe8 Qxa8 43. Rxa8 Rxa8 44. gxf5 Kg7 $17 {(Black can still try for a long time but the b-pawn seems just sufficient to keep the balance.)}) 42. Rxe8 Qxg4 43. Kf1 $6 {(I spent a lot of time on this move and still I make the wrong decision as I am not able to fully scrutinize the position.)} (43. Kh2 $1 Rxe8 $5 44. Rxe8 Kg7 45. f6 {(I missed this beautiful saving intermediary move in the game.)} Kxf6 $5 46. b5 $17 { (Again my analyses show that whites b-pawn is sufficient to save ones bacon.)}) 43... Rxe8 $6 {(I mainly feared Qd1 and elaborated analysis indeed justified my intuition. Black nonetheless a grandmaster chose after a long thought for the wrong game-continuation as he considered the b-pawn too strong.)} (43... Qd1 $1 44. Kg2 Qd5 $1 45. Kg3 Qxa8 46. Rxa8 Rxa8 47. fxg6 hxg6 $17 {(However this time the b-pawn is insufficient although I immediately should add that the win is certainly not trivial. The win consists to sacrifice the rook at the right moment for the bishop so a won pawn-endgame remains. This is possible at contrary with the analyses of move 41 due to the missing white h-pawn.)}) 44. Rxe8 Kg7 45. fxg6 Qd1 46. Kg2 Qd5 47. Kh2 e4 48. Re7 Kxg6 49. Ra7 $5 {(I focus on achieving a fortress but abandoning a pawn is rather debatable. In any case white now has to prove that a fortress is existing.)} Qd6 50. Kg2 Qxb4 51. Ra6 Kf5 52. Rh6 Qe7 53. Kh3 Qf7 54. Kh4 Qe7 55. Kh3 Qd7 56. Kg3 Qd1 57. Kh2 Qd7 58. Kg3 Qg7 59. Kh2 Kg4 60. Kg2 Qd7 61. Rf6 Qc7 62. Rh6 Qg7 63. Rh3 Kf5 64. Kh2 Qg8 65. Rh6 Ke5 66. Rh5 Kd6 67. Bf4 Kc6 68. Be3 Qf7 69. Rc5 Kd6 70. Rg5 Qf3 71. Ra5 h5 72. Rg5 h4 73. Ra5 Ke6 74. Rg5 Kf7 75. Rg1 Kf6 76. Rg8 Ke6 77. Rg5 Kf7 78. Rg1 Qf5 {(White has found an optimal setup and there seems to exist no break-through. Black keeps trying and hopes mainly that I will still commit a mistake in the final-phase due to time-trouble.)} 79. Rg5 Qe6 80. Rg1 Kf6 81. Rg5 Qd7 82. Rg8 Qd6 83. Kg2 Qe6 84. Rf8 Ke5 85. Kh2 Kd5 86. Rf4 Kc4 87. Rxh4 {(Now it is evident for everybody that this can not be won anymore without a big blunder. However black still refuses my draw-offer and wishes to squeeze further.)} Kd3 88. Kg2 Ke2 89. Rh3 Qg4 90. Rg3 Qf5 91. Rh3 Qd5 92. Rg3 ( 92. Rh1 $4 {(It is never too late to commit errors.)} Qg8 93. Kh2 Kf3 {(and check-mate can only be postponed for 1 move.)}) 92... Qd1 93. Rh3 Qf1 94. Kh2 Qa1 95. Kg2 Qf6 96. Rg3 Qa1 97. Rg5 {(Only now black had enough and proposed a draw despite having the move. Likely the fact that black had only a minute left on the clock also played a role.)} 1/2-1/2

I was a little afraid that he would like to win solely on time which is one of the biggest disadvantages of playing without an increment but fortunately it remained sportsmanlike. Anyway I won't complain about my endgame-results.

Can we train ourselves in these type of endgames? Well if we review the different positions then we remark that they are all unique. It are practical endgames which you can't find in any books. It is even doubful if it is interesting to study them as the chance is minimal that something can be reused in another endgame. I like delving in the complexities of endgames so I spend a lot of time on it but I won't recommend the work to players which hate these endgames.

On the other hand a certain basic knowledge of endgames sounds to me not redundant. However also that is not anymore a certainty if you hear Nakamura telling us how he didn't know the Vancura position so he had to distillate the right moves on the board himself. A completely different sound gives the recent work Grandmaster preparation endgame play of Jacob Aagaard in which the study of the endgame has been raised to a new (higher) level. He expects from the ambitious student that the tasks are solved in a peaceful environment. That way skills are acquired which are useful for tournament-chess.

On the book clearly an enormous effort was inserted. Nevertheless quickly a lot of critics were given of which the author clearly wasn't pleased, see his reaction on Quality chess. The tasks are too difficult was one of the most heard critics. Even the German grandmaster Joerg Hickl complained that he could only solve 10%. The fact that in many examples the correct move was missed even by players like Ivanchuk, indeed proofs that it is difficult. On the other hand finding the right move in a problem is easier than in practice as mentioned by Glen in a reaction on my article 'the expert'.

We can only speak of serious training  if done in a peaceful environment in which it is possible to concentrate properly on solving and studying endgames. This is how the author refutes the critics. However today we see that a lot of practical endgames must be dashed. Surely with the ever more becoming popular 30 seconds increments we can mainly trust our instincts and some minimal calculations.

With a slower tempo (like in the Belgian interclub) I still believe such trainings have their usefulness. It is not that we can copy some combinations or schemes in practice but it does stimulate our thought process how to solve certain problems. I also believe the effect of the trainings is temporarily and must be often repeated. You could maybe compare it with making IQ-tests. You won't become smarter with solving IQ-tests but you can improve your score on such test with making similar tests in advance. It was also detected that the effect quickly diminishes when no further tests were done. Studying practical endgames is not something trivial. Each amateur has to decide for himself if the effort is valuable or not.

Brabo