Sunday, April 21, 2019

Exchange pawns when standing worse

In my previous article I introduced the book Applying logic in chess and wrote that the content is often rather abstract. This means it is not always straight forward for the reader to figure out how to materialize the training guidelines into concrete activities. Still this doesn't mean you can't find any simple advice in the book. Personally I was surprised that the author advised several times in the book how important it is to exchange pawns when you have an inferior position. He considers it a basic-rule to improve the defense.

Well I have to admit that I never heard about such rule before. I do know that you have to exchange material when being ahead and you have to avoid exchanges when being down in material like I demonstrated successfully in the great escape. However I never heard about making a distinction between pawns and pieces. So as FM and having more than 20 years of tournament-experience I wondered if the American writer wasn't exaggerating again. Nonetheless only a month later in the February-edition of Chessmagazine the Dutch International master and senior fide trainer Jeroen Bosch wrote something similar in one of his articles. Also he recommends to exchange pawns when standing worse.

More than 20 years I never heard this so called basic-rule but now it suddenly pops up twice in a couple of months. I checked around me and it appears that I am not the only one so maybe Jeroen just read the book "Applying logic in chess". This wouldn't surprise me as I strongly recommend this book to any serious trainer as it will be a standard work in the future for training pedagogically chess. Basic or not, known or unknown, fact is that Jeroen thinks the rule is useful for anybody so we should not ignore it.

Besides if I would've considered it nonsense then obviously I would not spent time writing about it. Also unconsciously I am sure we all are already applying sometimes that rule via the endgames. Many endgames are a draw even when one side has a material advantage of +3 when no pawns remain on the board. I am thinking about only knight or only bishop but also rook+ knight against rook or rook + bishop against bishop. I am talking purely from the theoretical point of view as there always remains practical chances as happened recently in the game Veselin Topalov against Ding Liren played in Shamkir, Azerbaijan. It is incredible how 2 absolute world-class-players managed to misplay an endgame of rook against knight despite having sufficient time remaining on the clock.

Mistakes are human especially if you need to calculate after many hours of play. Nobody is immune. However I also see many mistakes in the endgame which have nothing to do with calculations but are rather a lack of knowledge. I already wrote about this before see quicker part 2 that our youth is lagging in that domain and this once again became clear in the endgame occurring in my standard-game played in the Christmas-tournament of Deurne end of last year against fresh FM Sim Maerevoet. The 17 year old made in the 3 previous years a rating jump of no less than 600 points ! Contrary to my students he works hard at chess so is also more successful but the endgame still remains something special.

I advised Sim to work at it and I think he will do. This was shown when we did a long post-mortem after the game while all other players already left the tournament-room. While others were enjoying drinks and making a lot of noise, we tried to investigate deeper the complexity of our endgame. I told him that I wasn't sure if the endgame was won against best play so I hesitated during the game to enter it. Sim was surprised but admit that a win without exchanging rooks was not simple at all. Eventually I was able to find a narrow path to the victory after using the best engines for several hours. Clearly in practice it would've been unlikely to find all those moves.
[Event "Kersttornooi Deurne"] [Site "?"] [Date "2018.??.??"] [Round "?"] [White "Brabo"] [Black "Maerevoet, S."] [Result "1-0"] [SetUp "1"] [FEN "6k1/5p1p/3r2p1/3B4/8/8/6PP/4R1K1 w - - 0 39"] [ECO "B92"] [WhiteElo "2310"] [BlackElo "2220"] [PlyCount "33"] [EventDate "2018.??.??"] [CurrentPosition "6k1/5p1p/3r2p1/3B4/8/8/6PP/4R1K1 w - - 0 39"] 39. Rd1 Rd7 40. Bf3 Ra7 41. Rb1 Rd7 42. Kf2 Kg7 43. Rb7 Rd6 44. Ke3 h5 45. Rc7 Rd8 46. Rc5 Rd7 47. Bd5 Kf6?! {(Black wants to have the king active when the rooks are exchanged but this is insufficient as white easily wins by using zugzwang. More important is to have the f-file free so the black rook can switch from a horizon to a vertical defense.)} (47... Ra7 48. Rb5 Ra3+ 49. Kd4 Ra4+ 50. Ke5 Ra1 51. Rb7 Re1+ 52. Kd6 Rf1 {(Black managed to make the switch but he will not succeed next time.)} 53. h3 Rf6+ 54. Kc5 Rf1 55. Ra7 {(We will immediately see why it will be important to position the king at d6 and the rook at the a-file before starting to switch.)} 55... Rc1+ 56. Kd6 Rf1 57. Ra3 {(White threatens to switch to the f-file but all squares at the 7th row are covered. This forces black to make an important concession.)} 57... Rf6+ 58. Kc5 Rf2 59. Rf3 Rc2+ 60. Kd6 f6 61. Rb3 Re2 62. Rb7+ Kh6 63. Re7 Rb2 64. Ke6 {(Now the weakness of the e6 square allows white to infiltrate.)} 64... Rd2 65. Rd7 Kg5 66. Rb7 Kh6 67. Bf3 f5 68. Rd7 Rb2 69. Kf7 Kg5 70. Rd6 h4 71. Rxg6+) 48. Kd4 Re7?! {(Black has only a few minutes on the clock in the K.O.-phase but after this move the win becomes easy as black needs to exchange rooks or loses a pawn. The best was Kf5.)} 49. Rb5 h4 50. Rb7 Rxb7 51. Bxb7 Kg5 52. Ke5 f6+ 53. Ke6 f5 54. Ke5 f4 {(In the postmortem I still showed h3 as a last try to resist. An anti-dote was also quickly discovered.)} (54... h3 55. g3 Kg4 56. Ba6 Kf3 57. Bf1 Kf2 58. Bxh3 Kg1 59. Bxf5 Kxh2 60. g4 +-) 55. h3 1-0

Some would categorize my judgment as intuition but I believe it is not just that. I was trying to force an exchange of rooks as that would make the win much easier. I would only exchange pawns when all other options were exhausted. From earlier experiences I know that winning such endgames against optimal defense would be a close call and that is also proven in my long analysis. It is not easy to keep the activity of the black rook within limits, defend the white pawns and simultaneously cause a weakness in black's camp.

Ok that is all nice but how can this be studied somehow? I am no specialist of endgame-books but I don't think this type of endgames has been analyzed deeply somewhere. No I think a healthy curiosity is important to improve. I wrote in my previous article that I spend (too) much time at analyzing endgames. However it is never useless doing such research. For this type of endgame I made an extra mile by analyzing similarly endgames which were played recently. During the Christmas-holidays  I was spending family-time in Russia so anyway also had a lot of free time. I don't have chessbase but by downloading scid which can be done fully legally, I was able to make a selection of games in which the endgame of Rook + Bishop against Rook popped up but in which there were also pawns on the board at one side of the board and the side without bishop has one extra pawn. Some endgames were very interesting stuff to analyze. Below you can find the most interesting ones. I start with an endgame played in 2018 between 2 Cuban grandmasters.
[Event "Guillermo Garcia Prem 1"] [Site "Santa Clara CUB"] [Date "2018.05.23"] [Round "2.1"] [White "Quesada Perez, Luis Ernesto"] [Black "Oliva, K."] [Result "1-0"] [SetUp "1"] [FEN "8/R4pk1/2B1p1p1/4K1P1/5P2/8/8/3r4 b - - 0 43"] [ECO "A37"] [WhiteElo "2462"] [BlackElo "2511"] [PlyCount "28"] [EventDate "2018.05.22"] [Eventtype "tourn"] [CurrentPosition "8/R4pk1/2B1p1p1/4K1P1/5P2/8/8/3r4 b - - 0 43"] 43... Re1+ 44. Kd6 Rf1 45. Bd7 Kf8? {(Black wants to avoid Be8 but he does it the wrong way.)} (45... Rd1+ 46. Ke5 Rd5+ 47. Ke4 Kf8 48. Bc6 Rc5 49. Ra6 Ke7 50. Ra7+ Kf8 51. Ba4 Rc4+ 52. Ke5 Rc5+ 53. Kd4 Rf5 54. Ke4 Rc5 55. Bb3 Rc3 56. Rb7 Rc5 57. Ke3 Rf5 58. Rb8+ Kg7 59. Re8 Rb5 60. Bc2 Rb4 61. Bd3 Rb3 = {(Stockfish still shows +3 but can't make any progress. White's pawns are too advanced.)}) 46. Ke5 Rf2 47. Bb5 Rf3 48. Ra8+ Kg7 49. Be8 Rc3 50. Ra7 Rc5+ 51. Kd6 Rd5+ 52. Kc6 Kf8 {(Also Rf5 leads to a lost pawn-endgame. This theme of giving away the material-plus to achieve a won pawn-endgame occurs regularly.)} (52... Rf5 53. Rxf7+ Rxf7 54. Bxf7 Kxf7 55. Kd7 +-) 53. Bxf7 Rf5 54. Bxe6 Rxf4 55. Rf7+ Rxf7 56. Bxf7 Kxf7 57. Kd7 1-0

Despite the large evaluation of the engine, I don't see a win against a correct defense. White's pawns are too advanced so the winning mechanism as shown in my game against Sim is not possible. Nevertheless black still managed to lose the game which I regularly noticed in such type of endgames. In practice many people falter as defending such positions is far from easy. This is also the case in the next example. This time we see the Latvian grandmaster Toms Kantans collapsing while a draw was feasible.
[Event "Baltic zt III Stage 2018"] [Site "Liepaja LAT"] [Date "2018.07.15"] [Round "9.8"] [White "Bernotas, A."] [Black "Kantans, T."] [Result "1-0"] [SetUp "1"] [FEN "8/2r2pkp/3R2p1/8/5P2/7P/4B3/6K1 w - - 0 35"] [ECO "B38"] [WhiteElo "2413"] [BlackElo "2529"] [PlyCount "51"] [EventDate "2018.07.09"] [CurrentPosition "8/2r2pkp/3R2p1/8/5P2/7P/4B3/6K1 w - - 0 35"] 35. Kf2 Ra7 36. Bf3 Ra2+ 37. Kg3 Ra3 38. Rd7 Kf6 39. Rb7 Rc3 40. Rb6+ Kg7 41. h4 Ra3 42. f5 {(The scattered white pawns make the win theoretical impossible but in practice we see that black still often makes a mistake due to the pressure.)} 42... gxf5 43. Kf4 Ra4+ 44. Kg5 Rd4 45. Rb7 Rd3 46. Bh5 h6+ 47. Kxf5 Rd5+ 48. Kg4 Rd4+ 49. Kg3 Rd3+ 50. Kf2 Rd5? {(The Latvian grandmaster commits an error.)} (50... Rd4 51. Rxf7+ Kg8 52. Kg3 Rd3+ 53. Kf2 {(Here we see the problem of pawn at the edge + wrong bishop after exchanging the rooks. However the alternative is giving up the last pawn which generates a theoretical draw.)} 53... Rd4 54. Kg3 Rd3+ 55. Kf4 Rd4+ 56. Ke5 Rxh4 {(White can still try but the tablebases tell us it is a draw.)}) 51. Rxf7+ Kg8 52. Bg6 Rd6 53. h5 Rd5 54. Ra7 Rb5 55. Ke3 Rc5 56. Kf4 Rg5 57. Bf5 Rxh5 58. Ke5 Rg5 59. Kf6 Kh8 60. Bg6 1-0

Here the problem were not the advanced pawns but rather that they were not anymore connected. This doesn't allow to coordinate attack and defense. Now it are not always the defenders making mistakes. In the next example we see a very favorable version of the endgame for white but black manages to defend. It is nice performance of the Argentinean grandmaster Federico Perez Ponsa.
[Event "Liga Nacional Superior"] [Site "Villa Martelli ARG"] [Date "2018.09.01"] [Round "9.1"] [White "Tomas Falcon, Jesus"] [Black "Perez Ponsa, F."] [Result "1/2-1/2"] [SetUp "1"] [FEN "8/1R3pkp/6p1/2r5/8/5BP1/5PK1/8 b - - 0 38"] [ECO "E60"] [WhiteElo "2152"] [BlackElo "2547"] [PlyCount "54"] [EventDate "2018.07.07"] [Eventtype "team-swiss"] [Whiteteam "Tren Verde"] [Blackteam "Obras AYSA"] [CurrentPosition "8/1R3pkp/6p1/2r5/8/5BP1/5PK1/8 b - - 0 38"] 38... h5 {(White just took a bishop at b7.)} 39. Rd7 Rf5 40. Be4 Rf6 41. Rc7 Kf8 42. f3 {(White has one of the best setups to win. His pawns are connected. He doesn't have a pawn at the edge of the board and they are not far advanced. Still this time black defends which is naturally also linked to the fact that he is much higher rated than white.)} 42... Rd6 43. Kh3 Rd4 44. Bb7 Kg7 45. f4 Kf8 46. Ba6 Rd6 47. Bc4 Rf6 48. Kh4 Kg7 49. Bd5 Kf8 50. Bf3 Kg7 51. Be4 Re6 52. Bd3 Rd6 53. Bc2 Rd2 54. Kh3 Rd6 55. Bb3 Rf6 56. Kg2? {(Kh4 still wins. Now black gets an opportunity to find a draw which he manages to do nicely.)} 56... h4 {(If you are behind then you should exchange pawns.)} 57. Rxf7+ {(White tries the pawn-endgame with an extra pawn but it is not won.)} 57... Rxf7 58. Bxf7 Kxf7 59. gxh4 Kf6 60. Kf2 Kf5 61. Kf3 Ke6 62. Kg4 Kf6 63. Kf3 Kf5 64. Kg3 Ke6 65. Kg4 1/2-1/2

Black executes nicely the basic-rule of exchanging pawns when standing worse. Besides we also see that the drawing chances immediately improve when 1 pair of pawns disappear. Still it doesn't mean a draw is given easily even when black is the super-grandmaster Peter Svidler is.
[Event "80th Tata Steel GpA"] [Site "Wijk aan Zee NED"] [Date "2018.01.14"] [Round "2.7"] [White "Wei Yi"] [Black "Svidler, P."] [Result "1/2-1/2"] [SetUp "1"] [FEN "5k2/5p2/2r3p1/8/7R/8/7P/3B2K1 b - - 0 63"] [ECO "D80"] [WhiteElo "2743"] [BlackElo "2768"] [PlyCount "122"] [EventDate "2018.01.13"] [CurrentPosition "5k2/5p2/2r3p1/8/7R/8/7P/3B2K1 b - - 0 63"] 63... Kg7 64. Rd4 Rf6 65. Kg2 Kh6 66. Bf3 Kg7 67. Kg3 Kh6 68. Kf2 Kg7 69. Rd7 Kg8 70. Ke3 Kg7 71. Bd5 Kg8 72. Bc4 Kg7 73. Rc7 Rf5 74. Ra7 Rf6 75. Rb7 Rf5 76. Ba2 Rf6 77. Bb3 Rf5 78. Rd7 Rf6 79. Bc4 Rf5 80. Bd5 Rf6 81. Rb7 Rf5 82. Ke4 Kg8 83. Ra7 Kg7 84. Kd4 Rf2 85. h3 Rf1 86. Kc5 Rf6 87. Rb7 Kg8 88. Kd4 Rf5 89. Kc4 Kg7 90. Kc5 Rf6 91. Rc7 Kf8 92. Be4 Kg7 93. Kd5 Re6 94. Kd4 Rf6 95. Ra7 Kf8 96. Ra3 Kg7 97. Kd5 Re6 98. Rc3 Rf6 99. Bc2 Kf8 100. Bb3 Kg7 101. Ke5 Rf5+ 102. Kd6 Rf4 103. Bd5 Rf5 104. Rc7 Kf8 105. Ra7 Rf6+ 106. Kd7 Rf5 107. Ra8+ Kg7 108. Kd6 Rf6+ 109. Ke7 Rf5 110. Rd8 Re5+ 111. Kd6 Rf5 112. Be4 Rf6+ 113. Ke5 Re6+ 114. Kf4 Rf6+ 115. Kg5 Re6 116. Rd4 Rf6 117. h4 Re6 118. Bf3 Rf6 119. Bg4 Rf1 120. Rd7 Rf2 121. h5 gxh5 122. Bxh5 Rf1 123. Bxf7 Rxf7 124. Rxf7+ {(Black didn't blink but white definitely tried all tricks.)} 1/2-1/2

Black didn't blink. White tried all his tricks and waited as long as possible till it was not possible anymore to avoid pawn-moves. In the next example we see again it is a draw but both players can't avoid making mistakes.
[Event "Bundesliga 2016-17"] [Site "Munich GER"] [Date "2017.02.19"] [Round "8.2"] [White "Gerigk, E."] [Black "Lampert, J."] [Result "0-1"] [SetUp "1"] [FEN "8/8/2Rbk1p1/8/8/5KP1/5P1r/8 w - - 0 54"] [ECO "C50"] [WhiteElo "2381"] [BlackElo "2500"] [PlyCount "56"] [EventDate "2016.10.15"] [Eventtype "team"] [Whiteteam "MSC Zugzwang"] [Blackteam "HSK"] [CurrentPosition "8/8/2Rbk1p1/8/8/5KP1/5P1r/8 w - - 0 54"] 54. Rc2 Kf5 55. Rd2 Be5 56. Ra2 g5 57. Ra4 Rh8 58. Re4 Rd8 59. Re2 Bd4 60. Re4 Bc5 61. Rc4 Bd4 62. Rc7 Rb8 63. Kg2? {(This is too passive as now black can win with a special maneuver.)} (63. Rd7 g4+ 64. Kg2 Rb4 65. Rf7+ Bf6 66. Rf8 Kg6 67. Rg8+ Kf5 68. Rf8 Rb2 69. Rc8 Rb3 70. Rc5+ Be5 71. Rc8 Rf3 72. Rf8+ Bf6 73. Kf1 Kg6 74. Rg8+ Kh5 75. Ra8 Kg6 76. Rg8+ Bg7 77. Ke1 Rc3 78. Kf1 Rc1+ 79. Ke2 {(Stockfish keeps showing almost +2,5 but can't make any progress.)}) 63... Rb2 64. Rf7+ Ke4 65. Kf1 g4 66. Rf4+ Kd3 67. Rxg4 {(The alternative is a lost pawn-endgame.)} (67. Kg2 Rxf2+ 68. Rxf2 Bxf2 69. Kxf2 Kd2 70. Kf1 Ke3 71. Kg2 Ke2 72. Kg1 Kf3 73. Kh2 Kf2 74. Kh1 Kxg3 75. Kg1 Kh3 76. Kh1 g3 77. Kg1 g2 78. Kf2 Kh2 -+) 67... Rxf2+ 68. Ke1 Rf3 69. Rf4 Bc3+ 70. Kd1 Re3 71. Rf1 {(Funnily without the white pawn it is a draw but now it is won for black.)} 71... Bd4 72. Rf2 Bc3 73. Rf1 Re5? {(Ld2 wins but it is not simple.)} (73... Bd2 74. g4 Re2 75. Rf3+ Be3 76. Rf7 Rb2 77. Rd7+ Bd4 78. Rc7 Rg2 79. Re7 Rxg4 80. Re2 Rg1+ 81. Re1 Rg2 82. Re7 Rh2 83. Re8 Rb2 84. Rc8 Bb6 85. Rc6 Be3 86. Rd6+ Bd4 87. Rc6 Rd2+ 88. Kc1 Rh2 89. Kb1 Rb2+ 90. Kc1 Rb5 91. Rd6 Ra5 -+) 74. Rf3+ Kc4 75. Rf8 Re1+ 76. Kc2 Re2+ 77. Kd1 Rd2+ 78. Kc1 Ra2 79. Rc8+? {(White permits black to access d3. A draw was e.g. Rf3.)} 79... Kd3 80. Rd8+ Bd4 81. Rb8 Rg2 {(White resigned as he can only postpone mate by sacrificing the rook.)} (81... Rg2 82. Kb1 Rg1+ 83. Ka2 Ra1+ 84. Kb3 Rb1+) 0-1

So you always need to be alert in this endgame even if you know which position is a draw or not. For me analyzing such endgames is fun and it also extends my horizon of the endgame. Only in 2018 I found dozens of this type of endgame in the big database. Some of them were an exact copy of my position against Sim. Also I do think some conclusions are valid for other type of endgames.

Finally I find it very important to think via concepts instead of concrete moves. You first need to figure out what you want to achieve and then you need to match the right moves to your idea. In my recent courses in KMSK I obliged my students to train such endgames by playing them out against each other. I opened a new world for them as they never tried to play chess in such way.

Brabo

No comments:

Post a Comment