< Earlier Kibitzing · PAGE 3 OF 3 ·
|Dec-30-06|| ||euripides: <capanegra> A good deal has been learned about endings since Capablanca wrote his book. One of the endings that changed as a result of computer analysis was Q vs R, which turned out to be harder than had been thought (though generally won). Many of the basic books pretend this ending is simpler than it actually is. |
Morozevich's failure to win it was a slip, probably relaxing after doing the hard part, not lack of knowledge - everyone knows this stalemate trap.
|Dec-30-06|| ||capanegra: Fine folks, maybe I overreacted a little bit. I didn't pretend to discredit Morozevich, for he is a great fighter, and certainly one of the most imaginative players of today. But I still believe that failing to win a Q vs R ending for a top GM is grave, no matter how hard it might be.|
Having said that let me wish you a Happy New year!
|Dec-31-06|| ||DutchDunce: Amazingly, Moro was on the losing end of a Q vs R in 2003, against Shirov. Also amazingly, that game was also 114 moves: Morozevich vs Shirov, 2003|
|Dec-31-06|| ||FHBradley: Elementary or not, playing through this endgame with the help of Nalimov is most instructive to a Patzer like me. Around move 80 Morozevich suddenly begins to lose the thread, having played the endgame immaculately before that: 81. Qf6 would have given white an easy and comfortable victory.|
|Dec-31-06|| ||TylerD: Capanegra: happay new year to you too!|
|Jan-02-07|| ||FJCF: yes 69.b6! Kb5 70.b7 Kc6 71.Nc5! followed Kf7 is 1-0 easy.Is the time... :-)|
|Jan-02-07|| ||FJCF: But 71.Nc5 Kc5 72.Kf7 Rb8 73.g8=Q Rb7 and another ending Q+R|
|Jan-02-07|| ||malthrope: <FJCF: yes 69.b6! Kb5 70.b7 Kc6 71.Nc5! followed Kf7 is 1-0 easy.Is the time... :-)>|
Followed by... <FJCF: But 71.Nc5 Kc5 72.Kf7 Rb8 73.g8=Q Rb7 and another ending Q+R>
The idea behind 69. b6! Kb5 70. b7 Kc6 (or Kb6 and if 70... Ka6 then 71. Nc5+ followed by 72. Kf7 wins) is (instead of 71. Nc5) 71. b8=Q Rxb8 72. Nf8! wins. This was all pointed out by <Mameluk:It looks almost impossible that Moro overlooked 69. b6-b7-b8. Something was already wrong.> on the previous page.
We surmised that Jako noticed this and decided it best to play on and try his luck once the 'Q vs. R' ending was achieved on move #71 with Kxb5, etc. Anyway, 'Good Luck' to Sasha in the '1st ACP World Rapid Cup' starting on January 4th, 2007! ;-) - Mal
|Jan-08-07|| ||Domdaniel: Q-vs-R looks very easy when you play through it with an engine, and see the wins on move 81, move 111 etc. It's a lot harder in practice, although clearly a super-GM like Moro should have the technique.|
I saw the same thing happen in the 1980s, in the deciding game of the Irish championship between two players rated about 2300. The guy with the Queen had to win the game for the title, but it ended in a draw and somebody else won on tie-break.
The really incredible thing about this game is the way Jakovenko defends for so long with a lone Rook -- first against knight and four pawns, then against a queen. It's one of the finest sustained defensive efforts in the history of chess. The psychological factor -- when your opponent is dead but refuses to lie down -- should never be underestimated.
|Jan-09-07|| ||soberknight: |
click for larger view
White to play at move 72: tablebase says mate in 28 (standard for this distribution of the pieces). Maximum distance to mate for this material is 35 moves in this position:
click for larger view
. Maximum distance to conversion, according to Monty Newborn and David Levy in "How Computers Play Chess," is 30 moves. GM Walter Browne played two queen versus rook games against a computer (maybe BELLE) in the 1980s, with the task for him to reach the conversion within the standard 50-move limit. Browne got stuck the first time, but he came back the next week after intensive practice, and captured the rook on the 50th move with no time to spare.
|Jan-09-07|| ||soberknight: Now tablebase comments on the moves of the ending:
72. Qc8 /28 (optimal)
...Rd5 /23 (suboptimal, makes mate 4 moves closer, hereafter -4)
73. Ke6 /26 (Qc3 was /23)
74. Ke5 /24
75. Qc2 /23
...Rd8 /22 (optimal by both players)
76. Qb3+ /22
77. Qc3+ /21
78. Ke6 /19
79. Qc4 /18
80. Qd4+ /17
So far the play on both sides has been superb, almost without mistakes. Now Morozevich gets stuck.
81. Qc3+ /18 (-2, Ke7 or Qf6 were optimal)
82. Kd6? /26 (-9, one of the worst moves that doesn't lose the queen, even though it looks aggressive. Qe5+ was optimal.)
83. Kc7 /25
...Rg4! /24 (The only move. The next best alternative was Ra6, allowing mate in 13.)
84. Qc6+ /26 (-2)
85. Qd6+ /28 (-2, but a good try)
...Kc3! /27 (other moves lose the rook quickly. Ka5 is mate in 10 after Qd2+!)
So now Morozevich must start all over again.
86. Kc6 /27
87. Qa3+ /27
88. Kc5 /26
89. Kd5 /24
90. Qf3 /23
91. Kc5 /22
...Ra4 /21 (Several consecutive optimal moves have just passed.)
92. Qf6 /25 (But not this one! Qf2 was best.)
...Kd3 /20 (Returning the favor, instead of Ke3.)
93. Qd6+ /28 (-8, Qf1! was mate in 20)
I'm too tired to finish this analysis. You can do it yourself by using the Lokasoft tablebase at http://www.lokasoft.nl/uk/tbweb.htm. A couple of critical mistakes by Moro, and some beautiful defensive moves by his opponent, did the trick.
|Jan-10-07|| ||positionalgenius: Poor Morozevich...I've done this more than once.|
|Jan-10-07|| ||malthrope: <soberknight: [...] GM Walter Browne played two queen versus rook games against a computer (maybe BELLE) in the 1980s, with the task for him to reach the conversion within the standard 50-move limit. Browne got stuck the first time, but he came back the next week after intensive practice, and captured the rook on the 50th move with no time to spare.>|
I can verify that story... :-)
Shawn - Walter Shawn Browne - is an old friend (today is his birthday, btw, born January 10, 1949). It was all based on a bet (could he or couldn't he?). So, that's exactly what happened he drew the first 'Q vs. R' game (failed to win in 50 moves). Then he studied the 'clever computer defenses' and in the next game won the rook on move 50 (just in the nick of time)... ;-) Regards, - Mal
PS: Thanks for posting the 'Q vs. R' endgame Nalimov tablebase analysis highlights. Note: <Mameluk> did a very nice job previously (I also added some stuff here in the 'Moro vs. Jako' game in the past few kibitz pages), and also posted a recap over at Mig's ChessNinja Forums - International Events / "XVI Torneo Internacional - 'Cuidad De Pamplona' - Dec. '06" // http://www.chessninja.com/cgi-bin/u...
|Jul-23-07|| ||Ashram64: good learning game|
|Jul-24-07|| ||argishti: WAT A BLOW!! wowww|
|Nov-01-08|| ||Eggman: <<soberknight>> Those tablebase figures are misleading (though thanks for the effort), because they give the number of moves to reach checkmate when in fact, for intents and purposes, winning the Rook is good enough.|
|Jan-12-09|| ||Lucid Faia: 114. Kxd2
|Jul-09-09|| ||YouRang: Yes, K+Q vs. K+R is not easy, but on the other hand, Moro had already done the hard part, and then let it get away at the end by allowing a crazy-rook stalemate. |
In positions like the one Moro had after 110 moves, white must use zugzwang tactics to force black into tighter positions until the king can't defend the rook any longer, and eventually the rook falls to a queen fork, skewer or pin.
In this case, 111.Qe5! maintains the grip on a square next to the king (h2) while setting up checks at e1 or a1. This is typicaly of what to look for in such situations.
|Jul-09-09|| ||aragorn69: Was this rapid??|
|Jul-09-09|| ||plang: <Was this rapid??>
|Jul-10-09|| ||popski: Oh man, that's a great ending to learn from!|
|Aug-17-12|| ||geniokov: Is there any useful golden rule to drawn the game of K+Q vs K+R ending in favor who handles the K+R?|
|Aug-17-12|| ||kasputine: 63. Kf4 would have easily won,
63. ... Rxa4+
|Aug-17-12|| ||Black Vampire: <geniokov> Unfortunately, given the fact the ending K+Q vs K+R is a mathematical win for the strong side, there´s no useful golden rule for the weak side, except at certain stalemate positions, such as the one which happened in the game.|
|Jul-08-15|| ||David2009: You can try your luck at winning the ending against Crafty End Game Trainer:
< Earlier Kibitzing · PAGE 3 OF 3 ·