|
< Earlier Kibitzing · PAGE 3 OF 4 ·
Later Kibitzing> |
| Jun-12-10 | | TheBish: R Calvo vs D Byrne, 1968 White to play (68.?) "Very Difficult"
This only seems difficult, until you see the zugzwang theme. 68. Be2 (also winning is 68. b8=Q!, as we will see later, but I believe the text is quicker) Be5 Other bishop moves transpose, and a pawn push by Black hastens the end. 69. b8=Q!! Bxb8+ 70. Kb7!
The point! Of course not 70. Kxb8?? d1=Q 71. Bxd1 Kxa6 draw. Now Black will have to give up both pawns (he's in zugzwang), and then either give up the bishop (after the black king moves away from the white pawn) in which case the white pawn queens, or move the bishop and let the pawn queen anyway. 70...g4 71. Bxg4 Kb5 72. Bf3 Ka5 73. Be2! (second zugzwang) d1=Q 74. Bxd1 Kb5 75. Be2+ Ka5 76. Kd3 (third zugzwang) and now Black can resign, if he hasn't already! White queens after the bishop moves, and then the mating game begins. |
|
| Jun-12-10 | | sethoflagos: 
click for larger view69. ... ?
Crafty 23.00 grasps the winning principle immediately, religiously and repeatedly delaying Be2 until BK is on a5. It sees 69. ... Kb5 and 69. ... g4 as equivalent and losing. Same with Comet B68. In fact both engines rapidly give up on trying to prevent a8=Q in favour of increasing survival time by saving the bishop. Fritz 5.32, 12 and Rybka 3 all seem to need prompting before they see the light. |
|
Jun-12-10
 | | LIFE Master AJ: How weird is this? (Started at 06:12 AM Saturday morning.) I dreamed about this position last night! And I woke up knowing that White should win, I guess I was too tired last night to do it right. ***
ONCE
(Your post was so "right-on" and hit the nail on the head so exactly, that I decided to reproduce it here.) < One interesting point here is how a non-tablebase engine can be fooled by this positions like one.
After 69. Kb7, Fritzie is saying that 69...Kb5 is 0.00, giving only 70. Be2+ as white's reply. All other moves lose for black.But as soon as we play the appparently drawing 69...Kb5, Fritzie gets the message and comes over all priapic about about 70. Bh5, 70. Bf3 and 70. Bg4. So why should the engine be so dumb? One answer might be the horizon effect. As the computer only analyses so many moves ahead, it doesn't see a win that happens just beyond its analysis. <I have been humbled by Fritzie so many times, that today is a rare treat. And whenever he is rude to me in future as in "I told you you gotta study opening theory", <I will quietly whisper: <"yes, but let's not forget that you wanted to play Be2+" >>>> By the way, I have all the tablebases, Once. But as far as I know, there are no TB's for eight pieces. So its impossible to START with a TB engine ... which I know how to use. *****
dzechiel
<You are most correct, sir.> The engines don't see it, at least not right away. In fact, Fritz has to get almost all the way to the correct position before it "SEES" the win. However, my brain refused to let go of this one, even in sleep. It is certainly a ZUGZWANG position, if ever there was one! To me, its tricky and kind of transpositional. It is also very strange that we finally have an opposite-colored Bishop ending that is NOT a draw. (hee-hee) ***
Anyway, I admit my mistake. Part of the reason that I believed I was correct was I too have come to regard the chess engines as virtually infallible. When they (initially) showed that the game was a draw, I simply took that as proof that I was correct. (This is a common human failing.) By the way, I almost never set the position up on any kind of board. I usually do all my analysis (BLINDFOLD) from the diagram. I know that many of you set the position up on a chess board - and even move the pieces ... but I don't. |
|
| Jun-12-10 | | shatranj7: To me, it seems that 68.Be2 is a better continuation. Also, I don't understand why white wins. Was it on time? |
|
Jun-12-10
| | LIFE Master AJ: <Jun-12-10 <sethoflagos:> <Crafty 23.00 grasps the winning principle immediately, religiously and repeatedly delaying Be2 ... > ... ... ...
<Fritz 5.32, 12 and Rybka 3 all seem to need prompting before they see the light.>> I will ignore the negative comments of some and simply point out that I have Fritz 12, Rybka 3, etc. And I will simply confirm what Mr. <setoflagos> has already pointed out, that Fritz 12 and Rybka 3 do NOT see the win ... at least, not initially. (I did let Fritz run while I was sleeping. And it WAS showing the endgame as a win - WHEN i FIRST WOKE UP, AFTER RUNNING THE ENTIRE TIME i WAS SLEEPING. But then I shut everything off and I re-booted my computer, which is something that I do every morning. Then I re-fired up Fritz 12, but first I cleared the hashtables and reset the engine back to its original defaults. Guess what? Its back to SHOWING THAT THE ENDGAME IS A DRAW!!!!) |
|
Jun-12-10
| | LIFE Master AJ: One more note. I have a whole set of DVD's with the endgame TABLE-BASES on them. They are installed on my hard drive, and Fritz does have them in the pathway and is acccessing them. I tested a few simple positions ... like ♔+♕ vs. ♔+♖ ... and Fritz got them right. So I have no clue as why Fritz would mess up this ending ... at least not as far as table-bases are concerned. |
|
| Jun-12-10 | | JG27Pyth: Superb puzzle today! And quality kibitzs too! :) |
|
Jun-12-10
 | | OBIT: <AJ>OK, but if Fritz has access to the tablebases, shouldn't it be finding the rather obvious 72. Bxg4 after 68.b8Q Bxb8+ 69.Kb7 Kb5 70.Bg4 Ka5 71.Bd1 (Fritz' analysis, but best is 71. Be2!) g4? After the capture, the game is down to six pieces. |
|
| Jun-12-10 | | CHESSTTCAMPS: In this opposite-colored bishops ending, white has an obvious edge with the connected passed pawns near promotion. But how does white actually make progress to force a promotion? The lines 68.b8=Q Bxb8+ 69.Kxb8 Kxa3 and 68.Be2 Bf4 69.Ka8? d1=Q 70.Bxd1 Kxa3 cost white his all-important a-pawn. Somehow, the black king must be forced to abandon his defensive post on a5. A little nore subtlety is needed: 68.b8=Q! (anyway) Bxb8 69.Kb7!
Now black is almost in zugswang. The black bishop is frozen, black's pawns have no value, and the K has one other defensive square - for now. 69... Kb5 70.Bg4! Ka5 71.Be2!
Now black can prolong the agony by giving up pawns, but it is obvious that black can't stop the a-pawn. My guess is that Byrne resugned after 69.Kb7 if Calvo found the right line. Time to check... |
|
| Jun-12-10 | | CHESSTTCAMPS: ...or he might have resigned instead of resugned. |
|
| Jun-12-10 | | Once: I don't know, resugnation sounds quite a funky way to end a game! |
|
| Jun-12-10 | | johnlspouge: Saturday (Very Difficult)
R Calvo vs D Byrne, 1968 (68.?) White to play and win.
Material: Even, with Bs of opposite color. The White Bd1 must stay on the d1-h5 diagonal to stop Pd2 and Pg5. The Black Bc7 must stay on the b8-h2 diagonal to stop Pb7. Moreover, the Black K must stay on either a5 or b5, to prevent the White sequence Pb7-b8=Q Bc7xb8 Ka7xb8 with the reply Kxa6. White also has the possibility Ka7-a8 Pa6-a7 Pb7-b8=Q. Both possibilities point to the candidate 68.Be2, to create a zugzwang. Candidates (68.): Be2
[68.Be2 Bf4 69.b8=Q Bxb8 70.Kxb8 d1=Q 71.Bxd1 Kxa6]
I missed this one cleanly. |
|
| Jun-12-10 | | randomsac: Got the idea, since only the bishop could stop promotion, white only has to put black in zugwang and trap the bishop on the edge of the board. white will eventually play Be2 to force either the king away (leaving the bishop free), or the bishop to let the pawn promote. |
|
| Jun-12-10 | | newzild: I found this much easier than yesterday (obviously - given that I got this one and missed yesterdays...) |
|
| Jun-12-10 | | kevin86: The first COY move is to promote the pawn-the second is to NOT take the bishop. The third will come after 70...♗a5 71 ♗e2! and black is in zugzwang! |
|
Jun-12-10
 | | Peligroso Patzer: <mertens: The "direct" attempt 68. Be2 Bf4 (or somewhere else along the diagonal h2-b8) 69. b8=Q Bxb8 <70. Kxb8> fails to 70. - d1=Q and 71. - Kxa6, so White has to free his King in order to promote the a-Pawn. After some thought it strangely appears that the only square that leads out of prison is b7, so 68. b8=Q Bxb8 69. Kb7. Now the black bishop can't move due to the advance of White's a-Pawn, and the black King is tied to White's a-Pawn (watching for d1=Q and Kxa6 as above) and consequently has to shift between a5 and b5. So White might operate with zugzwang, e. g. 69. - Kb5 70. Bf3 Ka5 71. Be2! (zugzwang) g4 72. Bxg4 Kb5 73. Bd1 Ka5 74. Be2! (once again zugzwang) d1=Q 75. Bxd1 Kb5 76. Be2+ and 77. Kxb8 and wins. <I hope I did not overlook anything> ...> The immediate 68. Be2 <is winning>, but in the line given at the beginning of the above-quoted post, White must play <70. Kb7> (rather than Kxb8?). Play then could continue 70. … g4 71. Bxg4 Kb5 72. Bd1 Ka5 73. Be2 d1=Q 74. Bxd1 Kb5 75. Bf3 Ka5 76. Be2 and Black is in zugzwang. White’s winning plan from the position given as today’s puzzle involves both (I) an unusual pawn sacrifice (i.e., the promotion of the b-pawn when it can be taken by the Black DSB, after which KxB(b8) is a legal move but a blunder instead of which the winning move is Kb7) and (II) tempo moves with the LSB to put Black in zugzwang. Since the second element of the plan requires White to play Be2 when the Black King is on a5, it is most accurate to do this on move #68, but either 68. Be2 or 68. b8=Q preserves a theoretical win for White. |
|
| Jun-12-10 | | lost in space: Missed this completely |
|
| Jun-12-10 | | turbo231: I thought I had this one figured out, but I didn't. Next time. I saw the zugwang but didn't think I needed it, silly me. |
|
Jun-12-10
| | Peligroso Patzer: To pick one final nit re: today’s puzzle, the assertion that 68. Be2 is more accurate than Calvo’s move (68. b8=Q) is that, if Black plays the ending out “to the bitter end” by sacrificing both his pawns before White is able to create the following zugzwang position (with Black to move, of course): 
click for larger view… if 68. Be2 had been played, the above position would have occurred after <76.> Be2, but with Calvo’s move (68. b8=Q), it would occur one move later (after <77.> Be2). This extra move arises due to the fact that with Calvo's move order, White has to burn a tempo with 70. Bg4 to force Black to move his King back to a5 (having missed the chance to play 68. Be2 after 67. … Ka5). The distinction between 68. Be2 and 68. b8=Q is academic in this case, but the thought process involved in making that distinction involves concepts that sometimes are necessary to choose between moves with different theoretical outcomes, so I believe this exercise in nit-picking is "worth the candle". |
|
Jun-12-10
 | | Jimfromprovidence: I wonder why black did not play 62...Bf4 (or 62...Bd6 or 62...Bd8) instead of 62...d2. 
click for larger viewThis looks like it was the way for black to hold. Now, white has to play 63 Bxd3 (to prevent 63...d2) and black can safely follow with 63...g4. 
click for larger viewNow it looks like neither side can make progress. For example, if 64 b7 then 64...g3 65 Bf1 Bc7 should follow. 
click for larger view |
|
Jun-12-10
| | doubledrooks: Zugzwang is the theme today, as white wins after 68. b8=Q Bxb8+ 69. Kb7 Kb5 and now white simply has to make a move on the d1-h5 diagonal other than Be2, and then after black plays 70...Ka5 play 71. Be2. The theme just repeats itself for several moves as black gives up his pawns, and after the final Be2 black is in zugzwang. After the black king moves away from the a6 pawn white can capture the bishop and then queen the a pawn. The posts were very interesting today, and goes to show that the engines are not always trustworthy. |
|
| Jun-12-10 | | Ferro: Las erratas son de tipo "gonzo": Que hago yo aqui? Donde me he metido? Una directa exclusiva del señor descartes a punto: toma! "Las palabras son ingeniosas se las lleva el viento" Donde esta el tipo motor ese? Que se lo lleva el viento. Un Alfil al cuadrado debe ganar! Como seA! Especifique el tipo de Sufragio... 1. Ab2 x Casilla Inicial bz, Antes tower of London a mi Fitzer me parece un Kennedy algo bonzo (sin calvo Sotelo) Raphodinggen:
70 …Ka5
71 Be2 g4
78 Bxg4 Kb5
79 Bf3? Ka5
80 Be2 d1=Q
76 Bxd1 Kb5
78 Etc... |
|
Jun-12-10
| | OBIT: <doublerooks>I have to say, I'm very skeptical of the claims that these high level machines are misanalyzing the position. Assuming the machines have access to a six-piece tablebase, it only takes four moves to get there by the most direct path: 68. Be2 Bd6 69. b8=Q Bxb8 70. Kb7 g4 71. Bxg4, and we now have a tablebase win. Four moves can't be beyond the computer's horizon. Um, any chance this is just human error? |
|
Jun-12-10
| | OBIT: Just for the fun of it, would anyone with Fritz or Rybka like to try the following position?
It is White to move. See if the bot gets the right answer:
click for larger view |
|
Jun-12-10
 | | Sneaky: <Peligroso Patzer: <Sneaky: Please note...
click for larger view
WHICH IS A DRAW!! Anytime White gets the nerve to try Kxb8 then Black immediately plays d1=Q, game drawn. Poor Byrne, he resigned when he could have salvaged half the point. *** >
The foregoing analysis is <not correct>.> Thank you for pointing that out, I tried to help clarify what's going on in this position last night, but now I find myself retracting virtually every single point I made. That means it that nobody resigned inappropriately in the game either--the final position really is 1-0. <Tablebases (http://www.k4it.de/index.php?topic=...) verify that the above-diagrammed position from the post by <Sneaky> is 1-0 with either side to move. Black to move is in zugzwang; White to move wins with Bishop tempo moves along the d1-h5 diagonal until he can play Be2 when Black's King is on a5.> I certainly am not going to argue with tablebase, but at first I still didn't see it. Now I'm pretty sure I do. At some point White has to play a waiting move like Bf3 or Bh5, then when Black plays Ka5 White responds with Be2! cutting off the king--and if the king can't move the bishop has to move, and if the bishop moves it's time for a7 and a8=Q. Good problem, it's got more layers than an onion. |
|
 |
|
|
< Earlier Kibitzing · PAGE 3 OF 4 ·
Later Kibitzing> |
|
|
|
Kibitzing
