< Earlier Kibitzing · PAGE 3 OF 3 ·
|May-02-10|| ||VincentL: "Insane".
I see only one way to stop black's a pawn queening: 36. Ra1 Kxa1 37 Kc1.
This blocks tha black king in permanently, since c1 is a dark square and black has a light
Black has no other threats (providing white keeps the h pawn on h2, h4 or possibly h6).
Now we have to work out how white wins.
White can in fact threaten mate in two from this position, with Nd4 followed by Nc2#/Nb6#.
This is the key, Black's bishop cannot easily prevent both mates, and also stop white's f pawn.
For example if black plays 37.....Ba4 then 38. Nd4, and white simply advances the
f pawn to f8. Black cannot pay Bb3, since Nxb3 will be mate.
Black's other a pawn cannot help - black would need one tempo more. After 37....a5 38. Nd4 and
there is not time to play Bb3 and a4.
We must check the defence 37....Be2 followed by Bd1 ! Of course white cannot respond Kxd1, since black would then play Kb1 and a1=Q. But white can simply push the f pawn
forwards. So 37..... Be2 38. f4, etc.
This all seems much too easy for Sunday. Am I missing something?
The only other line to consider is at the beginning. What happens if black declines the rook and
does not play Kxa1? The most obvious option would be 36.... Bc4. Now, if in any subsequent
move black plays Kxa1, again white responds with Kc1 and we return to the scenario above
(assuming white keeps his knight and f pawn on the board).
Here, I think the way forward for white is to play Nd4, with the option Nc2. If at any
time black plays Bxc2, white plays Rxa2+ !, and after Kxa2 white has Kxc2, and the f pawn
will march up the f file to queen (before black's other a pawn can do the same).
There are some other variations we need to consider in this line (37. Ra1 Bc4 38. Nd4).
White continues by advancing the f pawn. When it gets to f6, white must put his knight on a square
defending f7, and then advance the pawn to f7. After black plays Bxf7, and white responds with
Nxf7, white then moves his knight round the board, capturing both the other a pawn and the
h pawn. Then white queens the h pawn. Of course if at any time black plays Kxa1, then Kc1
and we continue as above.
Finally black has the option 36.....Ka3. Now the easiest win for white is by manoeuvring his
knight to c3, and then playing Nxa2. After black responds Bxa2, white has Rxa2+, and then
after black plays Kxa2 white simply marches the f pawn forward to queen. If in the knight
manoeuvre black exchanges bishop for knight, again white marches the f pawn forward.
I think this exhausts all possibilities.
Time to check.
|May-02-10|| ||VincentL: I think this is only the second time I have solved a Sunday puzzle.|
Much depends on whether you set off on the right track. If, for example, I had found some devilish defence to Ra1, and realised after 15 -20 minutes of analysis that I must start Nd4 (or with some other move), I would probably have thrown the towel in and looked at the game.
|May-02-10|| ||Marmot PFL: 36 Ra1 is think out of the box. I didn't really have much time to look at it this morning and must admit it didn't occur to me. Estrin was a WC correspondence player and probably thought of this plan many moves before. Probably Boey did too at some point but too late to turn back.|
|May-02-10|| ||Domdaniel: Suppose Black refuses the Rook on a1, but plays instead to frustrate White? Can he run himself out of moves and force a stalemate?|
Not unless White goes wrong. For example, 36.Ra1 Bc4 37.Nd4 a5 -- White's smothered mate threat is ready, the f-pawn can move, and Black's pawn or Bishop moves achieve nothing -- 38.f4 a4 39.f5 a3 40.Nc6 h5 41.h4 Bf7 42.f6 Bc4 43.Ne5 Bf7
Now, obviously, 44.Nxf7 spoils everything by allowing 44...Kxa1 45.Kc1 stalemate. Alternatives to the stalemate, eg 45.Ng5 Kb1 46.f7 a1Q 47.f8Q Qb2+, are good for Black: White's Knight is much less valuable than Black's 2nd a-pawn.
So we don't play 44.Nxf7, but 44.Nd3+:
click for larger view
44...Kxa1 leads to mate after 45.Kc1 Be8 46.Nb4 Bg6 47.f7 -- which either decoys the Bishop to permit Nc2# or allows White to play f8B(!) and Bg7#.
If Black *still* refuses the Ra1 after 44.Nd3+, White wins after 44...Kb3 45.Nc1+ Kb4 46.Rxa2! -- the Rook is still untouchable and White has an easy win.
|May-02-10|| ||SharpAttack: It took me 10 seconds to see the move Ra1!! Yeah I am so happy!!|
|May-02-10|| ||wals: Rybka 3 1-cpu: 3071 mb hash: depth 18:
Calculates Ra1, + 5.12, and Nd4 are equal.
I understood anything can be used in
correspondence games, if so,
32...g5, +2.96. better Kxb2 +0.74.and
33...g4, +6.46. better Kxb2 +2.96,
are hard to fathom.
|May-02-10|| ||sfm: 1I don't think the average club player will need more than 3 seconds to get the idea and less than 5-10 seconds more to convince himself that Black only can watch as the f-pawn marches. Then comes 10 seconds of considerations whether Black can give the bishop and stalemate himself.
Am I right?
Nice with some easy "Insane"'s so the incurable addicts among us do not see half an hour of a good Sunday go down the drain.
|May-02-10|| ||johnlspouge: Sunday (Insane)
Estrin vs J Boey, 1987 (36.?)
White to play and win.
Material: R+N for B+P. The Black Kb2 has 2 legal moves. The Black Pa2 threatens a1=Q, forcing the sacrifice of the White Rd1. The cramped position of Kb2 suggests an interesting candidate.
Candidates (36.): Ra1
(1) Black cannot accept the sacrifice:
36…Kxa1 37.Kc1 (threatening 38.Nd4 then 39.Nb3# or 39.Nc2#)
37...Ba4 (now or next move)
38.Nd4 (threatening 39.Nc2+ Bxc2 40.Kxc2)
A count shows that after White executes the threat with Kc1xc2, White requires 6 moves to mate, and Black is stalemated on his 5 move. Because Black moves first, he seems to achieve stalemate. He must, however, clear his Ba4 to permit Pa5 to move, and the B move drops a tempo and permits White to mate. In the only other defense, Black captures Pc3 with his B, but then permits Nd4-c2#.
(2) If Black does not accept the sacrifice, his Ra1 ties Kb2 to a1. White plays Ne6-d4 and then advances his Pc3. Black can sacrifice Bb5 for Pc3 at c5, where the Nd4 supports it, but clearly loses. If Black does not sacrifice at c5, White advances his Pc3 to c6, and then can play Nd4-c2, stopping Pa7 at a8. If Black dares to capture, White plays Pc6-c7-c8=Q and wins.
|May-02-10|| ||agb2002: White has a rook and a knight for a bishop and a pawn.|
Black threatens 36... a1=Q 37.Rxa1 Kxa1 followed by the advance of the a-pawn.
White can stop this with 36.Ra1:
A) 36... Kxa1 37.Kc1 Ba4 38.Nd4 and the f-pawn promotes or White delivers mate with Nc2.
B) 36... a5 37.Nd4
B.1) 37... Ba4 38.f4 is similar to A.
B.2) 37... Be8 38.f4 Bg6 39.f5 a4 40.fxg6 hxg6 41.h4 a3 42.Kd1 Kxa1 (42... Kb3 43.Kc1) 43.Kc1 g5 44.Nc2#.
C) 36... Ba4 37.Nd4 Be8 38.f4 Bd7 39.f5 Ba4 40.f6 Be8 41.Nc2 Bg6 42.Rc1 a4 (42... Bxc2 43.Rxc2+ Kb1 44.Rxa2 Kxa2 45.f7 + -) 43.Na1 h6 (43... Bb1 44.f7 Kxa1 45.f8=Q Kb2 46.Qb4+ Ka1 47.Qc3#) 44.Rf1 + -.
|May-02-10|| ||hedgeh0g: A really beautiful rook sac which I completely overlooked. Accepting it would leave Black helpless against a swift checkmate of his trapped king. Very nice puzzle!|
|May-02-10|| ||Domdaniel: <wals> -- < I understood anything can be used in correspondence games>|
You mean engines? Check the date of the game. The best computers in the early 1980s were rated below 2000, or were too erratic to rate properly. Not much use for CC either way.
|May-02-10|| ||Domdaniel: In fact this game seems to be wrongly dated. It's given as 1987, but Estrin's other games from the same event are all 1978. Even if that's the starting date of the event, it was over by 1984.|
|May-02-10|| ||wals: I just read a comment from GM Patrick Wolff
" The eye is less quick to pick up 'long moves' ".
Hope this changes your life forever.
|May-02-10|| ||wals: <Domdaniel> Thanks for the info re engines.
I can now rest easy.|
|May-02-10|| ||TheBish: Estrin vs J Boey, 1987|
White to play (36.?) "Insane"
If I'm correct on this, it will be the quickest I ever solved a Sunday problem! I don't think I ever saw this before, but it's possible, or maybe I have seen a similar endgame study. If Black is allowed to queen, then it's probably just a draw, but still problematic. How do you stop Black from queening?
An amazing resource! Black can't take the rook, because if 36...Kxa1 37. Kc1 seals the tomb on Black's king, just awaiting the final nail in the coffin from the white knight, i.e. 37...a5 38. Nd4 Ba4 39. f4 and it's just a few moves from mate, as Black's bishop can't stop the pawn and prevent mate at the same time. So...
36...Bd7 37. Nd4 a5 38. f4 a4 39. f5 a3
Or 39...Bxf5 40. Nxf5 Kxa1 41. Kc1 a3 42. Ne3 h5 43. Nc2#.
40. f6 Be8 41. Nc2 Bg6 42. Nb4 h5
Of course, 42...Kxa1 still loses to 43. Kc1.
Now 43...Bxd3 loses to 44. f7 and if 44...Kxa1 45. Kc1 and mate will come soon from the new queen.
43...Kb3 44. Ne5 Be8 45. Kc1!
This puts matters to rest, as Black has no more drawing chances. I would imagine Black resigns long before now. Of course, there are many other variations, but the principle is the same -- Black gets mated if he takes the rook on a1!
|May-02-10|| ||dzechiel: <Once: If the opponent had been desperate to win, he could probably have argued that it was too late to change the outcome. Under a strict interpretation of the rules of chess the original game result ought to have stood.>|
It might not be clear from the write up, but the game was still underway when Mike hunted me down. When his opponent's knight suddenly prevented the king from shuttling between c1 an c2, Mike stood up and allowed his clock to tick while we reviewed the game score.
Once the "irregularity" was found, the rules mandated that the game be reset to that point and a legal move made in lieu of the illegal move. When we told the guy this, he immediately "accepted" the draw (even though it wasn't technically on the table at that moment).
|May-02-10|| ||dzechiel: <Rama: Hey dz, in your story did both score sheets match? Did the N-guy also faithfully record Nc5-b8?>|
I never saw the opponent's score sheet, but he would have been hard pressed to get to the position on the board without recording at least one illegal move.
|May-02-10|| ||FSR: "Insane"? I'm probably missing something, but 36.Ra1 Kxa1 37.Kc1 looks like a trivial win. In response to the coming Nd4, Black will have to play ...Ba4 to stop the threats of Nb3# and Nc2#. Then White simply promotes his f-pawn (a bishop will suffice) and mates.|
|May-02-10|| ||muralman: Well, not a club player, but this one was solvable, anyway. I could see creating a log jam with the rook, allowing the passed pawn to move forward.|
|May-03-10|| ||Once: <dzechiel> A pedantic oppponent might call that external assistance:|
"12.3: ... During play the players are forbidden to make use of any notes, sources of information or advice, or analyse on another chessboard"
|May-03-10|| ||turbo231: This is the first time that I have solved a Sunday puzzle. And it only took about a minute to do it. I saw if black takes the rook then the black King is trapped. A neat puzzle.|
|May-03-10|| ||dzechiel: <Once: <dzechiel> A pedantic oppponent might call that external assistance:>|
Perhaps, but this guy was hardly in a position to claim that others were cheating.
|May-03-10|| ||gofer: I really liked this one. As soon as you see that Ra1 is going to make all
black's pawns impotent, then everything else is obvious. I wonder whether it is better to play Ra1 immediately or play it after Nd4?! |
36 Nd4 Ba4 (to stop Nc2 - also a1=Q is not playable as 37 Rxa1 Kxa1 38 Nxb5 axb5 39 Kc3 is a win for white)
winning as all black's pieces are misplaced and Pf3 simply marches up the board.
The beauty and simplicity of Kxa1 Kc1! is difficult to see, but wonderful to behold!
At some point the Ba4 is going to be deflected to stop Pf3's charge down the f file and at that point Nc2 becomes playable and from there Rf1 and Na1 winning.
|May-03-10|| ||Marmot PFL: <I have seen a game end with a mad time scramble where one of the players accidentally left his king in check and neither player spotted it. The spectators had, but of course couldn't say anything.>|
I've seen it too, but in that case one of the spectators did point it out. After a brief argument the illegal move was retracted, and the player soon resigned as moving through check was his only way to catch a passed pawn. The TD also warned the spectator.
Another incident I heard about but didn't see was in a team game time scramble where a team mate created a distraction while a player moved a pawn from the 3rd to the 5th rank and won. I probably would have pointed that out too, and if they want to kick me out, go ahead. I saw things similar to that too often.
|May-03-10|| ||Marmot PFL: 80s comps weren't too good, but if you let your Novag or Fidelity run all night it might find 36 Ra1.|
< Earlier Kibitzing · PAGE 3 OF 3 ·