|
< Earlier Kibitzing · PAGE 4 OF 4 ·
Later Kibitzing> |
Mar-06-10
 | | OhioChessFan: <patzer2: However, I wonder if the relatively quick find (23-ply on Fritz compared to 28-ply on Hiracs or 26-ply on Rybka) this time was influenced by having forced 52. g4!! a number of times previously. > If you didn't clear the hash tables, it might have. One other weak point of Fritz is an amazing propensity to make Rook moves along the 1st/8th rank. |
|
Mar-06-10
 | | OBIT: <agb2002>In your B1 line 52. g4 hg 53. h5 gh 54. Kxf5 Kg7 55. Kg5 Bc1: On page 1 of this thread, I suggest 56. Rd7+ Kg8 (or 56...Kf8 57. Rd1 Be3 58. Kxh5 Bxf4 59. Rf1 wins) 57. Kxh5! Bxf4 58. Kg6! Kf8 (necessary to stop Rd8#) 59. Rf7+ and wins. |
|
Mar-06-10
| | OhioChessFan: <scormus: From this experience and also from what others have said, endings which require a strategy (rather than technique) separate the best from the rest.> I might be splitting semantic hairs, but I think of it more as endings that require a plan and not just a series of tactical shots. Computers are notoriously poor in that, and cover it up in large measure by end game databases. |
|
| Mar-06-10 | | agb2002: <johnlspouge:
....
That said, one has to grant the machines a thoroughness that we humans (<agb2002>'s efforts included :) cannot match.>It would be very inconsiderate on our part to ruin the software industry ;-) . |
|
| Mar-06-10 | | kb2ct: The position is a draw, but not solveable by chess engine.
A humam would amswer Pg4 with fxg4 ensuring the strong side does not get an f-pawn passer but an h-pawn passer. I believe this to be a draw and confirable by tableebase Opposition prevents progress.
Just one of those positions where human knowledge outpreforms engine calculation. :0) |
|
| Mar-06-10 | | tacticalmonster: For those of you who did not solve the puzzle, I will share my method for anyone interested in chess improvement. This is an endgame position. One of the key skill in ending is the creation of " fantasy position". For those of you who read "how to reassesss your chess", You know what I am talking about. When I first look at the position, I did not calculate any variation at all. I just try to come up with different won position for white. At first, I come up with three winning tries: a) sac the rook for three pawns
b) get the bishop off the a1-h8 diagonal so white king can play Kf6-Kxg6 c) sac the exchange on e5 to go into winning king and pawn ending. I have to verify my ideas with calculation. a) 1 Rd5 Bc3 2 Rxf5 gxf5 3 Kxf5 Be1=
b) black bishop will always at least two squares in the diagonal, the rook cannot force the bishop off c) 1 Rd5 Bc3 2 Re5 Kg7! Black does not need to capture the rook. In this position, black can just shuffle his bishop back and forth. I was a little frustrated but I did not give up. After five min, I came up with forth ideas that most of the kibitzer missed. d) trade all the pawns to go into pure R vs B because the black king is in the wrong corner! To win that ending I need the pieces in this configuration: WK(g6), BK(h8), WR ( seventh rank) and BB (a1-h8 diagonal). The key idea is to get the WK on g6! But in the positon, It is almost impossible to do so. So I finally come up with the idea of g4!!. At first, I did not have the idea of creating passer. My fantasy position involving sac all the white pawns just to get my king on g6. Unfortunately, it does not work because I cannot force the black king to h8 because it can escape via f8. One idea leads to another. Since the king and rook alone cannot force the BK to h8 so I reasoned I need a pawn to help me out. This conclusion leads me to the right path e) breakthrough pawn sac to create pass pawn. The rest is just a matter of analysis. This motif is similar to the concept " attack your opponent's strong point" in the " attacking manual" by Jacob Aagaard. It all makes sense. The weakest point in black camp is the g6 pawn. Right now it can only be attacked once by WR and defended once by BK. The idea of g4 is to free the h or f pawn with f4-f5 or h4-h5. It also looses up the control of f5 so white king can get to g6 via f5. I explain my method to the best of my knowledge. I hope anyone can benefit from it. |
|
| Mar-06-10 | | agb2002: <OBIT: <agb2002>In your B1 line 52. g4 hg 53. h5 gh 54. Kxf5 Kg7 55. Kg5 Bc1: On page 1 of this thread, I suggest 56. Rd7+ Kg8 (or 56...Kf8 57. Rd1 Be3 58. Kxh5 Bxf4 59. Rf1 wins) 57. Kxh5! Bxf4 58. Kg6! Kf8 (necessary to stop Rd8#) 59. Rf7+ and wins.> Thank you. I saw a number of problematic lines involving ... g3 and bishop maneuvers to reach f2, even trapping the rook. In your line, Black can play 57... g3 instead of 57... Bxf4. |
|
| Mar-06-10 | | patzer2: <kb2c5t> Can you provide your best play line leading to a table base draw? Until proven otherwise, I'm convinced 52. g4!! fxg4 is a win for White. |
|
| Mar-06-10 | | wordfunph: i took a glance at 52.g4 but didn't mind it...twas a sorry glance afterall. i didn't get the key move huh! |
|
| Mar-06-10 | | kb2ct: 
click for larger viewI was wrong. I have the complete set of tablebases and this position is a win in 33 I would however play for it.
:0) |
|
Mar-06-10
 | | Tabanus: 52.g4 <fxg4> 52.f5 gxf5 53.Kxf5 Kg7 54.Kg5 Be5 55.Kxh5 g3
click for larger viewTablebase win, also with king on e7, e8, f7, f8, g8, h7, and h8. Draw with king on f6, but that's impossible. Else there is no way for black to save the pawn after fxg4. |
|
Mar-06-10
| | OBIT: <agb2002>Hmm... after 52. g4 hg 53. h5 gh 54. Kxf5 Kg7 55. Kg5 Bc1 56. Rd7+ (btw, <tacticalmonster> suggests 56. Rd1 here. I didn't look at it thoroughly, but it also seems to work) Kg8 57. Kxh5 g3!? is interesting. How about 58. Kg6! Kf8 (if 58...Ba3 59. Rd8+ Bf8 60. Rd3) 59. f5! g2 60. Rd8+ Ke7 61. Rd1 Be3 62. f6+ Ke6 (or 62...Ke8 63. f7+ Ke7 64. Kg7) 63. f7 Bc5 64. Rg1! Bxg1 65. f8=Q Bd4 66. Qf5+ Kd6 67. Qf4+ Kc5 68. Qg5+, finally picking off the g-pawn. * * * * * *
Well, <tacticalmonster's> idea 56. Rd1 Be3 57. Re1 Bd2 58. Re2 Bc1 59. Rg2 may not actually win. At least, I don't see the win after 59...Kh7. |
|
| Mar-06-10 | | johnlspouge: < <agb2002> wrote: <johnlspouge:
....
That said, one has to grant the machines a thoroughness that we humans (<agb2002>'s efforts included :) cannot match.>It would be very inconsiderate on our part to ruin the software industry ;-) > I have no doubt that if anyone could do it... ;>) |
|
| Mar-06-10 | | tacticalmonster: <OBIT> white wins after 59 Kh7 60 Rh2 Kg7 (60 g3 61 Rxh5+ Kg7 62 Rh3 g2 63 Rg3  )61 Rxh5 g3 62 Kf5 g2 63 Rg5+ |
|
| Mar-06-10 | | gofer: Posting late today, v busy decorating. Well I eventually looked at Pg3-g4, but only because it was "pawn-storm" week... 52 g4
52 ... hxg4
53 h5 gxh5
54 Kxf4 and now black realises that he is going to lose the remain two pawns and there is nothing
he can do about it! So this will leave white with a pawn up and the R v B exchange up, which is enough
for a win...
52 ... fxg4
53 f5 gxf5
54 Kxf4 again we have exactly the same scenario, but this time white's remaining pawn is on h4 and not f4.
The outcome is the same either way...
The next move is important. If Black doesn't pay Kh7 the white will play Kg6 (threatening mate) and from there
take Ph5. So Kh7 is forced, but then white just plays Kg5 and again white is winning... Time to check... |
|
Mar-06-10
| | Tabanus: 52.g4 hxg4 53.h5 gxh5 54.Kxf5 Kg7 55.Kg5 Bc1
56.Rc3! (engine based)
56...Bd2 57.Rc7+ Kf8 (57...Kg8 58.Kg6 Kf8 59.f5 wins) 58.Kf6 Ke8/g8 f5 wins 56...Bb2 57.Rc7+ Kf8 (57...Kg8 58.Kxh5 wins) 58.Rc2 Ba3 (58...Bd4 59.f5 wins) 59.Kxh5 wins |
|
Mar-06-10
 | | Jimfromprovidence: <Tabanus> and <kb2ct>, after 52.g4 <fxg4> 52.f5 gxf5 53.Kxf5 Kg7 54.Kg5 Be5 55.Kxh5 g3, indicate that the position below is a tablebase win.
click for larger view
The follow-through is amazing and leads to this position after 22 ply.
click for larger viewI am not at all familiar with this type of ending but I think I get it. White has the black king trapped. White now needs to force the bishop between the rook and the h file so the bishop has to retreat to the h file to protect against mate, as in the below position. 
click for larger viewAmazing. A rook vs. bishop ending that is a forced win. Thank you Mr Nalimov. |
|
| Mar-06-10 | | kb2ct: <Tabanus>
Wrong color bishop for the dark colored corner.
I have to admit that I didn't see how to win with it. :0)
+ |
|
| Mar-06-10 | | amathus: An easy task for IM Ornithopoulos. Unforgotten days in Nikaia and Akadhmias , 72. I solved it thanks to my memories from that time, late sixties, early seventies |
|
| Mar-06-10 | | dancindazed: Can somone post the best moves endgame after 52.g4 <fxg4> 52.f5 gxf5 53.Kxf5 Kg7 54.Kg5 Be5 55.Kxh5 g3
...? because I was looking at the shredder endgame tablebase, trying to follow the conversation, and it says next move 56. Kg5 is a win in 31. Then once the move is played it inexplicably says g2 from black is a draw! The other winning 56th moves from white yield the same result. Once played, it says there's a drawing move from black. Something's wrong with shredder's online table. Can anyone explain? |
|
| Mar-06-10 | | dancindazed: Follow up: I'll leave my previous post because I still find it interesting what shredder's tablebase was doing but it was not giving Rd1 as an option and that's where the tablebase was messing it up. It was also giving moves that are not only impossible given the position, but impossible in the game of chess. Such as (pawn)c1-d1(???). So the tablebase was just glitching out. http://www.shredderchess.com/online... Go ahead and see if it does it to you in that position. |
|
| Mar-07-10 | | zanshin: <patzer2: <zanshin> I thought this article about Chuck Amato might be of interest> Thanks for the link. I had suspected something like that. I think he'll be fine. |
|
| Mar-07-10 | | zanshin: <scormus: Finally, interesting discussion about what computer programs are not so good at.> I have been interested in this for some time. Check out my forum header for some examples. |
|
Mar-07-10
| | OBIT: <dancindazed>When I set up the position after 52.g4 fxg4 52.f5 gxf5 53.Kxf5 Kg7 54.Kg5 Be5 55.Kxh5 g3 on shredderchess.com, 56. Kg5 comes up as a draw. The two winning moves are 56. Rd2 and 56. Rd1. Trying some lines with the tablebase, it appears Black can't hold the position because the bishop gets overworked. The bishop doesn't have enough squares on the b8-f4 diagonal to maintain contact with the g-pawn, and if the Black king gets pushed to the 8th rank, White can create threats based on back rank mates. For example, here's an interesting continuation I found playing around with Nalimov: 56. Rd2 Bf4 57. Re2 Bb8 58. Kg5 Bd6 59. Re6 (This forces the bishop off the b8-f4 diagonal, but he can still protect the pawn from e1 or f2.) Bc5 60. Rg6+ Kh7 61. Kh5! (This move is essential to make progress. If the king goes to f5, the position is drawn -- the White must be on the h-pawn to help the pawn get through. After 61. Kf5? he can't get to the h-file, because if he tries to cross the g-file Black can play ...g2.) Bf2 62. Rg4 Be1 63. Re4 Bf2 64. Re7+ Kg8 65. Kg6 Bc5 66. Re2 Bb4 67. Rg2 Be1 68. h5 Bf2 69. h6. 
click for larger viewBlack is in zugzwang! With White to move, this position is drawn, but Black to move loses. For example, 69...Kh8 70. h7! is a mate in four. |
|
| Mar-08-10 | | malvar: Uff!! |
|
 |
|
|
< Earlier Kibitzing · PAGE 4 OF 4 ·
Later Kibitzing> |
Kibitzing
