What's up with this endgame?

Discussion of anything and everything relating to chess playing software and machines.

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Paloma
Posts: 848
Joined: Thu Dec 25, 2008 8:07 pm

Re: What's up with this endgame?

Post by Paloma » Sat Jul 20, 2019 7:11 pm

Thanks

zullil
Posts: 5668
Joined: Mon Jan 08, 2007 11:31 pm
Location: PA USA
Full name: Louis Zulli

Re: What's up with this endgame?

Post by zullil » Sat Jul 20, 2019 7:21 pm

ChickenLogic wrote:
Sat Jul 20, 2019 6:17 pm


At depth 100/187 SF dev finally shows +2.96 with analysis contempt on (c=24 so real eval is probably around +3.2 is my guess). So this might just be mate in N from this position.
I doubt that this position is a win for White. Black should simply avoid gxf4 and not allow the white king to penetrate across the c-file.

But, who knows.

User avatar
Ajedrecista
Posts: 1399
Joined: Wed Jul 13, 2011 7:04 pm
Location: Madrid, Spain.
Contact:

Re: What's up with this endgame?

Post by Ajedrecista » Sat Jul 20, 2019 7:48 pm

Hello:
zullil wrote:
Sat Jul 20, 2019 7:21 pm
ChickenLogic wrote:
Sat Jul 20, 2019 6:17 pm


At depth 100/187 SF dev finally shows +2.96 with analysis contempt on (c=24 so real eval is probably around +3.2 is my guess). So this might just be mate in N from this position.
I doubt that this position is a win for White. Black should simply avoid gxf4 and not allow the white king to penetrate across the c-file.

But, who knows.
Could this position be solved by Freezer and/or FinalGen? I think it is a suitable position: 1 minor piece vs 1 minor piece and some blocked pawns... however, the elapsed time to get the solution must be huge, as well as RAM and storage space available, because there are 7 + 7 = 14 pieces OTB.

To get an idea of the size of the problem, there are 5,748,050 unique positions after only 8 plies, counted with JetChess [not to be confused with perft(8) of this position, which is 1,406,028,571].

Regards from Spain.

Ajedrecista.

zullil
Posts: 5668
Joined: Mon Jan 08, 2007 11:31 pm
Location: PA USA
Full name: Louis Zulli

Re: What's up with this endgame?

Post by zullil » Sat Jul 20, 2019 9:45 pm

zullil wrote:
Sat Jul 20, 2019 7:21 pm
ChickenLogic wrote:
Sat Jul 20, 2019 6:17 pm


At depth 100/187 SF dev finally shows +2.96 with analysis contempt on (c=24 so real eval is probably around +3.2 is my guess). So this might just be mate in N from this position.
I doubt that this position is a win for White. Black should simply avoid gxf4 and not allow the white king to penetrate across the c-file.

But, who knows.
Well, some forward and backward analysis suggests that I was wrong. :D Cfish-dev says:

+3.84 40... Kf7 41. Ke1 Ke7 42. Kd2 Be8 43. Kc3 Kd6 44. Bd3 Ba4 45. Kb4 Bd1 46. a4 Kc6 47. Bb5+ Kb7 48. Bf1 Bh5 49. g3 g5 50. Bd3 Bg6 51. a5 Ka7 52. Bb5 Bh5 53. Bd7 Kb7 54. Bxf5 Bf3 55. Bc2 Bg2 56. Bd3 bxa5+ 57. Kxa5 Kc6 58. Bc2 Bf1 59. Ba4+ Kc7 60. Bb5 Bh3 61. Be2 Kc6 62. Ba6 Bg4 63. Bd3 Bh5 64. Bb5+ Kc7 65. Ba6 Bd1 66. Bd3 Kc6 67. Bb5+ Kb7 68. Bf1 Kc6 69. Bh3 Be2 70. Bf5 Bc4 71. Be6 Be2 72. Bf7 Bd3 73. Be8+ Kc7 74. Bh5 Kb7 75. Bd1 Kc6 76. Bg4 Be4 77. Be6 Bc2 78. Ka6 Bg6 79. Ka7 Kd6 80. Bg4 Kc7 81. Bf3 Kc6 82. Kb8 Bf5 83. Bh5 Kb6 84. Be8 Bh3 85. Bf7 gxf4 86. gxf4 Kc6 87. Be8+ Kb6 88. Bg6 Bd7 89. Bc2 Be6 90. Bb3 Kc6 91. Ba4+ Kd6 (depth 87, 0:30:47)

Uri Blass
Posts: 8590
Joined: Wed Mar 08, 2006 11:37 pm
Location: Tel-Aviv Israel

Re: What's up with this endgame?

Post by Uri Blass » Sat Jul 20, 2019 11:00 pm

I wonder if the following position that happened in the game is also a win for white or maybe in the game white did a mistake and in the following position cannot win in case black play the right move.


Vinvin
Posts: 4378
Joined: Thu Mar 09, 2006 8:40 am
Full name: Vincent Lejeune

Re: What's up with this endgame?

Post by Vinvin » Sat Jul 20, 2019 11:02 pm

zullil wrote:
Sat Jul 20, 2019 9:45 pm
zullil wrote:
Sat Jul 20, 2019 7:21 pm
ChickenLogic wrote:
Sat Jul 20, 2019 6:17 pm


At depth 100/187 SF dev finally shows +2.96 with analysis contempt on (c=24 so real eval is probably around +3.2 is my guess). So this might just be mate in N from this position.
I doubt that this position is a win for White. Black should simply avoid gxf4 and not allow the white king to penetrate across the c-file.

But, who knows.
Well, some forward and backward analysis suggests that I was wrong. :D Cfish-dev says:

+3.84 40... Kf7 41. Ke1 Ke7 42. Kd2 Be8 43. Kc3 Kd6 44. Bd3 Ba4 45. Kb4 Bd1 46. a4 Kc6 47. Bb5+ Kb7 48. Bf1 Bh5 49. g3 g5 50. Bd3 Bg6 51. a5 Ka7 52. Bb5 Bh5 53. Bd7 Kb7 54. Bxf5 Bf3 55. Bc2 Bg2 56. Bd3 bxa5+ 57. Kxa5 Kc6 58. Bc2 Bf1 59. Ba4+ Kc7 60. Bb5 Bh3 61. Be2 Kc6 62. Ba6 Bg4 63. Bd3 Bh5 64. Bb5+ Kc7 65. Ba6 Bd1 66. Bd3 Kc6 67. Bb5+ Kb7 68. Bf1 Kc6 69. Bh3 Be2 70. Bf5 Bc4 71. Be6 Be2 72. Bf7 Bd3 73. Be8+ Kc7 74. Bh5 Kb7 75. Bd1 Kc6 76. Bg4 Be4 77. Be6 Bc2 78. Ka6 Bg6 79. Ka7 Kd6 80. Bg4 Kc7 81. Bf3 Kc6 82. Kb8 Bf5 83. Bh5 Kb6 84. Be8 Bh3 85. Bf7 gxf4 86. gxf4 Kc6 87. Be8+ Kb6 88. Bg6 Bd7 89. Bc2 Be6 90. Bb3 Kc6 91. Ba4+ Kd6 (depth 87, 0:30:47)
Please, could you check the eval by playing the main line move by move ?
I have some doubts with long lines in Stockfish. They are sometimes wrong.
In the line displayed here 77...Bc2 is losing but 77...Bd3 seems drawish.

User avatar
Ovyron
Posts: 2720
Joined: Tue Jul 03, 2007 2:30 am

Re: What's up with this endgame?

Post by Ovyron » Sun Jul 21, 2019 3:08 am

Wow! Very impressive developments on the analysis. If that was lost then I wonder what was black's losing move, but it seems Stockfish was giving drawish score for mate in N position.
Uri Blass wrote:
Sat Jul 20, 2019 11:00 pm
I wonder if the following position that happened in the game is also a win for white or maybe in the game white did a mistake and in the following position cannot win in case black play the right move.
This is doubtful. My opponent was using some weeks for the moves, in that time he played what he believed was winning, and he was right, so if the position in the OP is a mate in something, my opponent picked it because he figured that out already, so I don't think he messed up later on.

I still have private Stockfish 9 with learning around, now that I know the depths required to see the light it's possible I can solve it and answer these questions.
Great spirits have always encountered violent opposition from mediocre minds.

User avatar
Eelco de Groot
Posts: 4161
Joined: Sun Mar 12, 2006 1:40 am
Location: Groningen

Re: What's up with this endgame?

Post by Eelco de Groot » Sun Jul 21, 2019 7:32 am

Vinvin wrote:
Sat Jul 20, 2019 11:02 pm
zullil wrote:
Sat Jul 20, 2019 9:45 pm
zullil wrote:
Sat Jul 20, 2019 7:21 pm
ChickenLogic wrote:
Sat Jul 20, 2019 6:17 pm


At depth 100/187 SF dev finally shows +2.96 with analysis contempt on (c=24 so real eval is probably around +3.2 is my guess). So this might just be mate in N from this position.
I doubt that this position is a win for White. Black should simply avoid gxf4 and not allow the white king to penetrate across the c-file.

But, who knows.
Well, some forward and backward analysis suggests that I was wrong. :D Cfish-dev says:

+3.84 40... Kf7 41. Ke1 Ke7 42. Kd2 Be8 43. Kc3 Kd6 44. Bd3 Ba4 45. Kb4 Bd1 46. a4 Kc6 47. Bb5+ Kb7 48. Bf1 Bh5 49. g3 g5 50. Bd3 Bg6 51. a5 Ka7 52. Bb5 Bh5 53. Bd7 Kb7 54. Bxf5 Bf3 55. Bc2 Bg2 56. Bd3 bxa5+ 57. Kxa5 Kc6 58. Bc2 Bf1 59. Ba4+ Kc7 60. Bb5 Bh3 61. Be2 Kc6 62. Ba6 Bg4 63. Bd3 Bh5 64. Bb5+ Kc7 65. Ba6 Bd1 66. Bd3 Kc6 67. Bb5+ Kb7 68. Bf1 Kc6 69. Bh3 Be2 70. Bf5 Bc4 71. Be6 Be2 72. Bf7 Bd3 73. Be8+ Kc7 74. Bh5 Kb7 75. Bd1 Kc6 76. Bg4 Be4 77. Be6 Bc2 78. Ka6 Bg6 79. Ka7 Kd6 80. Bg4 Kc7 81. Bf3 Kc6 82. Kb8 Bf5 83. Bh5 Kb6 84. Be8 Bh3 85. Bf7 gxf4 86. gxf4 Kc6 87. Be8+ Kb6 88. Bg6 Bd7 89. Bc2 Be6 90. Bb3 Kc6 91. Ba4+ Kd6 (depth 87, 0:30:47)
Please, could you check the eval by playing the main line move by move ?
I have some doubts with long lines in Stockfish. They are sometimes wrong.
In the line displayed here 77...Bc2 is losing but 77...Bd3 seems drawish.
I can confirm that backwards solving the line from CFish seems to get stuck on move 77 and Bd3 so far is the only move that does not lose. But no tablebases etc. used. 1.23 I consider a draw. 77...Bc2 would lose.



Engine: Kaissa IV NoContempt (512 MB)
by T. Romstad, M. Costalba, J. Kiiski, G. Linscott

59 10:53 +1.23 77...Bd3 78.Kb4 Be4 79.Bg4 Bd3
80.Bh3 Be2 81.Be6 Bf3 82.Bf5 Be2
83.Kc3 Kd6 84.Bd3 Bf3 85.Kd2 Bg4
86.Bb1 Bd7 87.Bc2 Bg4 88.Bd3 Bc8
89.Kc3 Bg4 90.Bb5 Bf5 (1.839.438.988) 2814

59 10:53 +6.01 77...Bg6 78.Ka6 Be4 79.Bf7 Bd3+
80.Ka7 Bf5 81.Kb8 Kd6 82.Kb7 Bg4
83.Kb6 Bf5 84.Bh5 Bc8 85.Bd1 Bf5
86.Kb7 Bh3 87.Be2 Kd7 88.Bb5+ Kd6
89.Ba6 Bd7 90.Bd3 Bg4 (1.839.438.988) 2814

59 10:53 +6.01 77...Bc2 78.Ka6 Bg6 79.Ka7 Kd6
80.Bg4 Kc7 81.Bf3 Kc6 82.Kb8 Bf5
83.Bh5 Kb6 84.Be8 gxf4 85.gxf4 Bg4
86.Bg6 Bd7 87.Bd3 Kc6 88.Ba6 Bh3
89.Bb7+ Kd6 90.Bc8 Bf1 (1.839.438.988) 2814

(At this point it is 20 moves without a non reversible move says Shredder GUI but I don't think that is very important yet.) I could post the .pgn of the line but an old Shredder GUI bug then stops analysis :mrgreen:

Here is the pgn:



Debugging is twice as hard as writing the code in the first
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
-- Brian W. Kernighan

zullil
Posts: 5668
Joined: Mon Jan 08, 2007 11:31 pm
Location: PA USA
Full name: Louis Zulli

Re: What's up with this endgame?

Post by zullil » Sun Jul 21, 2019 9:36 am



+7.52 40... g5 41. Ke1 Kf8 42. Kd2 Ke7 43. Kc3 Be8 44. Kb4 Kd8 45. Bb5 Bh5 46. Bc6 Bf7 47. Kb5 Kc7 48. g3 Bg8 49. Be8 Kd8 50. Bg6 Kc7 51. Bh5 Be6 52. Kb4 Bc8 53. Be2 Bd7 54. a4 Be8 55. Bb5 Bg6 56. a5 bxa5+ 57. Kxa5 Bh5 58. Bd3 Kc6 59. Bc2 Be2 60. Bxf5 Bf3 61. Be6 Be2 62. Bg8 Bd3 63. Bf7 Be4 64. Be8+ Kc7 65. Kb5 Kd6 66. Bf7 Bd3+ 67. Kb6 gxf4 68. gxf4 Bf1 69. Bg6 Bh3 70. Kb7 Bg4 71. Bd3 Be6 72. Bb5 Bh3 73. Kb6 Bf5 74. Bc6 Be4 75. Be8 Ke7 76. Bb5 Kd6 77. Bc6 Bh1 78. Be8 Be4 79. Kb7 Bd3 80. Bf7 f5 81. Be8 Bf1 82. Kb6 Be2 83. Bc6 Bd1 84. Bb5 Bc2 85. Kb7 Bd1 86. Kc8 (depth 78, 8:57:10)

Looks like White wins. :wink:

User avatar
Eelco de Groot
Posts: 4161
Joined: Sun Mar 12, 2006 1:40 am
Location: Groningen

Re: What's up with this endgame?

Post by Eelco de Groot » Sun Jul 21, 2019 10:07 am

Now Kaissa seems to get stuck a little here Louis... :o Assuming I entered the line correctly into Uly's pgn so far. Could it be a tablebase thing? But 63...Be4 seems just a wrong move...

Or is it just because you disabled nullmove in CFish? That I can believe, if it is some sort of Zugzwang and Stockfish also would not see it on time..




Engine: Kaissa IV NoContempt (512 MB)
by T. Romstad, M. Costalba, J. Kiiski, G. Linscott

56 3:14 +1.23 63...Bf1 64.Kb4 Be2 65.Bg6 Bf3
66.Be8+ Kb6 67.Ba4 Be4 68.Kc3 Kc7
69.Bd1 Kd6 70.Be2 Kc6 71.Bd3 Bf3
72.Bc2 Kd6 73.Kd2 Bg4 74.Bd3 Bd7
75.Kc3 Bg4 76.Bb1 Kc6 (531.595.387) 2733

56 3:14 +1.23 63...Be2 64.Be8+ Kc7 65.Kb4 Kd6
66.Bg6 Kc6 67.Bb1 Kd6 68.Bf5 Kc6
69.Kc3 Bf3 70.Kd3 Kd6 71.Bg6 Bg2
72.Kd2 Bf3 73.Bf7 Be4 74.Kc3 Kc6
75.Be8+ Kd6 76.Bb5 Bf3 (531.595.387) 2733

55 3:14 +6.49 63...Bb1 64.Ka6 Bh7 65.Ka7 Bd3
66.Kb8 Bf5 67.Bh5 Kb6 68.Be8 Be6
69.Bg6 Kc6 70.Bd3 Bd7 71.Ba6 gxf4
72.gxf4 Bf5 73.Bb7+ Kd6 74.Bc8 Bd3
75.Kb7 Be2 76.Kb6 Bd1 (531.595.387) 2733

Debugging is twice as hard as writing the code in the first
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
-- Brian W. Kernighan

Post Reply