hard endgame test position

[BBauer]

Yes, finding h4 in 5 min on weak hardware is very good.
If a program finds h4 chances are high that it will win this position.
Congratulations to Christophe.

regards
Bernhard

[Stan Arts]

As we say in Holland, more luck then wisdom, but currently Neurosis "finds" this in about a minute on a P4-3.4 GHz. 64MB hash.

Code: Select all

Legal moves: 22   Static score: 27
Ply Score  Time     Nodes      Best move and expected line
------------------------------------------------------------
2   65     0        290        f6 Bc3
3   65     0        385        f6 Bc3 fxe5
4   68     0        1993       f6 h4 fxe5 dxe5
5   63     2        4884       f6 h4 Kg6 hxg5
6   62     3        14526      f6 h4 Kg6 hxg5 fxg5 Bb4
7   71     13       40109      f6 h4 Kg6 hxg5 fxg5 Ke2 f3+ Kd3
8   71     30       101235     f6 h4 Kg6 hxg5 fxg5 Bb4 f3 Ke3
9   72     81       297026     f6 h4 Kg6 hxg5 fxg5 Bc3 f3 Ke3
10  75     256      959736     f6 h4 Kg6 exf6 Kxf6 Kf1 Kf5 hxg5 Kxg5 h4 Ke2 Kd3
11  78     683      2624202    f6 h4 Kg6 exf6 Kxf6 Kg1 Be4 Kh2 Bc6 hxg5+ Kxg5 h4
12  79     1819     7170510    f6 h4 Kg6 exf6 Kxf6 Kf1 Kf5 hxg5 Kxg5 Ke2 h4 Be1 Kg4 Ba5 Kd3
13  81     4889     19684617   f6 h4 Kg6 exf6 Kxf6 Bc1 Kf5 hxg5 Kxg5 Ba3 h4 Ke2 Kf5 Kd3
13  82     5989     24289732   h4 < + > Bb4 f6 Be7 fxe5 dxe5 Kg6 Bc5 f3 Bd6 Be6 Bc5 Bxh3
13  82     6420     26029990   h4 Bb4 f6 Be7 fxe5 dxe5 Kg6 Bc5 f3 Bd6 Be6 Bc5 Bxh3 Kxf3
14  83     12381    50547298   h4 Bb4 f5 Kg1 Kg6 Bd2 f3 Be1 Kh5 Bb4 g4 Kf2 g3+ Ke3 f2
15  87     31052    127512680  h4 Bb4 f6 Be7 fxe5 dxe5 Kg6 Bc5 Be6 Kg2 f3+ Kh2 Kf5 Bd4 Kh1
16  101    81158    333772027  h4 Bb4 f5 exf6ep+ Kxf6 Kg1 Be6 Kh2 f3 Be1 Kf5 Bd2 g4 hxg4+ Kxg4 Kg1 h3
Total nodes:497250778 n/sec:415700  (q-nodes:48% max depth:29) 

Stan

[mclane]

• m0 d 2, 0 s, +2.73, g7g6 d2c1 g6f5
  m0 d 3, 0 s, +2.73, g7g6 d2c1 g6f5
  m0 d 4, 0 s, +2.88, g7g6 d2e1 g5g4 h3g4 h5g4
  m0 d 5, 0 s, +2.28, g7g6 h3h4 g6f5 h4g5 h5h4 d2c1 h4h3
  m0 d 6, 0 s, +2.28, g7g6 h3h4 g6f5 h4g5 h5h4 d2c1 f5g5
  m0 d 6, 0 s, +2.41, d5e4 d2c1 e4f5 f2g2 g7g6 c1d2 f5e6 d2c1 g6f5
  m0 d 6, 0 s, +2.44, d5c6 h3h4 f7f6 e5f6 g7f6 h4g5 f6g5 d2c1 h5h4
  m0 d 6, 0 s, +2.48, f7f6 h3h4 d5e4 e5f6 g7f6 h4g5 f6g5 d2c1 h5h4
  m0 d 7, 0 s, +2.24, f7f6 h3h4 f6e5 d4e5 g7g6 e5e6 d5e6 h4g5
  m0 d 8, 0 s,>+2.62, f7f6 h3h4
  m0 d 8, 0 s, +2.69, f7f6 e5f6 g7f6 h3h4 f6f5 h4g5 h5h4 d2c1 f5g5
  m0 d 9, 2 s, +2.11, f7f6 h3h4 f6e5 d4e5 g5h4 d2f4 g7g6 f2e3 g6f5
  m0 d 9, 2 s, +2.54, g7g6 h3h4 g6f5 h4g5 h5h4 d2b4 f5g4 b4e7 h4h3
  m0 d 10, 3 s, +2.56, g7g6 h3h4 g6f5 h4g5 h5h4 d2c1 f5g4 f2g1 g4g5
  m0 d 10, 5 s, +2.83, h5h4 d2c3 g7g6 c3e1 g6f5 f2g1 f5e4 e1f2 d5e6 g1g2 e6d7
  m0 d 11, 8 s, +2.83, h5h4 d2e1 g7g6 f2g1 g6f5 g1h2 f5e4 e1f2 d5e6 h2g2 e6d7
  m0 d 12, 14 s, +2.85, h5h4 d2e1 g7g6 f2g1 g6f5 e1f2 g5g4 h3g4 f5g5 g1h2 g5g4 h2g1 h4h3 g1h2
  m0 d 13, 38 s, +2.83, h5h4 d2b4 g7g6 f2g1 g5g4 h3g4 g6g5 b4d2 d5f3 d2e1 f3d5 e1f2 g5g4 g1h2 h4h3
  m0 d 14, 80 s, +2.83, h5h4 d2b4 g7g6 f2g1 g5g4 h3g4 g6g5 b4d2 d5f3 d2e1 f3d5 e1f2 g5g4 g1h2 h4h3

5 seconds on an AMD Sempron 1500 mhz

[playjunior]

Glauraung 1.2.1 SMP found ...h4 with -/+ 1.00(around) in 1 sec., on my Core2 Duo 2.16

[Murat]

Glauraung 1.2.1 SMP found ...h4 with -/+ 1.00(around) in 1 sec., on my Core2 Duo 2.16

Anto,

Pari yegar mer khentanotzin. (welcome to our board )


Murat:.

[smirobth]

As we say in Holland, more luck then wisdom, but currently Neurosis "finds" this in about a minute on a P4-3.4 GHz. 64MB hash.

Code: Select all

Legal moves: 22   Static score: 27
Ply Score  Time     Nodes      Best move and expected line
------------------------------------------------------------
2   65     0        290        f6 Bc3
3   65     0        385        f6 Bc3 fxe5
4   68     0        1993       f6 h4 fxe5 dxe5
5   63     2        4884       f6 h4 Kg6 hxg5
6   62     3        14526      f6 h4 Kg6 hxg5 fxg5 Bb4
7   71     13       40109      f6 h4 Kg6 hxg5 fxg5 Ke2 f3+ Kd3
8   71     30       101235     f6 h4 Kg6 hxg5 fxg5 Bb4 f3 Ke3
9   72     81       297026     f6 h4 Kg6 hxg5 fxg5 Bc3 f3 Ke3
10  75     256      959736     f6 h4 Kg6 exf6 Kxf6 Kf1 Kf5 hxg5 Kxg5 h4 Ke2 Kd3
11  78     683      2624202    f6 h4 Kg6 exf6 Kxf6 Kg1 Be4 Kh2 Bc6 hxg5+ Kxg5 h4
12  79     1819     7170510    f6 h4 Kg6 exf6 Kxf6 Kf1 Kf5 hxg5 Kxg5 Ke2 h4 Be1 Kg4 Ba5 Kd3
13  81     4889     19684617   f6 h4 Kg6 exf6 Kxf6 Bc1 Kf5 hxg5 Kxg5 Ba3 h4 Ke2 Kf5 Kd3
13  82     5989     24289732   h4 < + > Bb4 f6 Be7 fxe5 dxe5 Kg6 Bc5 f3 Bd6 Be6 Bc5 Bxh3
13  82     6420     26029990   h4 Bb4 f6 Be7 fxe5 dxe5 Kg6 Bc5 f3 Bd6 Be6 Bc5 Bxh3 Kxf3
14  83     12381    50547298   h4 Bb4 f5 Kg1 Kg6 Bd2 f3 Be1 Kh5 Bb4 g4 Kf2 g3+ Ke3 f2
15  87     31052    127512680  h4 Bb4 f6 Be7 fxe5 dxe5 Kg6 Bc5 Be6 Kg2 f3+ Kh2 Kf5 Bd4 Kh1
16  101    81158    333772027  h4 Bb4 f5 exf6ep+ Kxf6 Kg1 Be6 Kh2 f3 Be1 Kf5 Bd2 g4 hxg4+ Kxg4 Kg1 h3
Total nodes:497250778 n/sec:415700  (q-nodes:48% max depth:29) 

Stan

Yes, luck indeed, since the follow up moves it gives (f6 or f5) I believe lead to only a draw.

[smirobth]

Glauraung 1.2.1 SMP found ...h4 with -/+ 1.00(around) in 1 sec., on my Core2 Duo 2.16

What were the follow up moves?

[smirobth]

5 seconds on an AMD Sempron 1500 mhz

WOW!

[tiger]

Yes, finding h4 in 5 min on weak hardware is very good.
If a program finds h4 chances are high that it will win this position.
Congratulations to Christophe.

regards
Bernhard

Thanks. I guess that on modern hardware with more hash tables (I used only 48Mb) it would find it in 1 minute. So it could be able to actually play it in a game.


// Christophe

[playjunior]

To Murat:
Shnorhakal em! (Thank you!)


To Robin:
Here is analysis by Glaurung 1.2.1 SMP

New game
8/5pk1/8/3bP1pp/3P1p2/7P/3B1K2/8 b - - 0 1

Analysis by Glaurung 1.2.1 SMP:

1...Kg6 2.Bb4
-/+ (-0.79) Depth: 2 00:00:00
1...f6 2.Bb4 fxe5 3.dxe5
-/+ (-0.89) Depth: 2 00:00:00
1...f6 2.Bb4 fxe5 3.dxe5
-/+ (-0.89) Depth: 3 00:00:00
1...f6 2.exf6+ Kxf6 3.h4 Kf5 4.hxg5 Kxg5
-/+ (-0.89) Depth: 4 00:00:00
1...f6 2.exf6+ Kxf6 3.h4 Kf5 4.hxg5 Kxg5
-/+ (-0.89) Depth: 5 00:00:00
1...f6 2.h4 Kf7 3.exf6 Kxf6 4.hxg5+ Kxg5
-/+ (-0.89) Depth: 6 00:00:00 4kN
1...f6 2.h4 Kf7 3.exf6 Kxf6 4.hxg5+ Kxg5 5.Bb4
-/+ (-0.84) Depth: 7 00:00:00 6kN
1...f6 2.h4 Kf7 3.exf6 Kxf6 4.hxg5+ Kxg5 5.Bb4 Kf5
-/+ (-0.89) Depth: 8 00:00:00 9kN
1...f6 2.h4 Kf7 3.exf6 Kxf6 4.hxg5+ Kxg5 5.Bb4 Kf5 6.Bc5
-/+ (-0.89) Depth: 9 00:00:00 13kN
1...f6 2.h4 Kf7 3.exf6 Kxf6 4.hxg5+ Kxg5 5.Bb4 Kf5 6.Bc5 h4
-/+ (-0.92) Depth: 10 00:00:00 28kN
1...f6 2.h4 Kf7 3.exf6 Kxf6 4.hxg5+ Kxg5 5.Bb4 Kf5 6.Bd6 Ke4 7.Be5
-/+ (-0.92) Depth: 11 00:00:00 45kN
1...f6 2.h4 Kf7 3.exf6 Kxf6 4.Kg1 Kf5 5.hxg5 Kxg5 6.Bb4 Kg4 7.Kf2
-/+ (-0.90) Depth: 12 00:00:00 83kN
1...h4 2.Bb4 f6 3.Be7 fxe5 4.Bxg5 Kg6 5.Bxh4 e4 6.Be7 Kf5 7.Bc5
-/+ (-1.06) Depth: 12 00:00:00 151kN
1...h4 2.Bb4 f6 3.Be7 fxe5 4.Bxg5 Kg6 5.Bxh4 e4 6.Be7 Kf5 7.Bc5 Bc4
-/+ (-0.98) Depth: 13 00:00:00 204kN
1...h4 2.Bb4 f6 3.Be7 fxe5 4.Bxg5 Kg6 5.Bxh4 exd4 6.Ke2 f3+ 7.Kf2 d3 8.Ke3 Kh5
-/+ (-1.00) Depth: 14 00:00:00 394kN
1...h4 2.Bb4 f6 3.Be7 fxe5 4.dxe5 Kg6 5.Kg1 Kf5 6.Kf2 Bf7 7.Kf3 Be6 8.Ke2 Bd5
-/+ (-1.03) Depth: 15 00:00:01 670kN
1...h4 2.Ba5 f5 3.exf6+ Kxf6 4.Bd8+ Kf5 5.Kg1 Be4 6.Be7 Bf3 7.Kf1 Bd5 8.Kg1 Bb3 9.Kf2 Bc4 10.Bc5
-/+ (-1.03) Depth: 16 00:00:02 1943kN
1...h4 2.Ba5 f5 3.exf6+ Kxf6 4.Bd8+ Kf5 5.Kg1 Be4 6.Be7 Bf3 7.Kf1 Bd5 8.Kg1 Bb3 9.Kf2 Bc4 10.Bd8 Be6
-/+ (-1.03) Depth: 17 00:00:02 2654kN

(, 19.08.2007)
[/quote]