Gull 3.0 Syzygy in GitHub

[Nordlandia]

Gull Syzygy 3.1 is still not capable to solve Behting Study



Gull evaluation is 1. Ng7+ which loses as it transpose into 7-piece endgame.

Of course Gull's next best movie is 1. Kc6.


[d]8/8/7p/3KNN1k/2p4p/8/3P2p1/8 w - - 0 1

Analysis by Gull.3.1_JVx64:

1.Ng7+ Kg5 2.Nf3+ Kg4 3.Ke4 h3 4.Nf5 g1Q 5.Nxh6+ Kh5 6.Nxg1 h2 7.Nf3 h1Q 8.Nf5 Kg4 9.Ne3+ Kg3 10.Nf5+ Kg2 11.N5d4 Kf2 12.Kd5 Kg3 13.Ke4 Kg4 14.Ke3 Qh6+ 15.Ke4 Qh7+ 16.Kd5 Qg8+ 17.Ke4 Qa8+ 18.Ke5 Kg3 19.Kd6 Kf4 20.Kc5 Qf8+ 21.Kc6 Qe8+ 22.Kd5 Qb8 23.Kc6 Ke4 24.Kc5 Qc7+ 25.Kb5 Qd6 26.Kxc4
-+ (-3.17) Depth: 26/57 00:00:10 177MN, tb=232663
(03.07.2016)

[MikeB]

Gull Syzygy 3.1 is still not capable to solve Behting Study


Gull evaluation is 1. Ng7+ which loses as it transpose into 7-piece endgame.

Of course Gull's next best movie is 1. Kc6.


[d]8/8/7p/3KNN1k/2p4p/8/3P2p1/8 w - - 0 1

Analysis by Gull.3.1_JVx64:

1.Ng7+ Kg5 2.Nf3+ Kg4 3.Ke4 h3 4.Nf5 g1Q 5.Nxh6+ Kh5 6.Nxg1 h2 7.Nf3 h1Q 8.Nf5 Kg4 9.Ne3+ Kg3 10.Nf5+ Kg2 11.N5d4 Kf2 12.Kd5 Kg3 13.Ke4 Kg4 14.Ke3 Qh6+ 15.Ke4 Qh7+ 16.Kd5 Qg8+ 17.Ke4 Qa8+ 18.Ke5 Kg3 19.Kd6 Kf4 20.Kc5 Qf8+ 21.Kc6 Qe8+ 22.Kd5 Qb8 23.Kc6 Ke4 24.Kc5 Qc7+ 25.Kb5 Qd6 26.Kxc4
-+ (-3.17) Depth: 26/57 00:00:10 177MN, tb=232663
(03.07.2016)

Hawkeye ( derivative based on Gull 3 syzygy) - sees Kc6 in less than 2 minutes with single core:

Code: Select all

here's the full pv provided at 142 seconds:
info time 142626 depth 25 seldepth 53 score cp -321 nps 2642386 nodes 376873046 tbhits 2370480 pv d5c6 g2g1q f5h4 g1a7 h4f3 a7a5 c6d6 a5b5 d6e6 b5c5 e6f5 c5d5 f5f4 d5b7 f4f5 b7a6 f5e4 a6d6 e4f5 d6c7 f5f4 c7a7 f4e4 a7a4 e4d4 a4b4 d4d5 b4b5 d5d4 b5b3 d4c5 c4c3 d2c3 b3c3 c5d5 c3b3 d5e4 b3b5 f3d4 b5b4 e4e3 b4c3 e3e4 h5g5 d4e6 g5f6

[Nordlandia]

GUI: Arena 3.5.1
Hash: 2048 MB allocated per laptop
Ponder: On
Tablebases: Gull 6Men | SF Dev 4Men
CPU: i5-5200U | 2-Core
[pgn][Event "Auto 232 Game"]
[Site "DESKTOP-DD1B2QU"]
[Date "2016.07.03"]
[Round "1"]
[White "Gull.3.1_JVx64"]
[Black "Stockfish Dev"]
[Result "1/2-1/2"]
[BlackElo "3200"]
[ECO "A05"]
[Opening "Reti Opening"]
[Time "11:47:01"]
[Variation "KIA, 3.Bg2 Bg7 4.O-O O-O"]
[WhiteElo "3200"]
[TimeControl "1800+30"]
[Termination "normal"]
[PlyCount "118"]
[WhiteType "program"]
[BlackType "program"]

1. Nf3 Nf6 2. g3 g6 3. Bg2 Bg7 4. O-O O-O 5. c4 {+0.18/21 67} c5 6. Nc3
{+0.21/22 146} d5 7. cxd5 {+0.23/21 59} Nxd5 8. Nxd5 {+0.25/22 45} Qxd5 9.
d4 {+0.25/22 0} cxd4 10. Be3 {+0.25/22 0} d3 11. Ne1 {+0.22/21 134} Qd6 12.
Qxd3 {+0.28/21 0} Qb4 13. a3 {+0.39/21 72} Qg4 14. h3 {+0.36/21 0} Qa4 15.
b3 {+0.43/22 115} Qe8 16. Rd1 {+0.39/23 106} Nc6 17. Nf3 {+0.33/23 162} b6
18. b4 {+0.35/22 130} Bf5 19. Qa6 {+0.40/22 110} Rc8 20. g4 {+0.40/22 0}
Be4 21. Rc1 {+0.39/22 0} e6 22. Ng5 {+0.25/21 86} Bxg2 23. Kxg2 {+0.20/21
39} Ne7 24. Qxa7 {+0.32/21 36} Nd5 25. Rxc8 {+0.16/22 42} Qxc8 26. Bd2
{+0.16/22 17} Qc4 27. e4 {+0.08/22 17} Qd3 28. Bc1 {+0.08/23 86} f6 29. Nf3
{+0.08/23 0} f5 30. exf5 {+0.06/23 0} exf5 31. g5 {+0.07/23 0} Bc3 32. Qa4
{+0.08/23 0} Rd8 33. Qb3 {+0.08/23 0} Qe2 34. Ng1 {+0.09/23 0} Qd3 35. b5
{+0.09/23 0} Kf8 36. Nf3 {0.00/24 82} Qe2 37. Rh1 {0.00/24 0} Ke8 38. Kg3
{+0.13/22 39} f4+ 39. Kg2 {+0.23/22 0} Kf8 40. Rg1 {+0.19/22 0} Kg7 41. a4
{+0.24/21 33} Kf8 42. Ba3+ {0.00/23 212} Kg7 43. Rc1 {0.00/25 87} Ne3+ 44.
Kg1 {0.00/25 0} Nd1 45. Rc2 {0.00/25 0} Qxf3 46. Bb4 {0.00/25 0} Rd5 47.
Bxc3+ {0.00/25 0} Kf7 48. Qc4 {0.00/25 0} Ne3 49. fxe3 {0.00/25 0} Qd1+ 50.
Qf1 {0.00/25 0} Qxc2 51. Qxf4+ {0.00/25 0} Rf5 52. Qc7+ {0.00/25 0} Ke8 53.
Qc6+ {0.00/25 0} Kf7 54. Qd7+ {0.00/25 8} Kf8 55. Qc8+ {0.00/25 0} Kf7 56.
Qc4+ {0.00/25 0} Ke8 57. Qc8+ {0.00/25 0} Kf7 58. Qc4+ {0.00/25 0} Ke8 59.
Qc8+ {0.00/25 0} Kf7 {3-fold repetition} 1/2-1/2
[/pgn]

[Nordlandia]

GUI: Arena 3.5.1
Hash: 2048 MB allocated per laptop
Ponder: On
Tablebases: Gull 6Men | SF Dev 5Men
CPU: i5-5200U | 2-Core
TC: 30m+30s

[pgn][Event "Auto 232 Game"]
[Site "DESKTOP-DD1B2QU"]
[Date "2016.07.03"]
[Round "1"]
[White "Official SF Dev 02.07.16 POPCNT"]
[Black "Gull.3.1_JVx64"]
[Result "1-0"]
[BlackElo "3200"]
[ECO "A05"]
[Opening "Reti Opening"]
[Time "13:54:55"]
[Variation "KIA, 3.Bg2 Bg7 4.O-O O-O"]
[WhiteElo "2400"]
[TimeControl "1800+30"]
[Termination "normal"]
[PlyCount "167"]
[WhiteType "program"]
[BlackType "program"]

1. Nf3 Nf6 2. g3 g6 3. Bg2 Bg7 4. O-O O-O 5. c4 c6 {-0.16/21 122} 6. b3 d5
{-0.11/21 54} 7. Bb2 Nbd7 {-0.11/21 0} 8. d4 b5 {-0.06/22 119} 9. Nbd2 bxc4
{-0.19/20 98} 10. bxc4 Rb8 {-0.11/21 39} 11. Ba3 Ne4 {+0.11/20 57} 12. e3
Qa5 {0.00/22 158} 13. Qc1 Nd6 {0.00/22 0} 14. cxd5 cxd5 {+0.08/22 0} 15.
Re1 Re8 {+0.09/21 52} 16. Nb3 Qa4 {+0.05/23 195} 17. Bxd6 exd6 {+0.05/22
51} 18. Ng5 h6 {+0.02/21 219} 19. Nh3 Nf6 {+0.02/21 0} 20. Nf4 Bb7 {0.00/21
0} 21. Qb2 Bf8 {+0.01/20 134} 22. Rab1 Be7 {+0.05/22 76} 23. Ne2 Bc6
{+0.08/22 76} 24. Rec1 Kg7 {0.00/22 93} 25. Nc3 Qb4 {-0.07/23 83} 26. Qe2
Rb6 {0.00/23 72} 27. Qd1 Bd8 {0.00/23 82} 28. Rc2 Qa3 {0.00/23 0} 29. Rbc1
Qb4 {-0.02/23 0} 30. h4 Bb7 {-0.08/22 128} 31. Nd2 Qa3 {-0.09/21 76} 32. h5
g5 {-0.23/21 0} 33. g4 Qa5 {-0.27/21 51} 34. Bf3 Bc6 {-0.42/22 20} 35. Nb3
Qb4 {-0.27/22 85} 36. Qf1 Rb8 {-0.40/22 181} 37. Rb1 Be7 {-0.40/22 0} 38.
Be2 Rec8 {-0.56/20 170} 39. Qc1 Re8 {-0.59/21 66} 40. a4 Rec8 {-0.73/18 28}
41. Ba6 Re8 {-1.01/22 63} 42. Bb5 Bxb5 {-1.01/22 0} 43. axb5 Bf8 {-1.07/22
342} 44. f3 Qc4 {-1.40/22 4} 45. Rbb2 Qc8 {-1.39/20 66} 46. Na5 Qe6
{-1.39/20 65} 47. Nc6 Ra8 {-1.46/19 25} 48. Re2 Qc8 {-1.49/20 33} 49. Rb1
Qb7 {-1.52/19 32} 50. Nb4 Rac8 {-1.69/19 29} 51. Qc2 Be7 {-1.46/18 20} 52.
Ra1 Bd8 {-1.70/19 40} 53. Nc6 Bb6 {-1.59/20 30} 54. Qf5 Re6 {-1.75/20 0}
55. Nxd5 Nxd5 {-1.82/21 18} 56. Qxd5 Qd7 {-1.81/21 0} 57. Ra3 Ree8
{-1.77/22 46} 58. Kf2 Rf8 {-1.77/22 45} 59. Qf5 Rc7 {-1.76/22 30} 60. Rea2
Qxf5 {-1.84/21 67} 61. gxf5 Kf6 {-1.95/22 29} 62. Nxa7 Bxa7 {-2.32/20 21}
63. Rxa7 Rxa7 {-2.47/21 8} 64. Rxa7 Rb8 {-2.55/22 24} 65. Ra6 Ke7 {-2.68/21
37} 66. b6 Kd7 {-3.32/21 15} 67. e4 f6 {-3.24/22 39} 68. Ke3 Kc6 {-3.61/19
46} 69. d5+ Kb5 {-4.03/19 0} 70. Ra7 Kxb6 {-3.95/22 29} 71. Rf7 Rh8
{-4.24/22 0} 72. Kd4 Ra8 {-4.59/21 76} 73. Rxf6 Ra4+ {-4.58/21 0} 74. Ke3
Ra3+ {-4.58/21 58} 75. Kf2 Kc5 {-5.25/21 7} 76. Rxh6 Kd4 {-5.22/21 69} 77.
Kg3 Ra8 {-5.54/19 44} 78. f6 Rf8 {-5.38/17 29} 79. Kg4 Rc8 {-7.15/18 30}
80. f7 Rb8 {-11.33/16 30} 81. Rf6 Rf8 {-16.68/15 30} 82. Kxg5 Ke3
{-18.66/17 29} 83. Kg6 Kd4 {-22.54/17 0} 84. Kg7 {Black resigns} 1-0
[/pgn]

[Nordlandia]

GUI: Arena 3.5.1
Hash: 2048 MB allocated per laptop
Ponder: On
Tablebases: Gull 5Men | SF Dev 5Men
CPU: i5-5200U | 2-Core
TC: 60m+30s
Book: None
[pgn][Event "Auto 232 Game"]
[Site "DESKTOP-DD1B2QU"]
[Date "2016.07.03"]
[Round "1"]
[White "Gull.3.1_JVx64"]
[Black "Houdini 4 Pro | Default"]
[Result "1/2-1/2"]
[BlackElo "3200"]
[ECO "C01"]
[Opening "French"]
[Time "20:17:24"]
[Variation "Exchange, 4.Nf3 Bd6 5.Bd3 Nf6"]
[WhiteElo "3200"]
[TimeControl "3600+30"]
[Termination "normal"]
[PlyCount "171"]
[WhiteType "program"]
[BlackType "program"]

1. e4 {+0.21/21 190} e6 2. d4 {+0.16/22 252} d5 3. exd5 {+0.16/23 198} exd5
4. Nf3 {+0.15/23 56} Bd6 5. Bd3 {+0.13/22 218} Nf6 6. O-O {+0.09/22 0} O-O
7. Re1 {+0.07/22 24} h6 8. a3 {+0.11/22 186} Nbd7 9. Nbd2 {+0.15/22 138}
Re8 10. Rxe8+ {+0.17/23 124} Qxe8 11. c4 {+0.17/23 0} c6 12. Qc2 {+0.16/24
513} Nf8 13. b4 {+0.16/24 0} a5 14. bxa5 {+0.09/23 127} Rxa5 15. c5
{+0.08/24 227} Bc7 16. Bb2 {+0.10/23 51} Be6 17. Ne5 {+0.11/22 105} Ng4 18.
Ndf3 {+0.17/22 97} Nxe5 19. Nxe5 {+0.18/22 39} Bxe5 20. dxe5 {+0.13/20 26}
Qe7 21. Qd2 {+0.03/22 313} Ra8 22. Rc1 {+0.03/22 0} Ng6 23. f3 {+0.03/22 0}
Nh4 24. Qf4 {0.00/24 39} h5 25. Be2 {-0.12/23 344} Ng6 26. Qb4 {-0.14/23
86} Bf5 27. f4 {-0.14/23 0} h4 28. h3 {-0.14/23 126} Nf8 29. Bg4 {-0.21/23
0} Bxg4 30. hxg4 {-0.21/23 0} Rb8 31. Bd4 {-0.07/23 77} Ne6 32. f5
{-0.07/23 0} Ng5 33. Qd2 {-0.07/25 67} Ne4 34. Qe3 {-0.07/25 0} Ra8 35. Rb1
{0.00/23 68} Ng3 36. Rb4 {0.00/24 55} Rf8 37. Rb3 {0.00/24 62} Kh7 38. Rb4
{0.00/24 89} Ra8 39. Qd3 {0.00/24 46} Kh8 40. Qd2 {0.00/24 91} Re8 41. Rb3
{0.00/24 64} Kg8 42. a4 {0.00/23 58} Ne4 43. Qc1 {0.00/24 51} Qc7 44. Qe1
{0.00/24 52} Ng3 45. a5 {0.00/23 64} Qe7 46. Qd1 {0.00/23 0} Ne4 47. Rb6
{0.00/22 46} h3 48. gxh3 {+0.08/20 27} Nxc5 49. Kg2 {0.00/21 27} Ne4 50.
Qb3 {0.00/23 28} Qc7 51. e6 {0.00/23 0} Re7 52. exf7+ {0.00/23 0} Rxf7 53.
Bg1 {0.00/22 3} Qd6 54. Qe3 {0.00/22 7} Re7 55. Qd3 {0.00/24 490} Qh6 56.
Qe3 {0.00/23 43} Qf6 57. Qf4 {0.00/24 34} d4 58. g5 {+0.03/22 57} Nxg5 59.
Bxd4 {+0.01/20 18} Re2+ 60. Kf1 {0.00/20 0} Qe7 61. Bf2 {0.00/20 0} Rc2 62.
Qe3 {0.00/23 39} Qe4 63. Qxe4 {0.00/23 0} Nxe4 64. Be1 {0.00/23 0} Nd6 65.
h4 {0.00/23 11} Rc5 66. f6 {0.00/23 0} gxf6 67. Bb4 {0.00/25 17} Rf5+ 68.
Kg2 {0.00/28 21} Nc4 69. Rxb7 {0.00/28 0} Nxa5 70. Rc7 {0.00/30 45} Nc4 71.
Rxc6 {0.00/30 0} Ne5 72. Rc8+ {0.00/29 49} Kg7 73. Bc3 {0.00/29 20} Rf4 74.
Bxe5 {0.00/29 0} fxe5 75. Rc7+ {0.00/38 3} Rf7 76. Rxf7+ {0.00/43 55} Kxf7
77. h5 {0.00/43 0} Kf6 78. h6 {0.00/127 1} Kg6 79. h7 {0.00/127 1} Kxh7 80.
Kg3 {0.00/127 0} Kh8 81. Kg4 {0.00/127 1} Kg8 82. Kf5 {0.00/127 1} Kh8 83.
Kg4 {0.00/127 1} Kh7 84. Kg3 {0.00/127 0} Kg8 85. Kf3 {0.00/127 1} Kh8 86.
Kg4 {0.00/127 1 3-fold repetition} 1/2-1/2[/pgn]

[velmarin]

Gull Syzygy 3.1 is still not capable to solve Behting Study

On my computer with 6 tables syzygy,
not solve it or Komodo, Arasan, Vajolet Syzygy or Gull version Basil.
StocFish View Kc6, but with a negative score of more than 3 cps.

It seems a fairly complicated position for engines.

Hawkeye ( derivative based on Gull 3 syzygy) - sees Kc6 in less than 2 minutes with single core:

Code: Select all

here's the full pv provided at 142 seconds:
info time 142626 depth 25 seldepth 53 score cp -321 nps 2642386 nodes 376873046 tbhits 2370480 pv d5c6 g2g1q f5h4 g1a7 h4f3 a7a5 c6d6 a5b5 d6e6 b5c5 e6f5 c5d5 f5f4 d5b7 f4f5 b7a6 f5e4 a6d6 e4f5 d6c7 f5f4 c7a7 f4e4 a7a4 e4d4 a4b4 d4d5 b4b5 d5d4 b5b3 d4c5 c4c3 d2c3 b3c3 c5d5 c3b3 d5e4 b3b5 f3d4 b5b4 e4e3 b4c3 e3e4 h5g5 d4e6 g5f6

Your version looks in that line Kc6, but the score is - 321, it is not a good line.
If we follow your line with the pgn Viewer we arrive to a loss position,
[pgn]
[Event "Blitz 1m"]
[Site "?"]
[Date "????.??.??"]
[Round "?"]
[White "Gull_BA"]
[Black "Gull_BA"]
[Result "*"]
[SetUp "1"]
[FEN "8/8/7p/3KNN1k/2p4p/8/3P2p1/8 w - - 0 1"]
[PlyCount "46"]
[TimeControl "60"]

{509MB, Mio4.ctg, DESKTOP-QFSOBQ2} 1. Kc6 g1=Q 2. Nxh4 Qa7 3. Nhf3 Qa5 4. Kd6 Qb5 5. Ke6 Qc5 6. Kf5 Qd5 7. Kf4 Qb7 8. Kf5 Qa6 9.
Ke4 Qd6 10. Kf5 Qc7 11. Kf4 Qa7 12. Ke4 Qa4 13. Kd4 Qb4 14. Kd5 Qb5+ 15. Kd4
Qb3 16. Kc5 c3 17. dxc3 Qxc3+ 18. Kd5 Qb3+ 19. Ke4 Qb5 20. Nd4 Qb4 21. Ke3 Qc3+
22. Ke4 Kg5 23. Ne6+ Kf6 *
[/pgn]

Remember that this version uses the original ThinkingALot code,
ThinkingALot changed the Tbprobe code to accommodate Gull,
I guess that that has its pros and cons, the code is there, and I think that functional.
The GitHub and the source code is there for anyone who wants to play with him.