MoveOrdering--

Discussion of chess software programming and technical issues.

Moderator: Ras

Henk
Posts: 7251
Joined: Mon May 27, 2013 10:31 am

MoveOrdering--

Post by Henk »

[d] r5nk/pp1br2p/2p1p3/3p4/3P4/2NB2R1/PPP3PP/2K2R2 w - - 2 22


Don't know what your engine or piece square table says about the rank of move Nc3-b1 in the move ordering. My chess book says Nc3-b1!!
User avatar
stegemma
Posts: 859
Joined: Mon Aug 10, 2009 10:05 pm
Location: Italy
Full name: Stefano Gemma

Re: MoveOrdering--

Post by stegemma »

Henk wrote:[d] r5nk/pp1br2p/2p1p3/3p4/3P4/2NB2R1/PPP3PP/2K2R2 w - - 2 22


Don't know what your engine or piece square table says about the rank of move Nc3-b1 in the move ordering. My chess book says Nc3-b1!!
Maybe is going to d2 and then f3, to control e5 square?
Author of Drago, Raffaela, Freccia, Satana, Sabrina.
http://www.linformatica.com
Henk
Posts: 7251
Joined: Mon May 27, 2013 10:31 am

Re: MoveOrdering--

Post by Henk »

stegemma wrote:
Henk wrote:[d] r5nk/pp1br2p/2p1p3/3p4/3P4/2NB2R1/PPP3PP/2K2R2 w - - 2 22


Don't know what your engine or piece square table says about the rank of move Nc3-b1 in the move ordering. My chess book says Nc3-b1!!
Maybe is going to d2 and then f3, to control e5 square?
Yes these moves were played. Stockfish does not see Nb1. Perhaps engines only understand tactical play or easy strategic play.

Position is taken from Lasker-Pillsbury Paris 1900. Lasker was world chess champion at that time.
zullil
Posts: 6442
Joined: Tue Jan 09, 2007 12:31 am
Location: PA USA
Full name: Louis Zulli

Re: MoveOrdering--

Post by zullil »

Henk wrote:
stegemma wrote:
Henk wrote:[d] r5nk/pp1br2p/2p1p3/3p4/3P4/2NB2R1/PPP3PP/2K2R2 w - - 2 22


Don't know what your engine or piece square table says about the rank of move Nc3-b1 in the move ordering. My chess book says Nc3-b1!!
Maybe is going to d2 and then f3, to control e5 square?
Yes these moves were played. Stockfish does not see Nb1. Perhaps engines only understand tactical play or easy strategic play.

Position is taken from Lasker-Pillsbury Paris 1900. Lasker was world chess champion at that time.
Stockfish "sees" the move. It just doesn't consider it best. Fifth best, in fact, at depth 35:

Code: Select all

info depth 35 seldepth 56 multipv 1 score cp 63 nodes 9404252950 nps 31655944 hashfull 980 tbhits 0 time 297077 pv c3e2 a8e8 g3f3 e6e5 d4e5 d7g4 f3e3 g4e2 e3e2 e7e5 e2e5 e8e5 f1f8 h8g7 f8a8 a7a6 g2g4 g8f6 a8a7 e5e1 c1d2 e1e7 d3a6 f6e4 d2d1 e4d6 a6e2 g7f6 a7a3 f6g5 a3h3 g5f4 h3h5 e7f7 c2c3 f7e7 h2h3 d6e4 e2d3 f4e3 d1c2 e7f7

info depth 35 seldepth 56 multipv 2 score cp 62 nodes 9404252950 nps 31655944 hashfull 980 tbhits 0 time 297077 pv g3g5 g8h6 g5h5 h6f5 c3e2 a8g8 d3f5 e6f5 e2f4 e7e4 c2c3 h8g7 h5g5 g7f8 g5g8 f8g8 c1d2 g8g7 g2g3 g7f6 f4d3 d7c8 d3f2 e4e7 f2g4 f6g7 g4e5 g7f6 d2d3 c8e6 f1e1 e7e8 d3d2 h7h5 e5d3 h5h4 g3h4 e8h8 d3f4 e6c8

info depth 35 seldepth 56 multipv 3 score cp 61 nodes 9404252950 nps 31655944 hashfull 980 tbhits 0 time 297077 pv g3f3 a8e8 c3e2 e6e5 d4e5 d7g4 f3e3 g4e2 e3e2 e7e5 e2e5 e8e5 f1f8 h8g7 f8a8 a7a6 g2g4 g8f6 a8a7 e5e1 c1d2 e1e7 d3a6 f6e4 d2d1 e4d6 a6e2 g7f6 a7a3 f6g5 a3h3 g5f4 h3h5 e7f7 g4g5 d6f5 c2c3 f7e7 e2d3 f4g4

info depth 35 seldepth 56 multipv 4 score cp 55 nodes 9404252950 nps 31655944 hashfull 980 tbhits 0 time 297077 pv h2h3 g8h6 g3f3 e6e5 d4e5 a8e8 c3e2 e7e5 e2f4 e8e7 g2g4 d7e8 f1g1 e7g7 f3f1 h6g8 d3f5 b7b6 h3h4 g8f6 f4d3 e5e7 c2c3 f6e4 d3f4 e8d7 f5d7 e7d7 f4e6 g7g8 c1c2 g8e8 e6d4 c6c5 d4f5 a7a5 f5e3

info depth 35 seldepth 56 multipv 5 score cp 46 nodes 9404252950 nps 31655944 hashfull 980 tbhits 0 time 297077 pv c3b1 g8h6 b1d2 e7g7 d2f3 h6f7 g3g7 h8g7 c2c4 a8c8 b2b4 a7a5 b4a5 c6c5 c4d5 e6d5 f1e1 c5c4 d3c2 c8e8 e1e8 d7e8 c1d2 f7d6 f3g5 e8d7 c2h7 d6b5 d2e3 b5c3 h7c2 c3a2 e3d2 a2b4 d2c3 b4c6 h2h3
Looks like Pillsbury should have played 22...Nh6, not 22...Re8.

Here's the game:
[pgn]
[Event "Paris"]
[Site "Paris FRA"]
[Date "1900.06.14"]
[EventDate "1900.??.??"]
[Round "15"]
[Result "1-0"]
[White "Emanuel Lasker"]
[Black "Harry Nelson Pillsbury"]
[ECO "A83"]
[WhiteElo "?"]
[BlackElo "?"]
[PlyCount "169"]

1. d4 f5 2. e4 fxe4 3. Nc3 Nf6 4. Bg5 c6 5. f3 exf3 6. Nxf3 e6
7. Bd3 Be7 8. Ne5 O-O 9. Bxf6 Rxf6 10. Qh5 g6 11. Nxg6 Qe8
12. Nxe7+ Qxe7 13. O-O-O d5 14. Rde1 Nd7 15. Re3 Rf7 16. Rg3+
Kh8 17. Bg6 Rg7 18. Rf1 Nf6 19. Qh4 Ng8 20. Qxe7 Rxe7 21. Bd3
Bd7 22. Nb1 Rae8 23. Nd2 e5 24. dxe5 Rxe5 25. Nf3 Re3 26. Ng5
Rxg3 27. hxg3 h6 28. Nf7+ Kg7 29. Nd6 Re7 30. Nxb7 Nf6 31. Nc5
Bg4 32. Rf4 Bc8 33. Ra4 Ng4 34. Ba6 Bf5 35. Rf4 Ne3 36. c3 Kg6
37. Rf2 Be4 38. b3 Bxg2 39. Bd3+ Kg5 40. Rf8 Kg4 41. Rg8+ Kf3
42. Rg6 Ng4 43. Bf5 h5 44. Rg5 Re1+ 45. Kb2 Rh1 46. Bg6 Kxg3
47. Bxh5 Bh3 48. Bxg4 Bxg4 49. Rg6 Rh2+ 50. Ka3 Rc2 51. Nd3
Kh4 52. Ne5 Bf5 53. Rxc6 Kg3 54. Rc5 Rd2 55. Nc6 Kf4 56. Nb4
d4 57. cxd4 Rxd4 58. Ra5 Rd7 59. Nc6 Be4 60. Nxa7 Rd2 61. Nb5
Rd5 62. Kb4 Bd3 63. Nc7 Rxa5 64. Kxa5 Ke5 65. Kb4 Kd6 66. Nb5+
Kc6 67. a4 Kb6 68. Na3 Be2 69. Nc4+ Ka6 70. Kc3 Bd1 71. Nb2
Bh5 72. b4 Be8 73. Kb3 Bc6 74. Kc4 Bd7 75. Kc5 Bg4 76. Nc4 Bd1
77. b5+ Ka7 78. a5 Bf3 79. Ne5 Bg2 80. Nc6+ Ka8 81. Kb6 Bh3
82. Nb4 Bg2 83. Na6 Bf3 84. Nc7+ Kb8 85. a6 1-0[/pgn]
zullil
Posts: 6442
Joined: Tue Jan 09, 2007 12:31 am
Location: PA USA
Full name: Louis Zulli

Re: MoveOrdering--

Post by zullil »

zullil wrote:
Henk wrote:
stegemma wrote:
Henk wrote:[d] r5nk/pp1br2p/2p1p3/3p4/3P4/2NB2R1/PPP3PP/2K2R2 w - - 2 22


Don't know what your engine or piece square table says about the rank of move Nc3-b1 in the move ordering. My chess book says Nc3-b1!!
Maybe is going to d2 and then f3, to control e5 square?
Yes these moves were played. Stockfish does not see Nb1. Perhaps engines only understand tactical play or easy strategic play.

Position is taken from Lasker-Pillsbury Paris 1900. Lasker was world chess champion at that time.
Stockfish "sees" the move. It just doesn't consider it best. Fifth best, in fact, at depth 35:

Code: Select all

info depth 35 seldepth 56 multipv 1 score cp 63 nodes 9404252950 nps 31655944 hashfull 980 tbhits 0 time 297077 pv c3e2 a8e8 g3f3 e6e5 d4e5 d7g4 f3e3 g4e2 e3e2 e7e5 e2e5 e8e5 f1f8 h8g7 f8a8 a7a6 g2g4 g8f6 a8a7 e5e1 c1d2 e1e7 d3a6 f6e4 d2d1 e4d6 a6e2 g7f6 a7a3 f6g5 a3h3 g5f4 h3h5 e7f7 c2c3 f7e7 h2h3 d6e4 e2d3 f4e3 d1c2 e7f7

info depth 35 seldepth 56 multipv 2 score cp 62 nodes 9404252950 nps 31655944 hashfull 980 tbhits 0 time 297077 pv g3g5 g8h6 g5h5 h6f5 c3e2 a8g8 d3f5 e6f5 e2f4 e7e4 c2c3 h8g7 h5g5 g7f8 g5g8 f8g8 c1d2 g8g7 g2g3 g7f6 f4d3 d7c8 d3f2 e4e7 f2g4 f6g7 g4e5 g7f6 d2d3 c8e6 f1e1 e7e8 d3d2 h7h5 e5d3 h5h4 g3h4 e8h8 d3f4 e6c8

info depth 35 seldepth 56 multipv 3 score cp 61 nodes 9404252950 nps 31655944 hashfull 980 tbhits 0 time 297077 pv g3f3 a8e8 c3e2 e6e5 d4e5 d7g4 f3e3 g4e2 e3e2 e7e5 e2e5 e8e5 f1f8 h8g7 f8a8 a7a6 g2g4 g8f6 a8a7 e5e1 c1d2 e1e7 d3a6 f6e4 d2d1 e4d6 a6e2 g7f6 a7a3 f6g5 a3h3 g5f4 h3h5 e7f7 g4g5 d6f5 c2c3 f7e7 e2d3 f4g4

info depth 35 seldepth 56 multipv 4 score cp 55 nodes 9404252950 nps 31655944 hashfull 980 tbhits 0 time 297077 pv h2h3 g8h6 g3f3 e6e5 d4e5 a8e8 c3e2 e7e5 e2f4 e8e7 g2g4 d7e8 f1g1 e7g7 f3f1 h6g8 d3f5 b7b6 h3h4 g8f6 f4d3 e5e7 c2c3 f6e4 d3f4 e8d7 f5d7 e7d7 f4e6 g7g8 c1c2 g8e8 e6d4 c6c5 d4f5 a7a5 f5e3

info depth 35 seldepth 56 multipv 5 score cp 46 nodes 9404252950 nps 31655944 hashfull 980 tbhits 0 time 297077 pv c3b1 g8h6 b1d2 e7g7 d2f3 h6f7 g3g7 h8g7 c2c4 a8c8 b2b4 a7a5 b4a5 c6c5 c4d5 e6d5 f1e1 c5c4 d3c2 c8e8 e1e8 d7e8 c1d2 f7d6 f3g5 e8d7 c2h7 d6b5 d2e3 b5c3 h7c2 c3a2 e3d2 a2b4 d2c3 b4c6 h2h3
Looks like Pillsbury should have played 22...Nh6, not 22...Rae8.
Actually, Stockfish now thinks 22...e5 is best for Black in reply to 22. Nb1.

Code: Select all

info depth 48 seldepth 73 multipv 1 score cp -35 nodes 171612704673 nps 35479044 hashfull 999 tbhits 0 time 4837016 pv e6e5 d4e5 a8e8 b1d2 e7e5 d2f3 e5e3 g3g5 g8f6 f3d4 e3e1 f1e1 e8e1 c1d2 e1e7 b2b4 e7e8 a2a3 f6e4 d3e4 e8e4 c2c3 e4e8 d4f3 e8e7 g5e5 e7g7 g2g3 h8g8 d2e3 g8f8 f3g5 h7h6 g5e6 d7e6 e5e6 g7e7 e6e7 f8e7 e3f4 e7e6 g3g4 b7b6 g4g5 h6g5 f4g5 c6c5 b4c5 b6c5 g5f4 c5c4 h2h4 d5d4 c3d4 a7a6
zullil
Posts: 6442
Joined: Tue Jan 09, 2007 12:31 am
Location: PA USA
Full name: Louis Zulli

Re: MoveOrdering--

Post by zullil »

zullil wrote:
zullil wrote:
Henk wrote:
stegemma wrote:
Henk wrote:[d] r5nk/pp1br2p/2p1p3/3p4/3P4/2NB2R1/PPP3PP/2K2R2 w - - 2 22


Don't know what your engine or piece square table says about the rank of move Nc3-b1 in the move ordering. My chess book says Nc3-b1!!
Maybe is going to d2 and then f3, to control e5 square?
Yes these moves were played. Stockfish does not see Nb1. Perhaps engines only understand tactical play or easy strategic play.

Position is taken from Lasker-Pillsbury Paris 1900. Lasker was world chess champion at that time.
Stockfish "sees" the move. It just doesn't consider it best. Fifth best, in fact, at depth 35:

Code: Select all

info depth 35 seldepth 56 multipv 1 score cp 63 nodes 9404252950 nps 31655944 hashfull 980 tbhits 0 time 297077 pv c3e2 a8e8 g3f3 e6e5 d4e5 d7g4 f3e3 g4e2 e3e2 e7e5 e2e5 e8e5 f1f8 h8g7 f8a8 a7a6 g2g4 g8f6 a8a7 e5e1 c1d2 e1e7 d3a6 f6e4 d2d1 e4d6 a6e2 g7f6 a7a3 f6g5 a3h3 g5f4 h3h5 e7f7 c2c3 f7e7 h2h3 d6e4 e2d3 f4e3 d1c2 e7f7

info depth 35 seldepth 56 multipv 2 score cp 62 nodes 9404252950 nps 31655944 hashfull 980 tbhits 0 time 297077 pv g3g5 g8h6 g5h5 h6f5 c3e2 a8g8 d3f5 e6f5 e2f4 e7e4 c2c3 h8g7 h5g5 g7f8 g5g8 f8g8 c1d2 g8g7 g2g3 g7f6 f4d3 d7c8 d3f2 e4e7 f2g4 f6g7 g4e5 g7f6 d2d3 c8e6 f1e1 e7e8 d3d2 h7h5 e5d3 h5h4 g3h4 e8h8 d3f4 e6c8

info depth 35 seldepth 56 multipv 3 score cp 61 nodes 9404252950 nps 31655944 hashfull 980 tbhits 0 time 297077 pv g3f3 a8e8 c3e2 e6e5 d4e5 d7g4 f3e3 g4e2 e3e2 e7e5 e2e5 e8e5 f1f8 h8g7 f8a8 a7a6 g2g4 g8f6 a8a7 e5e1 c1d2 e1e7 d3a6 f6e4 d2d1 e4d6 a6e2 g7f6 a7a3 f6g5 a3h3 g5f4 h3h5 e7f7 g4g5 d6f5 c2c3 f7e7 e2d3 f4g4

info depth 35 seldepth 56 multipv 4 score cp 55 nodes 9404252950 nps 31655944 hashfull 980 tbhits 0 time 297077 pv h2h3 g8h6 g3f3 e6e5 d4e5 a8e8 c3e2 e7e5 e2f4 e8e7 g2g4 d7e8 f1g1 e7g7 f3f1 h6g8 d3f5 b7b6 h3h4 g8f6 f4d3 e5e7 c2c3 f6e4 d3f4 e8d7 f5d7 e7d7 f4e6 g7g8 c1c2 g8e8 e6d4 c6c5 d4f5 a7a5 f5e3

info depth 35 seldepth 56 multipv 5 score cp 46 nodes 9404252950 nps 31655944 hashfull 980 tbhits 0 time 297077 pv c3b1 g8h6 b1d2 e7g7 d2f3 h6f7 g3g7 h8g7 c2c4 a8c8 b2b4 a7a5 b4a5 c6c5 c4d5 e6d5 f1e1 c5c4 d3c2 c8e8 e1e8 d7e8 c1d2 f7d6 f3g5 e8d7 c2h7 d6b5 d2e3 b5c3 h7c2 c3a2 e3d2 a2b4 d2c3 b4c6 h2h3
Looks like Pillsbury should have played 22...Nh6, not 22...Rae8.
Actually, Stockfish now thinks 22...e5 is best for Black in reply to 22. Nb1.

Code: Select all

info depth 48 seldepth 73 multipv 1 score cp -35 nodes 171612704673 nps 35479044 hashfull 999 tbhits 0 time 4837016 pv e6e5 d4e5 a8e8 b1d2 e7e5 d2f3 e5e3 g3g5 g8f6 f3d4 e3e1 f1e1 e8e1 c1d2 e1e7 b2b4 e7e8 a2a3 f6e4 d3e4 e8e4 c2c3 e4e8 d4f3 e8e7 g5e5 e7g7 g2g3 h8g8 d2e3 g8f8 f3g5 h7h6 g5e6 d7e6 e5e6 g7e7 e6e7 f8e7 e3f4 e7e6 g3g4 b7b6 g4g5 h6g5 f4g5 c6c5 b4c5 b6c5 g5f4 c5c4 h2h4 d5d4 c3d4 a7a6
And now Stockfish thinks Pillsbury was right!

Code: Select all

info depth 49 seldepth 79 multipv 1 score cp -41 nodes 258029106998 nps 35462629 hashfull 999 tbhits 0 time 7276085 pv a8e8 b1d2 e6e5 d4e5 e7e5 d2f3 e5e3 g3g5 g8f6 f3d4 e3e1 f1e1 e8e1 c1d2 e1e7 b2b4 e7e8 a2a3 f6e4 d3e4 e8e4 c2c3 e4e8 d4f3 e8e7 g5e5 e7g7 g2g3 h8g8 d2e3 g8f7 e5h5 f7f6 h5h6 f6f5 a3a4 g7e7 e3d2 a7a6 a4a5 f5g4 f3d4 g4g5 h6d6 d7h3 d6d8 h3f5 d8h8 f5e4 h8g8 e4g6 h2h3 g5f6 g8f8 f6g5 d4f3 g5h6 f8f4 g6e4 f3d4 h6g5 f4f2 g5g6
User avatar
stegemma
Posts: 859
Joined: Mon Aug 10, 2009 10:05 pm
Location: Italy
Full name: Stefano Gemma

Re: MoveOrdering--

Post by stegemma »

Henk wrote:
stegemma wrote:
Henk wrote:[d] r5nk/pp1br2p/2p1p3/3p4/3P4/2NB2R1/PPP3PP/2K2R2 w - - 2 22


Don't know what your engine or piece square table says about the rank of move Nc3-b1 in the move ordering. My chess book says Nc3-b1!!
Maybe is going to d2 and then f3, to control e5 square?
Yes these moves were played. Stockfish does not see Nb1. Perhaps engines only understand tactical play or easy strategic play.

Position is taken from Lasker-Pillsbury Paris 1900. Lasker was world chess champion at that time.
Wow... so I think as Lasker! :)
Author of Drago, Raffaela, Freccia, Satana, Sabrina.
http://www.linformatica.com
Henk
Posts: 7251
Joined: Mon May 27, 2013 10:31 am

Re: MoveOrdering--

Post by Henk »

[d] 3q1rk1/pr2bppp/1pp1pn2/8/2P1N3/1P4P1/PB1PQPKP/3R1R2 w - - 2 16

Another example that chess engines don't understand good positional play.

My book says Ne4-g5!

Stockfish can't find it.

Position is from Botwinnik - Stahlberg (Amsterdam 1954).

Book says that the knight has to go to f3 because it has to control the important squares d4 and e5.
Henk
Posts: 7251
Joined: Mon May 27, 2013 10:31 am

Re: MoveOrdering--

Post by Henk »

stegemma wrote:
Henk wrote:
stegemma wrote:
Henk wrote:[d] r5nk/pp1br2p/2p1p3/3p4/3P4/2NB2R1/PPP3PP/2K2R2 w - - 2 22


Don't know what your engine or piece square table says about the rank of move Nc3-b1 in the move ordering. My chess book says Nc3-b1!!
Maybe is going to d2 and then f3, to control e5 square?
Yes these moves were played. Stockfish does not see Nb1. Perhaps engines only understand tactical play or easy strategic play.

Position is taken from Lasker-Pillsbury Paris 1900. Lasker was world chess champion at that time.
Wow... so I think as Lasker! :)
Ok just for you Stefano. Find the next Lasker move. Stockfish finds it at depth 28 but rejects it. My book gives it two exclamation marks.

[d]r1br3k/pp2Npbp/6p1/4p1P1/5P1P/2n1B3/PPP5/R3KB1R w KQ - 2 18
User avatar
cdani
Posts: 2204
Joined: Sat Jan 18, 2014 10:24 am
Location: Andorra

Re: MoveOrdering--

Post by cdani »

Henk wrote:
Another example that chess engines don't understand good positional play.
Of course.

Most of the latest hundred of elo gained by engines where in tactical play (i.e. reductions and the like) or basic statistically based positional play, but not on deep understanding of the position.

I compare them to excavators; they clean the field, but don't construct much :-)

Much has to be done to reach 4000 elo :-)