The occasional victory

[sje]

The occasional victory

At present, I'm working on getting the Symbolic re-write to understand XBoard protocol version two so that it may soon be running on an ICS.

Today's sparring partner is Sjeng, which is the engine behind the Chess application provided as part of Apple's Mac OS/X. Symbolic is not doing very well here, but does gain the occasional victory:
[pgn][Event "Unknown event"]
[Site "gail"]
[Date "2015-08-26"]
[Round "10"]
[White "Sjeng"]
[Black "Symbolic 2015-08-26"]
[Result "0-1"]
[Termination "White is checkmated"]

1. e4 Nf6 2. e5 Nd5 3. c4 Nb6 4. d4 d6 5. Nf3 g6 6. exd6 cxd6 7. Be2 Bg7 8. O-O O-O 9. Nc3 Bg4 10.
c5 dxc5 11. dxc5 N6d7 12. Qd5 Nc6 13. Ne4 Nf6 14. Nxf6+ Bxf6 15. Qb3 Qc7 16. Re1 Be6 17. Qa3 Nd4
18. Nxd4 Bxd4 19. Be3 Rfd8 20. Bxd4 Rxd4 21. Rad1 Re4 22. Bf1 Rxe1 23. Rxe1 a6 24. Qe3 Qd7 25. a3
Rd8 26. b4 Qd5 27. Rc1 Qc6 28. Qg5 f6 29. Qh6 Rd4 30. Re1 g5 31. g3 Qd7 32. c6 Qxc6 33. h3 Rd2 34.
Bg2 Qb6 35. Re3 Bd5 36. Bxd5+ Rxd5 37. Qh5 Qd8 38. h4 gxh4 39. Qg4+ Kh8 40. Qxh4 Rd4 41. Qh6 Rd1+
42. Kg2 Qd5+ 43. Rf3 Kg8 44. Qe3 Kf7 45. Qb3 Qxb3 46. Rxb3 Rc1 47. b5 a5 48. Rd3 Ra1 49. Rd7 Rxa3
50. Rxb7 Rb3 51. b6 a4 52. Ra7 Rxb6 53. Rxa4 e5 54. Rh4 Kg6 55. Rg4+ Kf5 56. Rh4 Rb7 57. Kf3 Ra7
58. Rh1 Kg5 59. Ke4 h5 60. Kd5 Ra2 61. Rh2 Ra5+ 62. Ke6 f5 63. Rh1 f4 64. Re1 Kg4 65. Re4 Kf3 66.
Rxe5 Rxe5+ 67. Kxe5 fxg3 68. fxg3 Kxg3 69. Kf5 h4 70. Ke4 h3 71. Ke3 h2 72. Ke4 h1=Q+ 73. Ke5 Qd1
74. Kf6 Kf4 75. Kf7 Qd6 76. Kg8 Kg5 77. Kf7 Qd7+ 78. Kf8 Kf6 79. Kg8 Qg7# 0-1[/pgn]
I'm glad that Symbolic can sometimes beat the OS/X default chess program because I never could.

[sje]

And another a few rounds later:
[pgn][Event "Unknown event"]
[Site "gail"]
[Date "2015-08-26"]
[Round "18"]
[White "Sjeng"]
[Black "Symbolic 2015-08-26"]
[Result "0-1"]
[Termination "White is checkmated"]

1. Nf3 Nf6 2. Nc3 d5 3. e3 Bg4 4. Bb5+ Nbd7 5. O-O c6 6. Be2 Bxf3 7. gxf3 e6 8. d4 Bd6 9. e4 dxe4
10. fxe4 e5 11. d5 cxd5 12. exd5 Bb4 13. Bg5 Bxc3 14. bxc3 Qa5 15. c4 Ne4 16. Be3 Ke7 17. Bd3 Nef6
18. Qe1 Qxe1 19. Rfxe1 h6 20. Rab1 b6 21. Bd4 Kd6 22. Ba1 Rae8 23. Bf5 Rc8 24. Rb4 Rhe8 25. Re3 a5
26. Ra4 Rc7 27. Bb2 Nc5 28. Raa3 g6 29. Bd3 Ng4 30. Re1 f5 31. Be2 Nf6 32. Rh3 Rh7 33. Rb1 f4 34.
Ba3 Nd7 35. Rhb3 Rb8 36. Rb5 Rhh8 37. Bg4 Kc7 38. d6+ Kc6 39. Bh3 Rhf8 40. f3 g5 41. Rd1 Rbe8 42.
c3 Rf7 43. Bg4 Rh8 44. h3 h5 45. Bxd7+ Nxd7 46. Rdd5 Rc8 47. Rxe5 Nxe5 48. Rxe5 Rg8 49. c5 b5 50.
Kf2 h4 51. Ke2 Rff8 52. Re7 Re8 53. Re4 Rxe4+ 54. fxe4 Rb8 55. Kd3 b4 56. cxb4 axb4 57. Bb2 Kxc5
58. e5 Kd5 59. Ke2 Ra8 60. Kf3 Kc6 61. a4 Rxa4 62. Kg4 Ra2 63. Bd4 Rg2+ 64. Kf3 Rg3+ 65. Ke4 b3 66.
Kf5 Rd3 67. Bb2 f3 68. Ke4 f2 69. Kxd3 f1=Q+ 70. Kc3 Kb5 71. Ba3 Qc4+ 72. Kd2 Qd4+ 73. Ke1 Qxe5+
74. Kf2 Ka4 75. d7 Qd4+ 76. Kg2 Kxa3 77. d8=Q Qxd8 78. Kf3 Qd3+ 79. Kg4 Qxh3+ 80. Kxh3 b2 81. Kg4
b1=Q 82. Kxg5 h3 83. Kf4 h2 84. Kg5 Qe4 85. Kf6 h1=Q 86. Kf7 Qhh7+ 87. Kf6 Qef5# 0-1[/pgn]
For the moment, I'll omit any games where Symbolic gets clobbered. You'll get to see plenty of those once I get it running on an ICS.

[sje]

Just one more:
[pgn][Event "Unknown event"]
[Site "gail"]
[Date "2015-08-26"]
[Round "85"]
[White "Symbolic 2015-08-26"]
[Black "Sjeng"]
[Result "1-0"]
[Termination "Black is checkmated"]

1. e4 Nc6 2. Nf3 e5 3. Bb5 Bb4 4. c3 Bd6 5. d4 exd4 6. cxd4 Nge7 7. Nc3 O-O 8. d5 Ne5 9. Bf4 Nxf3+
10. Qxf3 Bxf4 11. Qxf4 c6 12. Bd3 cxd5 13. exd5 Ng6 14. Qd6 Re8+ 15. Be2 Qg5 16. g3 b6 17. f4 Qe7
18. Qxe7 Nxe7 19. Kf2 Bb7 20. Bf3 d6 21. h4 Rac8 22. h5 Rc5 23. Rad1 b5 24. a3 h6 25. Rhe1 Kf8 26.
Re3 Rec8 27. Be4 a5 28. Rd4 b4 29. axb4 axb4 30. Rxb4 Bxd5 31. Na4 Rc4 32. Rxc4 Rxc4 33. Nb6 Rxe4
34. Nxd5 Rxe3 35. Nxe3 g6 36. hxg6 fxg6 37. b4 Kf7 38. b5 Ke6 39. Ke2 Nd5 40. Nd1 h5 41. Kd2 h4 42.
gxh4 Nxf4 43. Ne3 Nd5 44. Nc4 Kd7 45. Kc2 Nf4 46. Nd2 Kc7 47. Kb3 Kb6 48. Ka4 Nd3 49. Nf3 Kc5 50.
Ka5 Ne5 51. Nxe5 dxe5 52. b6 Kd6 53. h5 gxh5 54. Ka6 e4 55. b7 Kd5 56. b8=Q Kd4 57. Kb5 e3 58. Qd8+
Ke4 59. Kc4 Kf5 60. Qd3+ Kf4 61. Kd4 e2 62. Qxe2 h4 63. Qe4+ Kg5 64. Ke5 h3 65. Qf5+ Kh6 66. Kf6 h2
67. Qg6# 1-0
[/pgn]
The match score:

Code: Select all

2015-08-26 100 1m+1s vs Sjeng    12-73-15 elo-246

[mar]

While this is nice, what's the point?
I mean I could post 10 wins against Critter, neglecting the fact that way over 50% games were actually losses..

[sje]

While this is nice, what's the point?
I mean I could post 10 wins against Critter, neglecting the fact that way over 50% games were actually losses..

The idea here is to post some real games from a program which is not yet available for competition on an ICS. Further, these early games are now on record and so later can be compared with the program's later games as features are added.

[bob]

Just for the record, you are playing the program called "sjeng.ChessEngine" using the default book as provided? I did not know it was even there until you mentioned it since I test against more common programs like senpai, stockfish and such on my macbook. This sjeng version looks to be very old and very weak.

I started up an xboard match, one core per engine (crafty vs sjeng) ponder=on, 1m+1s. as of right now on the xboard display

Crafty-25.0 vs Sjeng 11.2 (15-0-0).

Are you losing games due to bugs or crashes??

[elpapa]

While this is nice, what's the point?
I mean I could post 10 wins against Critter, neglecting the fact that way over 50% games were actually losses..

He didn't neglect that fact, though.

[sje]

Just for the record, you are playing the program called "sjeng.ChessEngine" using the default book as provided? I did not know it was even there until you mentioned it since I test against more common programs like senpai, stockfish and such on my macbook. This sjeng version looks to be very old and very weak.

I started up an xboard match, one core per engine (crafty vs sjeng) ponder=on, 1m+1s. as of right now on the xboard display

Crafty-25.0 vs Sjeng 11.2 (15-0-0).

Are you losing games due to bugs or crashes??

Yep, it's sjeng.ChessEngine alright.

Symbolic is losing games because it's currently lacking many search features. It's evaluation function is pretty much only material plus mobility. There is no pondering nor is there a multithreaded search. The move ordering is very simple and there are no depth adjustments beyond check extension. At present I'm concerned mostly with getting glitch-free operation under XBoard on a Mac locally and then with XBoard connected to an ICS.

[bob]

Just for the record, you are playing the program called "sjeng.ChessEngine" using the default book as provided? I did not know it was even there until you mentioned it since I test against more common programs like senpai, stockfish and such on my macbook. This sjeng version looks to be very old and very weak.

I started up an xboard match, one core per engine (crafty vs sjeng) ponder=on, 1m+1s. as of right now on the xboard display

Crafty-25.0 vs Sjeng 11.2 (15-0-0).

Are you losing games due to bugs or crashes??

Yep, it's sjeng.ChessEngine alright.

Symbolic is losing games because it's currently lacking many search features. It's evaluation function is pretty much only material plus mobility. There is no pondering nor is there a multithreaded search. The move ordering is very simple and there are no depth adjustments beyond check extension. At present I'm concerned mostly with getting glitch-free operation under XBoard on a Mac locally and then with XBoard connected to an ICS.

Just for the record, I forgot about the match and after about 150 straight wins I stopped it. No flag forfeits, no sjeng glitches, so xboard seems to work well for all the programs I use here...

[mar]

The idea here is to post some real games from a program which is not yet available for competition on an ICS. Further, these early games are now on record and so later can be compared with the program's later games as features are added.

I see, sorry.
Congratulations that Symbolic can now compete against other engines, will be interesting to follow as it progresses.