Chispa impressions and problems
Posted: 02 Jan 2005, 19:09Hi,
today I tried the engine Chispa 401 but I had a few problems with it (GUI WinBoard 4.2.7, Athlon64 3200+(2.0GHz) notebook, 512MB RAM, Windows XP SP2).
a.) Chispa refused to play me with the white pieces.
(tried with both executables, always reproducable - default opening book).
Chispa will play the first move out of its book, then reject answers as illegal. WinBoard.debug:
[code]
recognized 'normal' (-1) as variant normal
WinBoard 4.2.7 + chispa401-blend
Reset(1, 0) from gameMode 0
recognized 'normal' (-1) as variant normal
GameEnds(0, (null), 2)
StartChildProcess (dir="c:\chispa") chispa401-blend
531 >first : xboard
protover 2
531 <first : feature done=0
531 >first : accepted done
1156 <first : feature ping=1 setboard=1 usermove=1 draw=0 ics=1 name=1 sigint=0 sigterm=0 myname="Chispa 4.0.1" variants="normal,fischerandom,nocastle" done=1
1156 >first : accepted ping
1156 >first : accepted setboard
1156 >first : accepted usermove
1156 >first : accepted draw
1156 >first : accepted ics
1156 >first : accepted name
1156 >first : accepted sigint
1156 >first : accepted sigterm
1156 >first : accepted myname
1156 >first : accepted variants
1156 >first : accepted done
1171 >first : new
random
1171 >first : ics -
1171 >first : level 0 30 0
1171 >first : post
1171 >first : hard
1171 >first : ping 1
1171 <first : Error (Unknown command): ics -
1171 <first :
1171 <first : pong 1
4375 >first : name ba2axf
4375 >first : black
4375 >first : time 180000
otim 180000
4375 >first : white
go
4375 <first : Error (Unknown command): name ba2axf
4375 <first :
4375 <first : 0 1 0 0 Nf3 (Book move)
4375 <first :
4375 <first : move g1f3
4546 <first :
4578 <first : 0 1 0 0 Nf6 (Book move)
4578 <first :
4578 <first : move g8f6
Ignoring move out of turn by first, gameMode 1, forwardMost 1
4578 <first :
6343 >first : time 180000
otim 179803
6343 >first : usermove 6343 >first : g8f6
6343 <first : Illegal move: g8f6
GameEnds(0, (null), 2)
19359 >first : force
19359 >first : ping 2
19359 >first : quit
[/code]
b.) Chispa sometimes doesn't use its opening book
After it successfully played its first game with the black pieces (a long one with lots of tablebase hits, so potentially using lots of RAM (no idea if relevant)) , it started the second game just fine, but was out of book after 1. d4 , which it should not have been (no winboard.debug saved).
When starting and closing WinBoard with Chispa several times I also noted that sometimes WinBoard doesn't get the "name" from Chispa ( race condition?)
I suspect that this is all easily reproducable so I didn't try for long - if not, I can try to collect more helpful data. Probably playing a human ( so no genuine engine-engine game) has not been tested much ?!
c.) The game/ a few impressions
I let Chispa play a game/30 minutes against Hiarcs 9.5a on a Palm Tungsten T5 (2MB Hash) . It used 160MB Hash itself, 3-4-5-men tablebases.
I like Chispa's smiley faces very much when it displays its mainlines btw
- it also searches quite deep.
It underevaluated its queen in the game as many amateur engines so , in both controversial material constellations in this game. Hiarcs and Chispa disagreed by more than 2 pawns in eval. All commercial engines are mad about their queen actually - I guess they know what they are doing
Chispa still managed to get a promising position, outsearching Hiarcs some, but it doesn't know that you usually can't mate with 2 knights, which allowed Hiarcs to escape to a draw by 51. Qxa7, a manouvre Chispa didn't get at all, and it thought for several moves that it were up by about 6 pawns.
Hope some of this helps - thanks for this nice engine, Federico
Peter
[Event "Computer chess game"]
[Site "XP3200"]
[Date "2005.01.02"]
[Round "-"]
[White "Hiarcs 9.5a"]
[Black "Chispa 4.0.1"]
[Result "1/2-1/2"]
[TimeControl "1800"]
1. d4 Nf6 2. c4 e6 3. Nc3 Bb4 4. Qc2 O-O 5. a3 Bxc3+ 6. Qxc3 b6 7. Bg5 Bb7
8. f3 d6 9. e3 Nbd7 10. Nh3 c5 11. dxc5 bxc5 12. Be2 a5 13. Rd1 Qc7 14. O-O
d5 15. b3 a4 16. b4 cxb4 17. Qxb4 Ba6 18. Rc1 Rfc8 19. cxd5 Qxc1 20. Bxa6
Rxa6 21. Rxc1 Rxc1+ 22. Kf2 Nxd5 23. Qb7 Nc7 24. e4 Rc2+ 25. Kg3 f6 26. Be3
Rc4 27. Nf4 Rd6 28. Qc8+ Nf8 29. Nh5 Rd7 30. Bh6 Rf7 31. Bd2 e5 32. Kh3 g6
33. Ng3 Kg7 34. Nf1 Nfe6 35. Kg3 Rc2 36. Qb8 Rd7 37. Qb4 g5 38. Bxg5 Nxg5
39. Qxa4 Rdd2 40. Nxd2 Rxd2 41. Qb4 Rd7 42. a4 Nge6 43. a5 Nd4 44. Qb7 Rf7
45. a6 Ncb5 46. Qb6 Rd7 47. f4 Na7 48. fxe5 fxe5 49. Kf2 Nac6 50. g4 h5 51.
gxh5 Rf7+ 52. Ke3 Kh6 53. Kd3 Kxh5 54. Kc4 Ra7 55. Kc5 Kh4 56. Kd6 Kg4 57.
Qxa7 Nxa7 58. Kxe5 Ndc6+ 59. Ke6 Nb4 60. e5 Nxa6 61. Kf7 Nc8 62. h3+ Kh5
63. e6 Nc7 64. Kf6 Nd5+ 65. Ke5 Nde7 66. Kf6 Nc6 67. Kf7 Kh4 68. Kg6 N8e7+
69. Kf7 Nd5 70. Ke8 Kg5 71. h4+ Kh5 72. Kd7 Nce7 73. Kd8 Nf5 74. Ke8 Nf4
75. Kf7 Ng6 76. Kf6 Nfe7 77. Kf7 Nc6 78. Ke8 Kh6 79. h5 Nge7 80. Kf7 Kh7
81. Kf8 Kh8 82. Kf7 Nf5 83. Kf6 Nce7 84. Kf7 Kh7 85. Kf8 Kh6 86. Ke8 Kg7
87. Kd8 Kh8 88. Ke8 Kg8 89. h6 Kh7 90. Kd8 Nc6+ 91. Ke8 Ne5 92. Kf8 Ng6+
93. Ke8 Nfe7 94. Kf7 Nc8 95. Ke8 Nge7 96. Kd7 Kg8 97. Kd8 Kh8 98. Kc7 Na7
99. Kd7 Nd5 100. h7 Nf6+ 101. Ke7 Ne4 102. Kf8 Ng5 103. e7 Ne6+ 104. Kf7
Nc7 105. e8=Q+ Nxe8 106. Kxe8 Kxh7
{Insufficient material} 1/2-1/2
today I tried the engine Chispa 401 but I had a few problems with it (GUI WinBoard 4.2.7, Athlon64 3200+(2.0GHz) notebook, 512MB RAM, Windows XP SP2).
a.) Chispa refused to play me with the white pieces.
(tried with both executables, always reproducable - default opening book).
Chispa will play the first move out of its book, then reject answers as illegal. WinBoard.debug:
[code]
recognized 'normal' (-1) as variant normal
WinBoard 4.2.7 + chispa401-blend
Reset(1, 0) from gameMode 0
recognized 'normal' (-1) as variant normal
GameEnds(0, (null), 2)
StartChildProcess (dir="c:\chispa") chispa401-blend
531 >first : xboard
protover 2
531 <first : feature done=0
531 >first : accepted done
1156 <first : feature ping=1 setboard=1 usermove=1 draw=0 ics=1 name=1 sigint=0 sigterm=0 myname="Chispa 4.0.1" variants="normal,fischerandom,nocastle" done=1
1156 >first : accepted ping
1156 >first : accepted setboard
1156 >first : accepted usermove
1156 >first : accepted draw
1156 >first : accepted ics
1156 >first : accepted name
1156 >first : accepted sigint
1156 >first : accepted sigterm
1156 >first : accepted myname
1156 >first : accepted variants
1156 >first : accepted done
1171 >first : new
random
1171 >first : ics -
1171 >first : level 0 30 0
1171 >first : post
1171 >first : hard
1171 >first : ping 1
1171 <first : Error (Unknown command): ics -
1171 <first :
1171 <first : pong 1
4375 >first : name ba2axf
4375 >first : black
4375 >first : time 180000
otim 180000
4375 >first : white
go
4375 <first : Error (Unknown command): name ba2axf
4375 <first :
4375 <first : 0 1 0 0 Nf3 (Book move)
4375 <first :
4375 <first : move g1f3
4546 <first :
4578 <first : 0 1 0 0 Nf6 (Book move)
4578 <first :
4578 <first : move g8f6
Ignoring move out of turn by first, gameMode 1, forwardMost 1
4578 <first :
6343 >first : time 180000
otim 179803
6343 >first : usermove 6343 >first : g8f6
6343 <first : Illegal move: g8f6
GameEnds(0, (null), 2)
19359 >first : force
19359 >first : ping 2
19359 >first : quit
[/code]
b.) Chispa sometimes doesn't use its opening book
After it successfully played its first game with the black pieces (a long one with lots of tablebase hits, so potentially using lots of RAM (no idea if relevant)) , it started the second game just fine, but was out of book after 1. d4 , which it should not have been (no winboard.debug saved).
When starting and closing WinBoard with Chispa several times I also noted that sometimes WinBoard doesn't get the "name" from Chispa ( race condition?)
I suspect that this is all easily reproducable so I didn't try for long - if not, I can try to collect more helpful data. Probably playing a human ( so no genuine engine-engine game) has not been tested much ?!
c.) The game/ a few impressions
I let Chispa play a game/30 minutes against Hiarcs 9.5a on a Palm Tungsten T5 (2MB Hash) . It used 160MB Hash itself, 3-4-5-men tablebases.
I like Chispa's smiley faces very much when it displays its mainlines btw
- it also searches quite deep.
It underevaluated its queen in the game as many amateur engines so , in both controversial material constellations in this game. Hiarcs and Chispa disagreed by more than 2 pawns in eval. All commercial engines are mad about their queen actually - I guess they know what they are doing
Chispa still managed to get a promising position, outsearching Hiarcs some, but it doesn't know that you usually can't mate with 2 knights, which allowed Hiarcs to escape to a draw by 51. Qxa7, a manouvre Chispa didn't get at all, and it thought for several moves that it were up by about 6 pawns.
Hope some of this helps - thanks for this nice engine, Federico
Peter
[Event "Computer chess game"]
[Site "XP3200"]
[Date "2005.01.02"]
[Round "-"]
[White "Hiarcs 9.5a"]
[Black "Chispa 4.0.1"]
[Result "1/2-1/2"]
[TimeControl "1800"]
1. d4 Nf6 2. c4 e6 3. Nc3 Bb4 4. Qc2 O-O 5. a3 Bxc3+ 6. Qxc3 b6 7. Bg5 Bb7
8. f3 d6 9. e3 Nbd7 10. Nh3 c5 11. dxc5 bxc5 12. Be2 a5 13. Rd1 Qc7 14. O-O
d5 15. b3 a4 16. b4 cxb4 17. Qxb4 Ba6 18. Rc1 Rfc8 19. cxd5 Qxc1 20. Bxa6
Rxa6 21. Rxc1 Rxc1+ 22. Kf2 Nxd5 23. Qb7 Nc7 24. e4 Rc2+ 25. Kg3 f6 26. Be3
Rc4 27. Nf4 Rd6 28. Qc8+ Nf8 29. Nh5 Rd7 30. Bh6 Rf7 31. Bd2 e5 32. Kh3 g6
33. Ng3 Kg7 34. Nf1 Nfe6 35. Kg3 Rc2 36. Qb8 Rd7 37. Qb4 g5 38. Bxg5 Nxg5
39. Qxa4 Rdd2 40. Nxd2 Rxd2 41. Qb4 Rd7 42. a4 Nge6 43. a5 Nd4 44. Qb7 Rf7
45. a6 Ncb5 46. Qb6 Rd7 47. f4 Na7 48. fxe5 fxe5 49. Kf2 Nac6 50. g4 h5 51.
gxh5 Rf7+ 52. Ke3 Kh6 53. Kd3 Kxh5 54. Kc4 Ra7 55. Kc5 Kh4 56. Kd6 Kg4 57.
Qxa7 Nxa7 58. Kxe5 Ndc6+ 59. Ke6 Nb4 60. e5 Nxa6 61. Kf7 Nc8 62. h3+ Kh5
63. e6 Nc7 64. Kf6 Nd5+ 65. Ke5 Nde7 66. Kf6 Nc6 67. Kf7 Kh4 68. Kg6 N8e7+
69. Kf7 Nd5 70. Ke8 Kg5 71. h4+ Kh5 72. Kd7 Nce7 73. Kd8 Nf5 74. Ke8 Nf4
75. Kf7 Ng6 76. Kf6 Nfe7 77. Kf7 Nc6 78. Ke8 Kh6 79. h5 Nge7 80. Kf7 Kh7
81. Kf8 Kh8 82. Kf7 Nf5 83. Kf6 Nce7 84. Kf7 Kh7 85. Kf8 Kh6 86. Ke8 Kg7
87. Kd8 Kh8 88. Ke8 Kg8 89. h6 Kh7 90. Kd8 Nc6+ 91. Ke8 Ne5 92. Kf8 Ng6+
93. Ke8 Nfe7 94. Kf7 Nc8 95. Ke8 Nge7 96. Kd7 Kg8 97. Kd8 Kh8 98. Kc7 Na7
99. Kd7 Nd5 100. h7 Nf6+ 101. Ke7 Ne4 102. Kf8 Ng5 103. e7 Ne6+ 104. Kf7
Nc7 105. e8=Q+ Nxe8 106. Kxe8 Kxh7
{Insufficient material} 1/2-1/2
