cutechess-cli

[michiguel]

After building from the sources, I had to figure out that some libraries were generated and had to be installed in the system. It will be much easier if this is already included in the Makefile. Did I miss something? I copy them to /usr/local/lib and executed ldconfig. Is that all? it seems to work.

That works, but it's not necessary. You could also run "export LD_LIBRARY_PATH=path_to_cutechess_dir/projects/lib". Or you could link to libchess statically by running "qmake -config static" followed by "make".

It there a way to tell all engines to ponder on or off at once?

Unfortunately pondering isn't supported yet.

If I did not say this before, thanks a lot for this tool.

I ran a match to 100,000 games (just to get short pgn files to build a test set, I made the engines resign after 10 moves) and I got this error

Gaviota_1 wins the game as black
Score of Gaviota_1 vs Gaviota_2: 10109 - 10353 - 0
Started game 20463 of 100000
pure virtual method called
terminate called without an active exception

real 16m53.211s
user 0m19.073s
sys 0m21.729s

Miguel

BTW why is the score not evenly distributed after the first 20462 games?
Did they not resign in alternate mode or always for the same colour?

Guenther

Good observation!

Sometimes they got out of book earlier (For instance, when there is only one move in the book file that lead to a defeat). I and made them resign there.

Miguel

[jarkkop]

Can you also give starting positions in FEN format e.g. 1000 epd positions in a file?

This could be useful if you want to run endgame tournament.

[ilari]

Can you also give starting positions in FEN format e.g. 1000 epd positions in a file?

This could be useful if you want to run endgame tournament.

Yes. EPD isn't supported yet, but FEN is, so you can use a PGN file that looks like this:

Code: Select all

[FEN "fen 1"]
*

[FEN "fen 2"]
*

[FEN "fen n"]
*

Cutechess-cli needs the termination marker (*) to know when a game has ended and a new one begins.

[jarkkop]

Why is Fire wasting 0.12s move 10 when we are still in book(12% of its total time for 40 moves) and repeating the starting position? (tc=40/1) Note: Toga had 4 times more time allocated to 40 moves.

[Event "?"]
[Site "?"]
[Date "2010.06.02"]
[Round "2"]
[White "Fire 1.3 w32"]
[Black "Toga II 1.4 beta5c"]
[Result "0-1"]
[PlyCount "136"]
[EventDate "2010.??.??"]

1. c4 {book} Nf6 {book} 2. d4 {book} e6 {book} 3. Nf3 {book} b6 {book} 4. Nc3 {
book} Bb7 {book} 5. a3 {book} d5 {book} 6. Bg5 {book} Be7 {book} 7. Qa4+ {book}
c6 {book} 8. Bxf6 {book} Bxf6 {book} 9. cxd5 {book} exd5 {book} 10. g3 {book}
O-O {+0.12/9 0.12s} 11. Bg2 {book} Nd7 {book} 12. O-O {+0.16/8 0.007s} Re8 {
book} 13. Rfd1 {book} h5 {+0.23/8 0.081s} 14. e3 {+0.25/8 0.006s} g5 {
+0.15/8 0.11s} 15. Qc2 Qe7 {+0.15/8 0.085s} 16. Rac1 {+0.40/8 0.018s} g4 {
+0.31/9 0.10s} 17. Ne1 {+0.44/8 0.011s} h4 {+0.21/8 0.067s} 18. Nd3 {
+0.44/8 0.010s} h3 {+0.34/8 0.065s} 19. Bh1 {+0.33/8 0.015s} Bg5 {
+0.44/8 0.084s} 20. Ne2 Qd6 {+0.42/7 0.054s} 21. Ndf4 Bh6 {+0.39/8 0.16s} 22.
Rd2 Nf6 {+0.35/8 0.099s} 23. Nd3 {+0.33/8 0.019s} Kf8 {+0.12/8 0.14s} 24. Rdd1
Rac8 {+0.21/8 0.13s} 25. Qa4 {+0.25/8 0.018s} a5 {+0.38/9 0.11s} 26. Qb3 {
+0.21/8 0.019s} Nd7 {+0.59/9 0.086s} 27. Nef4 {+0.11/8 0.022s} Ba6 {
+0.55/9 0.099s} 28. Rc3 Bc4 {+0.69/8 0.083s} 29. Qc2 Kg8 {+0.41/8 0.16s} 30. b3
Ba6 {+0.37/8 0.081s} 31. Rc1 Bb7 {+0.34/7 0.11s} 32. Qe2 {+0.27/8 0.020s} f5 {
+0.38/7 0.11s} 33. Qc2 Kh8 {+0.64/7 0.056s} 34. b4 Ra8 {+0.55/7 0.064s} 35. b5
{+0.96/8 0.020s} cxb5 {-0.59/9 0.26s} 36. Rc7 {+0.94/8 0.017s} Rac8 {
-0.27/8 0.17s} 37. Nxd5 Qxa3 {+0.03/6 0.083s} 38. Qd2 Red8 {+0.09/7 0.064s} 39.
Nxb6 Nxb6 {+1.04/6 0.040s} 40. Bxb7 Rxc7 {+1.42/9 0.065s} 41. Rxc7 {
-1.70/8 0.008s} Rxd4 {+1.66/10 0.11s} 42. Rc3 {-1.76/8 0.008s} Qa1+ {
+1.55/10 0.26s} 43. Qe1 {-1.78/8 0.002s} Qxe1+ {+1.65/11 0.11s} 44. Nxe1 Rc4 {
+1.70/12 0.13s} 45. Rxc4 {-2.11/8 0.010s} Nxc4 {+1.76/11 0.096s} 46. Nc2 {
-2.07/8 0.007s} b4 {+2.85/7 0.056s} 47. Nd4 {-1.94/8 0.004s} a4 {+2.26/7 0.078s
} 48. Bd5 {-2.81/8 0.015s} Nxe3 {+2.05/11 0.055s} 49. fxe3 {-3.18/9 0.016s}
Bxe3+ {+4.72/10 0.066s} 50. Kf1 Bxd4 {+5.02/12 0.033s} 51. Be6 {-3.24/9 0.007s}
b3 {+6.11/12 0.11s} 52. Bc4 {-3.60/8 0.009s} b2 {+6.88/10 0.050s} 53. Bd3 {
-5.57/10 0.011s} a3 {+7.72/14 0.071s} 54. Bb1 {-4.60/10 0.006s} Kg7 {
+7.97/15 0.072s} 55. Ba2 {-5.76/10 0.014s} Be5 {+7.93/14 0.080s} 56. Kf2 {
-6.49/10 0.016s} f4 {+8.13/13 0.054s} 57. Kg1 {-8.11/9 0.015s} fxg3 {
+19.51/10 0.075s} 58. hxg3 {-9.84/10 0.016s} Bxg3 {+24.54/10 0.066s} 59. Bb1 {
-8.11/8 0.001s} Kf6 {+24.58/11 0.11s} 60. Ba2 {-9.66/9 0.015s} Bf4 {
+28.61/9 0.098s} 61. Bb1 {-18.60/8 0.005s} g3 {+31.03/8 0.095s} 62. Kf1 {
-21.01/9 0.013s} g2+ {+33.64/8 0.11s} 63. Kf2 {-23.92/8 0.007s} Be3+ {
+M15/8 0.10s} 64. Kf3 {-26.50/8 0.015s} g1=Q {+M11/8 0.061s} 65. Ba2 {
-33.45/8 0.014s} Ke5 {+M7/21 0.080s} 66. Bb1 {-M12/8 0.007s} Kd4 {+M5/63 0.036s
} 67. Ke2 {-M4/15 0.004s} Qf2+ {+M3/63 0.022s} 68. Kd1 Qd2# {
+M1/63 0.020s, Black mates} 0-1

[ilari]

Why is Fire wasting 0.12s move 10 when we are still in book(12% of its total time for 40 moves) and repeating the starting position? (tc=40/1) Note: Toga had 4 times more time allocated to 40 moves.

Were you using a Polyglot opening book? It looks like the game went out of book at move 10, Toga castled, and the game went back into the book. I also noticed that there's a serious bug in the opening book code that was introduced in cutechess-cli 0.2.2. I'll have to fix this asap.