Kayra Tests

Discussion of computer chess matches and engine tournaments.

Moderator: Ras

mehmet123
Posts: 682
Joined: Sun Jan 26, 2020 10:38 pm
Location: Turkey
Full name: Mehmet Karaman

Kayra Tests

Post by mehmet123 »

Kayra vs Stockfish:

Program Elo + - Games Score Av.Op. Draws

1 Kayra 1.2 bmi2 : 2402 5 4 640 50.5 % 2398 97.2 %
2 Stockfish 190322 x64 bmi2 : 2398 4 5 640 49.5 % 2402 97.2 %

Individual statistics:

1 Kayra 1.2 bmi2 : 2402 640 (+ 12,=622,- 6), 50.5 %

Stockfish 190322 x64 bmi2 : 640 (+ 12,=622,- 6), 50.5 %

2 Stockfish 190322 x64 bmi2 : 2398 640 (+ 6,=622,- 12), 49.5 %

Kayra 1.2 bmi2 : 640 (+ 6,=622,- 12), 49.5 %


Game Conditions: Cutechess Gui, 1 Core (Core-i7 9750h), 5 min TC, SuperGM 4 moves Opening Book, 128 Mb Hash, Ponder Off
https://www.mediafire.com/file/4kzwg0wn ... 7.pgn/file
mehmet123
Posts: 682
Joined: Sun Jan 26, 2020 10:38 pm
Location: Turkey
Full name: Mehmet Karaman

Re: Kayra Tests

Post by mehmet123 »

Kayra 1.2 vs Kayra 1.1:

Program Elo + - Games Score Av.Op. Draws

1 Kayra 1.2 bmi2 : 2406 8 7 640 51.7 % 2394 92.5 %
2 Kayra 1.1 bmi2 : 2394 7 8 640 48.3 % 2406 92.5 %


Individual statistics:

1 Kayra 1.2 bmi2 : 2406 640 (+ 35,=592,- 13), 51.7 %

Kayra 1.1 bmi2 : 640 (+ 35,=592,- 13), 51.7 %

2 Kayra 1.1 bmi2 : 2394 640 (+ 13,=592,- 35), 48.3 %

Kayra 1.2 bmi2 : 640 (+ 13,=592,- 35), 48.3 %


Game Conditions: Cutechess Gui, 1 Core (Core-i7 9750h), 5 min TC, SuperGM 4 moves Opening Book, 128 Mb Hash, Ponder Off
https://www.mediafire.com/file/3ib3wgcx ... 8.pgn/file
Guenther
Posts: 4718
Joined: Wed Oct 01, 2008 6:33 am
Location: Regensburg, Germany
Full name: Guenther Simon

Re: Kayra Tests

Post by Guenther »

mehmet123 wrote: Tue Mar 22, 2022 3:58 pm Kayra vs Stockfish:

Program Elo + - Games Score Av.Op. Draws

1 Kayra 1.2 bmi2 : 2402 5 4 640 50.5 % 2398 97.2 %
2 Stockfish 190322 x64 bmi2 : 2398 4 5 640 49.5 % 2402 97.2 %

Individual statistics:

1 Kayra 1.2 bmi2 : 2402 640 (+ 12,=622,- 6), 50.5 %

Stockfish 190322 x64 bmi2 : 640 (+ 12,=622,- 6), 50.5 %

2 Stockfish 190322 x64 bmi2 : 2398 640 (+ 6,=622,- 12), 49.5 %

Kayra 1.2 bmi2 : 640 (+ 6,=622,- 12), 49.5 %


Game Conditions: Cutechess Gui, 1 Core (Core-i7 9750h), 5 min TC, SuperGM 4 moves Opening Book, 128 Mb Hash, Ponder Off
https://www.mediafire.com/file/4kzwg0wn ... 7.pgn/file
You haven't learnt much yet since last time in how to conduct useful tests.

You don't play pairwise openings, so its complete luck, if White/Black gets a very bad opening. (Sylwy style)
You play a sudden death time control, which leads to random wins in low time.
You play concurrency of 10 with just 12 hyper threads (and whatever is running simultaneously too, which kills the cpu)
,which helps even more to randomly lose at sudden death when time is extremely low despite complete draw position.
Of course this also manifests in strange low depths at random points.

So this was a first meaningless test...

Just one example for any of those non reverted lucky openings:
[pgn][Event "My Tournament"]
[Site "?"]
[Date "2022.03.21"]
[Round "50"]
[White "Kayra 1.2 bmi2"]
[Black "Stockfish 190322 x64 bmi2"]
[Result "1-0"]
[ECO "A04"]
[GameDuration "00:09:58"]
[GameEndTime "2022-03-21T08:21:40.885 Türkiye Standart Saati"]
[GameStartTime "2022-03-21T08:11:42.392 Türkiye Standart Saati"]
[Opening "Reti Opening"]
[PlyCount "201"]
[TimeControl "300"]

1. Nf3 {book} d6 {book}
2. d4 {book} Bg4 {book}
3. e4 {book} Nf6 {book}
4. h3 {book} Bh5 {book}
5. Nc3 {+1.31/27 11} e6 {-1.04/25 9.4}
6. Qe2 {+1.21/23 3.3} Bg6 {-1.08/27 11}
7. h4 {+1.30/23 4.7} h5 {-1.18/29 35}
8. Bg5 {+1.41/23 4.6} c6 {-1.02/23 3.2}
9. O-O-O {+1.43/22 3.5} Qa5 {-1.36/29 11}
10. Nd2 {+1.45/24 15} Nbd7 {-1.04/24 3.5}
11. f4 {+1.31/27 16} O-O-O {-0.99/24 3.3}
12. Nc4 {+1.17/25 2.8} Qc7 {-0.83/24 4.0}
13. Qf3 {+1.36/24 4.0} b5 {-1.15/27 7.4}
14. Nd2 {+1.20/26 12} d5 {-1.30/30 16}
15. exd5 {+1.19/26 4.5} cxd5 {-1.27/24 2.2}
16. f5 {+1.52/29 15} Bxf5 {-0.81/26 3.5}
17. Bf4 {+1.43/28 3.7} Bd6 {-1.00/29 11}
18. Bxd6 {+1.42/27 4.0} Qxd6 {-1.23/26 4.1}
19. Nxb5 {+1.46/26 4.4} Qb6 {-1.15/28 3.3}
20. Qa3 {+1.25/29 7.7} Kb8 {-1.15/28 3.7}
21. Nd6 {+1.31/27 4.2} Ka8 {-1.15/28 3.6}
22. c4 {+1.38/29 7.9} Rb8 {-1.19/27 9.0}
23. c5 {+1.38/26 4.0} Qb4 {-1.20/24 2.0}
24. Qxb4 {+1.34/26 3.4} Rxb4 {-1.25/26 3.1}
25. Bb5 {+1.28/24 3.8} Nb8 {-1.23/26 5.2}
26. Nxf7 {+1.56/24 3.6} Rc8 {-1.15/28 12}
27. Nd6 {+1.34/26 5.0} Rc7 {-0.96/24 2.5}
28. a3 {+1.45/27 24} Rxd4 {-1.37/27 6.5}
29. b4 {+1.50/23 2.2} Rg4 {-1.42/23 2.8}
30. Rhg1 {+1.57/24 3.7} Rxh4 {-1.33/24 11}
31. Nf3 {+1.41/22 2.3} Rf4 {-1.71/27 9.5}
32. Nd4 {+1.71/24 6.0} Re7 {-1.58/24 5.4}
33. Rgf1 {+1.70/24 11} Rxf1 {-1.66/26 5.9}
34. Rxf1 {+2.19/22 2.3} g6 {-1.88/21 2.2}
35. Re1 {+2.11/24 5.2} Ne4 {-2.04/21 2.4}
36. Kb2 {+2.13/22 2.6} h4 {-1.95/28 20}
37. Ba4 {+2.41/26 12} a6 {-2.00/21 1.0}
38. Bc2 {+2.87/23 1.8} Nxd6 {-1.99/27 7.1}
39. cxd6 {+2.50/23 5.8} Rd7 {-2.08/23 0.96}
40. Nxf5 {+2.25/24 2.9} exf5 {-2.15/24 2.8}
41. Re6 {+2.69/23 2.9} Kb7 {-2.19/25 9.5}
42. Rxg6 {+2.13/24 4.9} f4 {-1.90/22 1.3}
43. Bf5 {+2.59/24 4.4} Rd8 {-1.64/21 1.2}
44. Rh6 {+2.32/24 2.4} Nc6 {-2.57/23 6.0}
45. d7 {+2.40/22 1.9} Ne5 {-2.02/21 2.0}
46. Be6 {+2.48/21 1.4} Nxd7 {-2.53/23 4.0}
47. Bxd5+ {+2.66/21 1.5} Kc7 {-2.37/22 1.9}
48. Kc3 {+2.75/23 1.5} Re8 {-2.12/20 1.0}
49. Rc6+ {+3.00/22 1.6} Kd8 {-2.91/22 2.1}
50. Rxa6 {+3.36/21 1.6} Re3+ {-2.91/19 0.74}
51. Kd4 {+3.17/23 1.8} h3 {-2.60/21 1.1}
52. gxh3 {+3.20/21 1.9} Rxh3 {-3.30/20 1.9}
53. Be6 {+3.20/21 1.1} Rh4 {-3.67/26 6.7}
54. Bxd7 {+4.39/25 1.7} Kxd7 {-3.98/24 0.66}
55. Ke4 {+4.52/25 1.5} Kc7 {-4.42/25 1.5}
56. b5 {+4.53/32 2.5} Kb8 {-4.46/24 0.86}
57. a4 {+4.54/30 1.8} Kc7 {-4.47/22 0.24}
58. Kf3 {+4.82/28 1.3} Kb7 {-5.69/22 1.7}
59. Rf6 {+4.83/28 1.2} Rh3+ {-4.23/18 0.28}
60. Kg4 {+4.86/30 2.0} Rh1 {-4.90/22 1.4}
61. Rxf4 {+4.97/27 1.8} Re1 {-5.79/21 2.2}
62. Kf5 {+5.21/28 0.95} Kb6 {-5.89/20 0.38}
63. Re4 {+5.22/28 1.1} Ra1 {-5.89/16 0.10}
64. Ke5 {+6.48/26 3.5} Rd1 {-5.52/17 0.43}
65. Rd4 {+6.45/26 1.2} Re1+ {-6.20/21 0.25}
66. Kd5 {+6.46/27 1.0} Ra1 {-6.20/22 0.15}
67. Re4 {+6.99/22 0.68} Rd1+ {-6.65/24 0.69}
68. Kc4 {+7.00/23 0.83} Rc1+ {-6.66/18 0.16}
69. Kd3 {+7.02/24 1.3} Kc5 {-6.66/17 0.12}
70. Re5+ {+8.21/23 0.84} Kd6 {-6.69/18 0.11}
71. Rh5 {+9.98/21 1.7} Ra1 {-6.71/18 0.058}
72. Rh6+ {+9.98/19 0.81} Kc5 {-6.82/17 0.10}
73. Rc6+ {+10.17/20 0.59} Kb4 {-6.33/16 0.036}
74. Ra6 {+10.19/22 0.65} Kc5 {-6.33/14 0.13}
75. b6 {+10.35/21 0.61} Kc6 {-5.59/13 0.071}
76. a5 {+10.61/21 0.59} Rd1+ {-6.33/14 0.083}
77. Kc4 {+11.07/20 0.82} Rc1+ {-8.59/14 0.27}
78. Kb3 {+11.52/19 0.92} Rh1 {-7.87/14 0.12}
79. Ra7 {+11.52/18 0.63} Rb1+ {-8.57/12 0.034}
80. Kc3 {+11.76/17 0.32} Rb5 {-9.16/14 0.060}
81. Kc4 {+12.39/15 0.40} Rb1 {-9.27/13 0.017}
82. Rc7+ {+11.51/18 2.2} Kd6 {-9.10/12 0.012}
83. Ra7 {+10.23/16 0.44} Kc6 {-9.28/15 0.061}
84. Rf7 {+12.23/15 0.19} Rc1+ {-9.23/12 0.033}
85. Kd4 {+12.65/15 0.21} Rb1 {-5.12/11 0.009}
86. Rc7+ {+12.70/17 0.31} Kb5 {-5.40/10 0.003}
87. Rc8 {+13.41/18 1.3} Ka6 {-8.69/10 0.012}
88. Kd5 {+13.42/14 0.098} Rd1+ {-8.91/11 0.001}
89. Ke5 {+13.42/15 0.14} Rg1 {-10.17/9 0.006}
90. Ra8+ {+13.71/14 0.14} Kb7 {-8.74/9 0.006}
91. Ra7+ {+13.86/14 0.066} Kc8 {-12.61/7 0.002}
92. Rc7+ {+15.63/14 0.17} Kd8 {-8.03/5 0}
93. Kd5 {+16.59/12 0.097} Rg5+ {-10.71/6 0.001}
94. Kc6 {+54.50/20 0.044} Rg6+ {-56.96/6 0.002}
95. Kb7 {+72.11/19 0.063} Rg1 {-14.03/8 0.001}
96. a6 {+M19/30 0.041} Rg4 {-12.75/4 0}
97. a7 {+M9/49 0.040} Ra4 {-M8/18 0.001}
98. a8=Q+ {+M7/66 0.036} Rxa8 {-55.78/1 0}
99. Kxa8 {+M5/245 0.021} Ke8 {-55.78/1 0}
100. b7 {+M3/245 0.012} Kf8 {-M2/74 0.003}
101. b8=Q# {+M1/245 0.009}
{White mates}
1-0[/pgn]

An example of stupid sudden depth time control craziness:
[pgn]
[Event "My Tournament"]
[Site "?"]
[Date "2022.03.20"]
[Round "11"]
[White "Stockfish 190322 x64 bmi2"]
[Black "Kayra 1.2 bmi2"]
[Result "0-1"]
[ECO "B51"]
[GameDuration "00:09:59"]
[GameEndTime "2022-03-20T22:43:40.525 Türkiye Standart Saati"]
[GameStartTime "2022-03-20T22:33:41.192 Türkiye Standart Saati"]
[Opening "Sicilian"]
[PlyCount "266"]
[TimeControl "300"]
[Variation "Canal-Sokolsky (Nimzovich-Rossolimo, Moscow) attack"]

1. e4 {book} c5 {book}
2. Nf3 {book} d6 {book}
3. Bb5+ {book} Nd7 {book}
4. O-O {book} Nf6 {book}
5. Re1 {+0.78/27 7.8} a6 {-0.69/31 24}
6. Bf1 {+0.76/25 3.7} b6 {-0.75/26 2.6}
7. d4 {+0.85/28 4.9} cxd4 {-0.66/28 3.9}
8. Nxd4 {+0.75/30 4.3} Bb7 {-0.64/28 3.8}
9. f3 {+0.70/32 7.4} e6 {-0.62/28 4.4}
10. c4 {+0.83/28 4.4} Be7 {-0.53/28 3.9}
11. Nc3 {+0.76/28 4.7} O-O {-0.68/27 5.7}
12. Be3 {+0.82/25 3.8} Re8 {-0.75/25 4.9}
13. Rc1 {+0.62/27 13} Ne5 {-0.69/28 15}
14. Na4 {+0.71/26 11} Nfd7 {-0.55/25 3.2}
15. Qd2 {+0.63/24 3.5} Rc8 {-0.65/25 9.0}
16. h3 {+0.62/28 28} Qc7 {-0.51/24 4.4}
17. b3 {+0.70/26 11} Ng6 {-0.61/24 10}
18. Nc3 {+0.64/23 2.9} Qb8 {-0.62/24 3.0}
19. Red1 {+0.62/23 4.2} Red8 {-0.56/26 23}
20. Qf2 {+0.66/24 12} Bf6 {-0.66/24 8.6}
21. Bd3 {+0.62/24 11} Ba8 {-0.65/24 2.8}
22. Bb1 {+0.66/21 2.4} Nge5 {-0.67/22 3.7}
23. Qc2 {+0.59/21 3.0} Ng6 {-0.62/24 7.4}
24. Qd2 {+0.71/22 5.8} Ndf8 {-0.63/24 8.4}
25. Qf2 {+0.73/22 4.6} Bh4 {-0.57/23 5.2}
26. Qe2 {+0.82/25 2.8} Bf6 {-0.63/23 1.9}
27. Qf2 {+0.76/23 3.2} Nd7 {-0.61/26 16}
28. Kh1 {+0.76/24 2.7} Bh4 {-0.71/21 2.1}
29. Qd2 {+0.80/28 20} Bf6 {-0.72/22 2.9}
30. Nde2 {+0.80/25 1.6} Be7 {-0.62/21 2.4}
31. a4 {+0.88/22 3.9} Nf6 {-0.56/21 3.1}
32. Nf4 {+0.71/23 4.7} Ne5 {-0.70/23 8.6}
33. Qf2 {+0.65/23 3.6} Nfd7 {-0.60/22 1.9}
34. Nfe2 {+0.61/23 2.6} Nc6 {-0.66/25 16}
35. Nd4 {+0.69/21 2.1} Bf6 {-0.50/23 5.9}
36. Nxc6 {+0.78/23 2.4} Bxc6 {-0.40/21 1.7}
37. Qc2 {+0.66/20 2.4} g6 {-0.48/21 2.2}
38. Qd2 {+0.54/24 11} Nc5 {-0.31/23 2.6}
39. Bc2 {+0.50/21 1.3} a5 {-0.33/19 2.3}
40. Bg5 {+0.50/23 8.7} Bxg5 {-0.19/20 1.7}
41. Qxg5 {+0.44/24 1.2} Rd7 {-0.28/21 1.7}
42. Rd2 {+0.34/22 2.2} Re8 {-0.23/24 5.2}
43. Re1 {+0.32/23 4.5} Qd8 {-0.13/23 7.0}
44. Qe3 {+0.24/20 1.4} e5 {-0.25/19 1.5}
45. Rb1 {+0.29/18 1.6} h5 {-0.12/22 1.3}
46. Qe1 {+0.30/22 4.4} Kg7 {-0.20/18 1.8}
47. h4 {+0.22/20 2.5} Rg8 {0.00/20 2.8}
48. Qf2 {+0.18/18 2.6} Qf6 {0.00/22 1.3}
49. Kg1 {+0.16/16 0.92} Kf8 {-0.04/18 1.5}
50. Nd1 {0.00/24 5.6} Kg7 {0.00/20 1.2}
51. Kf1 {+0.26/20 3.9} Kf8 {-0.20/24 7.9}
52. Ke2 {+0.63/19 0.81} Ke7 {-0.23/18 0.76}
53. Ne3 {+0.24/20 3.1} Kd8 {-0.14/22 3.0}
54. Kf1 {+0.21/22 9.4} Kc7 {0.00/18 0.73}
55. b4 {+0.12/16 0.78} axb4 {-0.08/19 1.3}
56. Rxb4 {+0.27/18 0.90} Ra8 {0.00/22 2.1}
57. Kg1 {+0.22/18 0.76} Rdd8 {0.00/20 0.55}
58. Rb1 {+0.13/19 1.7} Rdb8 {0.00/24 2.9}
59. Kh2 {+0.24/18 1.5} Ra5 {-0.14/20 3.4}
60. Rbd1 {+0.20/19 1.5} Rd8 {-0.15/18 0.57}
61. g3 {+0.15/19 2.6} Ra7 {-0.08/19 2.6}
62. Kg2 {+0.22/20 2.7} Ra5 {-0.13/17 0.91}
63. Nd5+ {+0.50/17 1.3} Bxd5 {-0.24/17 0.56}
64. Rxd5 {+0.40/16 0.40} Nb7 {-0.26/19 2.3}
65. Rb1 {+0.49/17 0.45} Nc5 {-0.04/16 0.42}
66. Rbd1 {+0.09/22 3.4} Nb7 {-0.04/18 0.70}
67. Rb1 {+0.09/17 0.24} Nc5 {0.00/19 0.62}
68. Qd2 {+0.21/21 1.3} Qe7 {-0.02/19 0.59}
69. Rf1 {+0.15/18 0.30} Rg8 {0.00/22 0.99}
70. f4 {+0.10/17 0.29} f6 {0.00/20 0.27}
71. Rd1 {+0.18/16 0.27} Rd8 {-0.02/17 0.34}
72. Qf2 {+0.17/19 2.0} Raa8 {0.00/21 1.1}
73. Kh2 {+0.18/17 0.50} Qf8 {0.00/19 0.91}
74. Rf1 {+0.09/19 1.0} Qe7 {0.00/19 0.23}
75. Rfd1 {+0.13/17 0.30} Rd7 {0.00/21 0.64}
76. Qe3 {+0.09/19 0.86} Rdd8 {0.00/22 0.38}
77. R1d2 {+0.07/15 0.13} Ra5 {0.00/15 0.55}
78. Bd1 {+0.17/16 0.54} Qg7 {0.00/19 0.62}
79. Bc2 {+0.16/17 0.23} Qe7 {0.00/21 0.24}
80. Rd1 {+0.19/16 0.090} Raa8 {0.00/21 0.29}
81. Kg2 {+0.19/14 0.14} Ra5 {0.00/17 0.26}
82. Qf3 {+0.28/16 0.63} Qg7 {0.00/20 0.26}
83. Rb1 {+0.24/12 0.081} Ra6 {0.00/21 0.75}
84. Qe2 {+0.04/19 1.3} Raa8 {0.00/19 0.63}
85. Qd2 {+0.36/12 0.082} Qe7 {0.00/18 0.13}
86. Ra1 {0.00/19 0.70} Ra5 {0.00/17 0.18}
87. Rf1 {0.00/15 0.027} Rf8 {0.00/18 0.16}
88. Rd1 {0.00/17 0.078} Rd8 {0.00/20 0.16}
89. fxe5 {+0.18/17 0.16} fxe5 {-0.09/19 0.20}
90. Qh6 {+0.19/14 0.041} Qe6 {-0.08/22 0.28}
91. Rf1 {0.00/17 0.22} Raa8 {0.00/17 0.39}
92. Bd1 {0.00/15 0.031} Kb7 {-0.01/16 0.16}
93. Qg7+ {0.00/12 0.046} Rd7 {-0.02/15 0.16}
94. Qf6 {+0.11/12 0.047} Qe8 {0.00/15 0.20}
95. Bc2 {0.00/11 0.11} Kc7 {0.00/15 0.13}
96. Rfd1 {0.00/12 0.055} Ne6 {+0.25/14 0.19}
97. Rf1 {0.00/12 0.031} Qg8 {+0.13/13 0.18}
98. c5 {-0.06/12 0.10} Nxc5 {0.00/17 0.81}
99. Kg1 {-0.01/12 0.017} g5 {+1.39/15 0.093}
100. hxg5 {-0.38/12 0.060} Ne6 {+0.94/15 0.13}
101. Bd3 {-0.90/12 0.098} Rf8 {+1.63/16 0.73}
102. Qh6 {-2.18/11 0.018} Qxg5 {+0.61/17 0.73}
103. Rc1+ {-0.45/9 0.005} Kd8 {+0.75/15 0.065}
104. Qxg5+ {-0.48/10 0.011} Nxg5 {+0.61/12 0.037}
105. Rb1 {-0.59/9 0.009} h4 {+0.98/12 0.042}
106. Bb5 {-0.79/11 0.027} Nf3+ {+1.87/12 0.037}
107. Kh1 {-1.00/9 0.006} Rc7 {+1.58/13 0.15}
108. Rxd6+ {-0.47/10 0.003} Ke7 {+1.77/10 0.031}
109. Rc6 {-0.34/8 0.003} Rxc6 {+2.48/12 0.027}
110. Bxc6 {-1.44/9 0.005} Nd4 {+1.92/12 0.070}
111. gxh4 {-1.49/7 0.002} Nxc6 {+2.29/11 0.030}
112. Rxb6 {-1.82/9 0.006} Kd6 {+2.30/9 0.018}
113. Kg2 {-1.84/3 0} Rf4 {+2.38/10 0.031}
114. h5 {-1.70/4 0} Rh4 {+3.90/11 0.028}
115. a5 {-3.04/7 0.002} Rxh5 {+4.33/9 0.013}
116. a6 {-3.78/5 0.002} Rh7 {+4.08/11 0.028}
117. Kg3 {-3.11/3 0} Ra7 {+4.45/10 0.012}
118. Kg4 {-2.81/4 0} Kc5 {+4.39/10 0.019}
119. Rb1 {-3.80/6 0.002} Rxa6 {+4.32/10 0.017}
120. Rc1+ {-3.93/5 0.001} Kd6 {+4.41/9 0.010}
121. Rxc6+ {-3.88/5 0} Rxc6 {+10.98/10 0.010}
122. Kf5 {-8.47/6 0.002} Rc4 {+54.06/13 0.047}
123. Kg4 {-54.78/9 0.002} Rxe4+ {+54.33/11 0.009}
124. Kf3 {-54.90/8 0.001} Rc4 {+55.59/15 0.025}
125. Ke3 {-54.81/7 0.002} e4 {+55.59/12 0.003}
126. Ke2 {-54.87/3 0} Kd5 {+55.97/15 0.013}
127. Kd1 {-54.97/1 0} e3 {+65.45/14 0.002}
128. Ke2 {-55.24/6 0.001} Re4 {+M13/12 0.003}
129. Kf3 {-65.76/5 0} e2 {+M11/15 0.002}
130. Kg3 {-65.44/1 0} e1=Q+ {+M9/15 0.002}
131. Kh3 {-M8/6 0} Qc3+ {+M5/22 0.003}
132. Kg2 {-65.49/5 0.001} Re2+ {+M3/23 0.001}
133. Kf1 {-M2/12 0} Qe1# {+M1/12 0.001}
{Black mates}
0-1[/pgn]

I am too bored to post examples with games with too low depths due
to overloading of the cpu...

And despite all of this there still was a draw rate of 97.2% (and being inside the error margin) ;-)
Most people here know what that means and it makes looking at the source 'changes' even more boring.
https://rwbc-chess.de

[Trolls n'existent pas...]
mehmet123
Posts: 682
Joined: Sun Jan 26, 2020 10:38 pm
Location: Turkey
Full name: Mehmet Karaman

Re: Kayra Tests

Post by mehmet123 »

Guenther wrote: Tue Mar 22, 2022 7:29 pm
You don't play pairwise openings, so its complete luck, if White/Black gets a very bad opening. (Sylwy style)
You play a sudden death time control, which leads to random wins in low time.
You play concurrency of 10 with just 12 hyper threads (and whatever is running simultaneously too, which kills the cpu)
,which helps even more to randomly lose at sudden death when time is extremely low despite complete draw position.
Of course this also manifests in strange low depths at random points.
I can say that I rarely use dead time control and Gm 4 moves opening books in my tests. I mostly prefer incremental time controls and Balsa 5 Moves Opening book (500 openings) in my tests.
forum3/viewtopic.php?f=6&t=74518
But for these tests, I wanted to try something different. Mostly ı used concurrency 10 in my tests but since these tests are dead time tests ı set the concurrency 6 for these games.
These are only fun tests. 640 games is not a sufficient number to reach a clear conclusion
User avatar
AlexChess
Posts: 1562
Joined: Sat Feb 06, 2021 8:06 am
Full name: Alex Morales

Re: Kayra Tests

Post by AlexChess »

Image

Dear Mehmet,

Congratulations!

Just compiled Kayra 1.2 also for macOS M1 native (running 4 times faster with 8 CPUs than Windows 11 ARM64) and restarted all the Kayra 1.1 games 3 mins + 2 secs /all . (Now playing against ProteusChessSFV005) ...I will publish crosstable and all PGN games here: forum3/viewtopic.php?f=6&t=79238&sid=22 ... f0ba4a9480 :)

kind regards, Alex

PEACE for Ukraine!
Chess engines and dedicated chess computers fan since 1981 :D macOS Sequoia 16GB-512GB, Windows 11 & Ubuntu ARM64.
ProteusSF Dev Forum
mehmet123
Posts: 682
Joined: Sun Jan 26, 2020 10:38 pm
Location: Turkey
Full name: Mehmet Karaman

Re: Kayra Tests

Post by mehmet123 »

Kayra vs Stockfish:

Program Elo + - Games Score Av.Op. Draws

1 Kayra 1.2 bmi2 : 2411 23 22 450 53.1 % 2389 51.1 %
2 Stockfish 120322 x64 bmi2 : 2389 22 23 450 46.9 % 2411 51.1 %


Individual statistics:

1 Kayra 1.2 bmi2 : 2411 450 (+124,=230,- 96), 53.1 %

Stockfish 120322 x64 bmi2 : 450 (+124,=230,- 96), 53.1 %

2 Stockfish 120322 x64 bmi2 : 2389 450 (+ 96,=230,-124), 46.9 %

Kayra 1.2 bmi2 : 450 (+ 96,=230,-124), 46.9 %


Game Conditions: Cutechess Gui, 1 Core (Core-i7 9750h), 5 min TC, Drawkiller Unbalanced V4 4 moves Opening Book, 128 Mb Hash, Ponder Off,
https://www.mediafire.com/file/dufpjp7e ... 0.pgn/file


This is the first time ı used an unbalanced opening book in Kayra test and i got a strange result. It would be ridiculous to claim that Kayra is + 22 elo stronger than Stockfish. It is not easy to determine whether this elo difference is only due to lucky openings.
Guenther
Posts: 4718
Joined: Wed Oct 01, 2008 6:33 am
Location: Regensburg, Germany
Full name: Guenther Simon

Re: Kayra Tests

Post by Guenther »

mehmet123 wrote: Wed Mar 23, 2022 6:09 pm Kayra vs Stockfish:

Program Elo + - Games Score Av.Op. Draws

1 Kayra 1.2 bmi2 : 2411 23 22 450 53.1 % 2389 51.1 %
2 Stockfish 120322 x64 bmi2 : 2389 22 23 450 46.9 % 2411 51.1 %


Individual statistics:

1 Kayra 1.2 bmi2 : 2411 450 (+124,=230,- 96), 53.1 %

Stockfish 120322 x64 bmi2 : 450 (+124,=230,- 96), 53.1 %

2 Stockfish 120322 x64 bmi2 : 2389 450 (+ 96,=230,-124), 46.9 %

Kayra 1.2 bmi2 : 450 (+ 96,=230,-124), 46.9 %


Game Conditions: Cutechess Gui, 1 Core (Core-i7 9750h), 5 min TC, Drawkiller Unbalanced V4 4 moves Opening Book, 128 Mb Hash, Ponder Off,
https://www.mediafire.com/file/dufpjp7e ... 0.pgn/file


This is the first time ı used an unbalanced opening book in Kayra test and i got a strange result. It would be ridiculous to claim that Kayra is + 22 elo stronger than Stockfish. It is not easy to determine whether this elo difference is only due to lucky openings.
Bravo! Here (six days ago) you tested just 2 single opening lines (one more winning, one more drawish in sudden death) and even
those were not distributed equally in those 450 games (239:211)... Do you ever look at what you post here?

The only used two opening lines in 450 games:
[pgn][Event "My Tournament"]
[Site "?"]
[Date "2022.03.17"]
[Round "1"]
[White "Stockfish 120322 x64 bmi2"]
[Black "Kayra 1.2 bmi2"]
[Result "1-0"]
[ECO "A00"]
[GameDuration "00:09:53"]
[GameEndTime "2022-03-17T06:43:29.688 Türkiye Standart Saati"]
[GameStartTime "2022-03-17T06:33:36.233 Türkiye Standart Saati"]
[Opening "Mieses Opening"]
[PlyCount "175"]
[TimeControl "300"]

1. d3 {book} e6 {book}
2. Be3 {book} Bd6 {book}
3. Nc3 {book} Nh6 {book}
4. Qd2 {book} O-O {book}
5. Bxh6 {+2.13/26 8.3}[/pgn]

[pgn][Event "My Tournament"]
[Site "?"]
[Date "2022.03.17"]
[Round "1"]
[White "Kayra 1.2 bmi2"]
[Black "Stockfish 120322 x64 bmi2"]
[Result "1/2-1/2"]
[ECO "A00"]
[GameDuration "00:08:20"]
[GameEndTime "2022-03-17T06:41:56.370 Türkiye Standart Saati"]
[GameStartTime "2022-03-17T06:33:36.202 Türkiye Standart Saati"]
[Opening "Van't Kruijs Opening"]
[PlyCount "98"]
[TimeControl "300"]

1. e3 {book} d6 {book}
2. Nh3 {book} Nc6 {book}
3. Bd3 {book} Be6 {book}
4. O-O {book} Qd7 {book}
5. Nf4 {+0.81/28 7.8}[/pgn]

I don't believe anymore you are doing this unintentionally (if not this must be the worst test ever posted at talkchess).
This is just trolling and wasting our time and it will be reported.
https://rwbc-chess.de

[Trolls n'existent pas...]
Modern Times
Posts: 3710
Joined: Thu Jun 07, 2012 11:02 pm

Re: Kayra Tests

Post by Modern Times »

Where are the Windows binaries for this ?
mehmet123
Posts: 682
Joined: Sun Jan 26, 2020 10:38 pm
Location: Turkey
Full name: Mehmet Karaman

Re: Kayra Tests

Post by mehmet123 »

Guenther wrote: Wed Mar 23, 2022 7:34 pm
I don't believe anymore you are doing this unintentionally (if not this must be the worst test ever posted at talkchess).
This is just trolling and wasting our time and it will be reported.
I hope I do not enter the top 10 troll list after these Kayra tests and I hope citing sources will save me time :)

http://outskirts.altervista.org/forum/v ... &start=190
Modern Times
Posts: 3710
Joined: Thu Jun 07, 2012 11:02 pm

Re: Kayra Tests

Post by Modern Times »

I've run my own small test on Kayra out of curiosity.

Ryzen 9 5900X
Cutechess GUI, concurrency 12
Ponder off, engines 1 thread each, 256MB hash each
Time control 2'+1"

For the very first time I used an unbalanced book - Pohl's UHO_2022_6mvs_+110_+119.pgn comprising 3000+ openings. Whether that makes the results less or more valid, people have their own opinions. Cutechess will have made random selections from that, with each opening played reversed sides

Kayra did not get the better of Stockfish:

Code: Select all

1   Stockfish 2022-03-19 64-bit   +10  +150/=493/-127 51.49%  396.5/770
2   Kayra 1.2 64-bit              -10  +127/=493/-150 48.51%  373.5/770