New engine releases 2024 ... 6+3

[Frank Quisinsky]

15. Seer 2.8.0 NN ... v15 later this day available
16. Peacekeeper 2.20 NN ... canceled
17. Drofa 4.1.0 NN ... still running
18. Velvet 6.0.0 NN ... soon

Code: Select all

A)  Wasp 6.50 NN                                         released   = February 28th, 2023    Reference = 3300 Elo
B)  Wasp 6.63 NN dev                                     executable = December 02nd, 2023    Last dev
    John Stanback, USA
------------------------------------------------------------------------------------------------------------------------------------------


12/2023 - 2024


18. Velvet 6.0.0 NN                                      executable = December 21st, 2023    Test on Wasp-1 = soon
    Martin Honert (GER)
    https://github.com/mhonert/velvet-chess/

17. Drofa 4.1.0 NN                                       executable = December 24th, 2023    Test on Wasp-1 = still running
    Alexander Litov & Rhys Rustad-Elliott (RUS, CAN)
    https://github.com/justNo4b/Drofa

------------------------------------------------------------------------------------------------------------------------------------------

16. Peacekeeper 2.20 NN                                  executable = December 24th, 2023    Test on Wasp-1 = canceled
    Kyle Zhang (USA)
    https://github.com/Sazgr/peacekeeper

    - does not want to checkmate. Winning games runs often over 150 moves (3 times after the first 25 games)
    - during the games the Shredder GUI does nothing and the matches hangs (2 times after the first 25 games)
    - engine give advantage for draws in bad bishop endgames,
      50-moves rules will be avoided more times, very long draw games (3 times after the first 25 games)

15. Seer 2.8.0 NN                                        executable = December 31st, 2023    Test on Wasp-1 = v15
    Connor McMonigle (USA)
    https://github.com/connormcmonigle/seer-nnue/

14. Arasan 24.1 NN                                       executable = January 14th, 2024     Test on Wasp-1
    Jon Dart, USA
    https://www.arasanchess.org/index.shtml
    https://github.com/jdart1/arasan-chess

13. Renegade 1.0.0 NN                                    executable = January 13th, 2024     Test on Wasp-1
    Krisztián Peőcz (HUN)
    https://github.com/pkrisz99/Renegade/

12. Minic 3.40 NN                                        executable = January 14th, 2024     Test on Wasp-1
    Vivien CLAUZON, FRA
    https://github.com/tryingsomestuff/Minic

11. Lizard 10.1 NN                                       released   = January 13th, 2024     Test on Wasp-1
(2) Lizard 10.1 NN replaced Lizard 10.0 NN
    Liam McGuire, USA
    https://github.com/liamt19/Lizard

10. Avalanche 2.1.0 NN                                   executable = January 13th, 2024     Test on Wasp 2
    Yinuo Huang, CHN
    https://github.com/SnowballSH/Avalanche

    - during the games the Shredder GUI does nothing and the matches hangs (4 times)

09. RubiChess 20240112 NN                                executable = January 12th, 2024     Test on Wasp 1
    Andreas Matthies, GER
    https://github.com/Matthies/RubiChess/

08. Texel 1.11 NN                                        executable = January 12th, 2024     Test on Wasp-1
    Peter Österlund, SWE
    https://github.com/peterosterlund2/texel

07. Pawn 3.0 NN                                          executable = January 12th, 2024     Test on Wasp-1
    Rui Coelho, POR
    https://github.com/ruicoelhopedro/pawn

06. Counter 5.5 NN                                       released   = January 12th, 2024     Test on Wasp-1
    Vadim Chizhov, RUS
    https://github.com/ChizhovVadim/CounterGo/

05. Caissa 1.16 NN                                       executable = January 11th, 2024     Test on Wasp-2
    Michal Witanowski, POL
    https://github.com/Witek902/Caissa

04. Clarity 4.1.0 NN                                     executable = January 07th, 2024     Test on Wasp-2
    Joseph Pasfield, USA
    https://github.com/Vast342/Clarity

03. Lizard 10.0 NN                                       released   = January 04th, 2024     Test on Wasp-2
(1) Liam McGuire, USA
    https://github.com/liamt19/Lizard

02. Starzix 3.0 NN                                       executable = January 03rd, 2024     Test on Wasp-2
    Ricardo Pinto, POR
    https://github.com/zzzzz151/Starzix

01. Akimbo 0.8.0 NN                                      executable = January 02nd, 2024     Test on Wasp-2
    Jamie Whiting, GBR
    https://github.com/jw1912/akimbo

------------------------------------------------------------------------------------------------------------------------------------------

Watching (I believe not more as ~3150 Elo)

--. Kuma 1.2 NN                                         released   = January 12th, 2024      no test for the moment
    Kato Daichi, JPN
    https://github.com/kato-daichi/kuma

--. Mida 2.3 NN                                         executable = December 27th, 2023     no test for the moment
    Giacomo Porpiglia, ITA
    https://github.com/GiacomoPorpiglia/Mida

[Frank Quisinsky]

15. Seer 2.8.0 NN
16. Peacekeeper 2.2.0 NN ... canceled
17. Velvet 6.0.0 NN ... still running
18. Opsidian 10.0 NN ... soon
It seems that Drofa 4.1.0 is not stronger 3150 Elo, test stopped.

Here the Seer 2.8.0 NN results ...
v.15

Code: Select all

              Name                 Games     Win     Draw     Lose       Pts         S-B         %    wins-55m  lost-55m  AV-wins  AV-draws  AV-al

01. RubiChess 20240112 NN       :   750  :  349+  :  397=  :    4-  :   547.5  : 192778.00  :  73.00%     42       0        79       79       78
02. Caissa 1.16 NN              :   750  :  332+  :  414=  :    4-  :   539.0  : 189134.00  :  71.87%     20       0        85       79       82
03. Seer 2.8.0 NN               :   750  :  312+  :  419=  :   19-  :   521.5  : 182381.00  :  69.53%     19       0        85       81       83
04. Arasan 24.1 NN              :   750  :  202+  :  480=  :   68-  :   442.0  : 155165.75  :  58.93%     24       3        84       82       83
05. Minic 3.40 NN               :   750  :  141+  :  501=  :  108-  :   391.5  : 137649.75  :  52.20%      5      15        95       91       90
06. Pawn 3.0 NN                 :   750  :  139+  :  498=  :  113-  :   388.0  : 136112.25  :  51.73%      5       7        82       88       87
07. Akimbo 0.8.0 NN             :   750  :  149+  :  470=  :  131-  :   384.0  : 133104.50  :  51.20%      5      13        93       92       90
08. Starzix 3.0 NN              :   750  :  123+  :  491=  :  136-  :   368.5  : 127317.25  :  49.13%      3      17        92       91       89
09. Wasp 6.63 NN dev            :   750  :  116+  :  476=  :  158-  :   354.0  : 123724.50  :  47.20%     15       4        80       80       83
10. Texel 1.11 NN               :   750  :  120+  :  463=  :  167-  :   351.5  : 121883.50  :  46.87%     19       1        75       89       87
--------------------------------------------------------------------------------------------------------------------------------------------------
11. Wasp 6.50 NN                :   750  :  114+  :  474=  :  162-  :   351.0  : 121022.75  :  46.80%     27       0        77       81       83
12. Lizard 10.1 NN              :   750  :  112+  :  477=  :  161-  :   350.5  : 122136.50  :  46.73%      7      14        87       94       90
13. Renegade 1.0.0 NN           :   750  :   74+  :  448=  :  228-  :   298.0  : 104042.25  :  39.73%      2      24        91       97       92
14. Clarity 4.1.0 NN            :   750  :   51+  :  465=  :  234-  :   283.5  : 100279.75  :  37.80%      0      17        95       84       85
15. Avalanche 2.1.0 NN          :   750  :   42+  :  350=  :  358-  :   217.0  :  76296.50  :  28.93%      0      57        91       85       82
16. Counter 5.5 NN              :   750  :   35+  :  355=  :  360-  :   212.5  :  75962.75  :  28.33%      3      24        85      101       94


White Wins =  1.586 ( 26.43% )
Draws      =  3.589 ( 59.82% )
Black Wins =    825 ( 13.75% )
Average    = 172.24 ( 86,12 moves )

Code: Select all

   # Player                   :      Elo  Games  Score%  won  draw  lost  Points  Draw%  Error   OppAvg   OppE   OppD
   1 RubiChess 20240112 NN    :  3499.17    750    73.0  349   397     4   547.5   52.9  15.39  3311.85  13.94   15.0
   2 Caissa 1.16 NN           :  3489.38    750    71.9  332   414     4   539.0   55.2  15.93  3312.50  13.90   15.0
   3 Seer 2.8.0 NN            :  3469.68    750    69.5  312   419    19   521.5   55.9  15.67  3313.82  13.92   15.0
   4 Arasan 24.1 NN           :  3387.75    750    58.9  202   480    68   442.0   64.0  13.02  3319.28  14.10   15.0
   5 Minic 3.40 NN            :  3338.80    750    52.2  141   501   108   391.5   66.8  13.49  3322.54  14.07   15.0
   6 Pawn 3.0 NN              :  3335.45    750    51.7  139   498   113   388.0   66.4  13.42  3322.76  14.07   15.0
   7 Akimbo 0.8.0 NN          :  3331.62    750    51.2  149   470   131   384.0   62.7  13.58  3323.02  14.06   15.0
   8 Starzix 3.0 NN           :  3316.89    750    49.1  123   491   136   368.5   65.5  12.72  3324.00  14.12   15.0
   9 Wasp 6.63 NN dev         :  3302.88    750    47.2  116   476   158   354.0   63.5  13.35  3324.94  14.07   15.0
  10 Texel 1.11 NN            :  3300.48    750    46.9  120   463   167   351.5   61.7  13.42  3325.10  14.07   15.0
  11 Wasp 6.50 NN             :  3300.00    750    46.8  114   474   162   351.0   63.2  13.25  3325.13  14.08   15.0
  12 Lizard 10.1 NN           :  3299.52    750    46.7  112   477   161   350.5   63.6  13.05  3325.16  14.09   15.0
  13 Renegade 1.0.0 NN        :  3248.48    750    39.7   74   448   228   298.0   59.7  14.03  3328.56  14.03   15.0
  14 Clarity 4.1.0 NN         :  3234.02    750    37.8   51   465   234   283.5   62.0  13.69  3329.53  14.05   15.0
  15 Avalanche 2.1.0 NN       :  3163.93    750    28.9   42   350   358   217.0   46.7  15.17  3334.20  13.95   15.0
  16 Counter 5.5 NN           :  3158.87    750    28.3   35   355   360   212.5   47.3  15.28  3334.54  13.95   15.0

White advantage = 51.31 +/- 2.75
Draw rate (equal opponents) = 75.57 % +/- 0.74

[Frank Quisinsky]

Here the results from the stopped Drofa 4.1.0 NN test.
Engine need around 27% for 3150 Elo ... 85 games only but quit clear that 27% isn't possible.

Best
Frank

Code: Select all

Drofa 4.1.0 NN - Wasp 6.50 NN           1.0	-	5.0		16.67%		
Drofa 4.1.0 NN - Wasp 6.63 NN dev       1.5	-	4.5		25.00%		
Drofa 4.1.0 NN - Akimbo 0.8.0 NN        1.5	-	4.5		25.00%		
Drofa 4.1.0 NN - Starzix 3.0 NN         0.5	-	5.5		8.33%		
Drofa 4.1.0 NN - Clarity 4.1.0 NN       4.0	-	2.0		66.67%		
Drofa 4.1.0 NN - Caissa 1.16 NN         0.0	-	5.0		0.00%		
Drofa 4.1.0 NN - Counter 5.5 NN         2.5	-	2.5		50.00%		
Drofa 4.1.0 NN - Pawn 3.0 NN            0.0	-	5.0		0.00%		
Drofa 4.1.0 NN - Texel 1.11 NN          1.0	-	4.0		20.00%		
Drofa 4.1.0 NN - RubiChess 20240112 NN  0.5	-	4.5		10.00%		
Drofa 4.1.0 NN - Avalanche 2.1.0 NN     1.0	-	4.0		20.00%		
Drofa 4.1.0 NN - Lizard 10.1 NN         1.0	-	4.0		20.00%		
Drofa 4.1.0 NN - Minic 3.40 NN          0.5	-	4.5		10.00%		
Drofa 4.1.0 NN - Renegade 1.0.0 NN      0.5	-	4.5		10.00%		
Drofa 4.1.0 NN - Arasan 24.1 NN         1.0	-	4.0		20.00%		
Drofa 4.1.0 NN - Seer 2.8.0 NN          0.0	-	5.0		0.00%		

16.5	-	68.5		19.41%

Now Velvet 6.0.0 NN ...
Wasp 6.50 NN in danger to lose the first fast game!

To 95% sure the new record in average of short wins we will see with around 73-74 moves.

[Frank Quisinsky]

Here an example why I have canceled the Peacekeeper 2.20 NN test ...
Same for my 40 in 20 tournament I stopped, some of examples can be found in the databases, available in the download from my replay-zone to my 66+6 tournament.

https://www.amateurschach.de/fling/index.html

In the database "replaced engines" can be found the Peacekeeper 2.10 NN games.
Version 2.20 have the same problems!

[pgn] [Event "4 Minutes/Game + 2 Seconds/Move"] [Site "test-16, WASP-1"] [Date "2024.01.17"] [Round "1.7"] [White "Peacekeeper 2.20 NN"] [Black "Counter 5.5 NN"] [Result "1-0"] 1. e4 {book 0s} e5 {book 0s} 2. Nf3 {book 0s} Nc6 {book 0s} 3. Bb5 {book 0s} a6 {book 0s} 4. Ba4 {book 0s} Nf6 {book 0s} 5. Qe2 {book 0s} b5 {book 0s} 6. Bb3 {book 0s} Be7 {book 0s} 7. O-O {+0.13/27 17s} O-O {+0.09/18 6s} 8. d3 {+0.10/26 9s} a5 {+0.14/18 9s} 9. a4 {+0.12/27 16s} b4 {+0.13/20 7s} 10. Be3 {+0.05/25 6s} d5 {+0.14/20 6s} 11. Nbd2 {+0.11/25 6s} d4 {-0.17/22 5s} 12. Bg5 {+0.19/28 7s} Nd7 {0.00/23 6s} 13. Bxe7 {+0.27/27 7s} Nxe7 {0.00/25 14s} 14. Nh4 {+0.06/26 24s} Kh8 {-0.07/25 13s} 15. Bc4 {+0.17/25 8s} Nb6 {-0.25/21 8s} 16. Bb3 {+0.03/26 8s} f6 {-0.55/21 11s} 17. g3 {+0.08/27 7s} Bh3 {-0.67/20 8s} 18. Ng2 {+0.07/26 7s} Qe8 {-0.62/20 4s} 19. f4 {+0.15/26 5s} exf4 {-0.34/20 5s} 20. gxf4 {0.00/26 8s} Ned5 {-0.83/19 5s} 21. Rf3 {+0.04/26 9s} Bxg2 {-0.70/21 6s} 22. Qxg2 {+0.21/24 6s} Ne3 {-0.81/20 7s} 23. Qf2 {+0.21/25 12s} c5 {-0.78/19 4s} 24. f5 {+0.30/25 6s} Ra7 {-0.50/19 7s} 25. Rh3 {+0.53/23 5s} h6 {-0.44/18 5s} 26. Qf4 {+0.48/24 6s} Qd8 {-0.12/20 6s} 27. Kh1 {+1.83/20 5s} Qb8 {0.00/20 5s} 28. Qh4 {+1.93/21 5s} c4 {+0.54/17 6s} 29. dxc4 {+2.03/23 6s} Qc7 {+0.90/17 4s} 30. Rg1 {+3.10/19 4s} Qc5 {+1.75/16 8s} 31. Rg6 {+3.22/21 4s} Rc7 {+2.07/16 3s} 32. Qf4 {+4.79/17 5s} Rff7 {+3.27/16 9s} 33. Nf3 {+5.49/17 5s} Nexc4 {+4.39/17 10s} 34. Rgxh6+ {+5.44/19 4s} gxh6 {+5.46/22 10s} 35. Qxh6+ {+5.87/21 3s} Rh7 {+5.25/18 10s} 36. Qxf6+ {+5.69/24 6s} Kg8 {+5.41/21 3s} 37. Qe6+ {+5.89/23 4s} Kf8 {+5.72/22 3s} 38. Qf6+ {+6.07/23 6s} Kg8 {+5.51/22 3s} 39. Qe6+ {+5.91/21 3s} Kf8 {+5.81/20 4s} 40. Rg3 {+6.40/19 4s} Rcg7 {+6.12/22 7s} 41. Ng5 {+6.68/21 4s} Rxg5 {+7.54/23 9s} 42. Rxg5 {+6.74/23 4s} Rf7 {+7.86/22 2s} 43. Qh6+ {+6.82/19 5s} Ke8 {+6.11/19 3s} 44. Rg8+ {+6.84/19 3s} Rf8 {+5.44/19 2s} 45. Rxf8+ {+6.69/21 3s} Qxf8 {+6.10/21 3s} 46. Qc6+ {+6.90/22 3s} Ke7 {+6.05/21 4s} 47. Bxc4 {+7.09/20 5s} Nxc4 {+5.82/20 5s} 48. Qe6+ {+7.10/18 3s} Kd8 {book 0s} 49. Qxc4 {+7.34/19 3s} Qf6 {+6.32/18 4s} 50. Qc5 {+7.50/19 3s} Ke8 {+6.81/18 4s} 51. Qxa5 {+6.77/17 5s} Qg5 {+6.96/17 2s} 52. Qe5+ {+7.19/17 5s} Kf7 {+7.11/18 9s} 53. Qxd4 {+7.99/17 3s} Qc1+ {+7.92/19 8s} 54. Kg2 {+7.87/16 5s} Qxc2+ {+8.24/18 8s} 55. Kg3 {+8.51/15 2s} Qxa4 {+8.51/18 3s} 56. Qc4+ {+9.50/16 3s} Kg7 {+9.36/19 5s} 57. e5 {+12.71/19 2s} Qa7 {+11.38/20 5s} 58. f6+ {+14.37/18 3s} Kg6 {+14.76/17 7s} 59. Qe4+ {+14.26/16 3s} Kh6 {+11.82/18 3s} 60. Kf3 {+14.03/17 10s} Qf7 {+12.74/18 3s} 61. Qh4+ {+27.80/22 2s} Kg6 {+16.56/20 5s} 62. Qxb4 {+29.16/19 1s} Kh6 {+29.31/24 6s} 63. Qh4+ {+29.78/22 8s} Kg6 {+33.79/24 6s} 64. Qg4+ {+30.89/16 5s} Kh6 {+M25/28 5s} 65. Qg7+ {+33.93/22 1s} Qxg7 {+M16/31 2s} 66. fxg7 {+33.93/26 4s} Kh7 {+M13/31 1s} 67. Ke4 {+35.20/21 1s} Kxg7 {+M12/29 0s} 68. e6 {+36.19/21 3s} Kf6 {+M12/29 1s} 69. Kd5 {+36.19/18 2s} Kg7 {+M11/27 0s} 70. e7 {+35.96/16 2s} Kf7 {+M10/25 0s} 71. Kd6 {+M1/17 2s} Ke8 {+M9/23 0s} 72. b4 {+M1/16 2s} Kf7 {book 0s} 73. Kd7 {+M1/15 2s} Kg6 {+M8/21 0s} 74. e8=Q+ {+M1/13 1s} Kf5 {+M6/19 0s} 75. Qe3 {+M2/13 2s} Kf6 {+M5/15 0s} 76. Qe7+ {+M1/12 1s} Kf5 {+M6/17 0s} 77. Qe3 {+M1/13 2s} Kf6 {+M5/15 0s} 78. Kd6 {+M4/12 2s} Kf5 {+M5/19 0s} 79. Qf3+ {+M1/11 1s} Kg5 {+M4/13 0s} 80. Qf7 {+M1/12 2s} Kg4 {+M5/15 0s} 81. Qb3 {+M1/12 2s} Kf4 {+M6/18 0s} 82. Qg3+ {+M1/12 1s} Ke4 {+M6/17 0s} 83. Qc3 {+M1/13 1s} Kf4 {+M6/18 0s} 84. Qe5+ {+M1/12 2s} Kf3 {+M7/22 0s} 85. Kd5 {+33.91/13 2s} Kg2 {+M7/19 0s} 86. Qb2+ {+M1/13 2s} Kf3 {+M5/20 0s} 87. Qb3+ {+M1/13 2s} Kf4 {+M6/18 0s} 88. Qg3+ {+M1/12 1s} Kf5 {book 0s} 89. Qf2+ {+M1/12 2s} Kg4 {+M4/14 0s} 90. Ke5 {+M1/12 2s} Kg5 {+M3/13 0s} 91. Qf6+ {+M1/11 2s} Kg4 {+M3/13 0s} 92. Qf4+ {+M1/11 2s} Kh5 {+M3/12 0s} 93. Qf7+ {+M1/11 2s} Kh6 {+M4/15 0s} 94. Qf4+ {+M1/11 1s} Kg6 {+M4/15 0s} 95. Qg3+ {+M1/11 2s} Kf7 {+M4/14 0s} 96. Qg4 {+M1/14 2s} Ke8 {+M3/15 0s} 97. Kd6 {+M1/13 2s} Kf7 {+M4/13 0s} 98. Qe6+ {+M1/12 1s} Kg7 {+M5/15 0s} 99. Qd7+ {+M1/13 2s} Kf6 {+M5/17 0s} 100. Qh3 {+M1/13 2s} Kg5 {+M5/15 0s} 101. Qg3+ {+M1/13 2s} Kf6 {+M5/15 0s} 102. Qf4+ {+M1/12 2s} Kg7 {+M4/13 0s} 103. Qh4 {+M3/12 2s} Kg6 {+M4/15 0s} 104. Qd4 {+M1/12 2s} Kf5 {+M6/17 0s} 105. Qe5+ {+M1/12 2s} Kg4 {+M6/17 0s} 106. Qe4+ {+M1/12 1s} Kh3 {+M6/18 0s} 107. Qh1 {+M1/13 2s} Kg4 {+M6/17 0s} 108. Ke5 {+M1/12 2s} Kh3 {+M3/13 0s} 109. Qf1+ {+M1/12 2s} Kxh2 {+M4/19 0s} 110. Qf3 {+M0/13 2s} Kg1 {book 0s} 111. Ke4 {+M0/12 2s} Kh2 {book 0s} 112. Qe2+ {+M1/11 2s} Kg3 {+M3/15 0s} 113. Qe1+ {+M1/12 2s} Kg2 {+M4/13 0s} 114. Qe2+ {+M1/12 2s} Kg3 {+M3/11 0s} 115. Qe1+ {+M3/11 2s} Kg2 {+M4/15 0s} 116. Qd2+ {+M1/12 1s} Kg3 {+M5/15 0s} 117. Qc3+ {+M2/11 2s} Kf2 {+M6/17 0s} 118. Qd4+ {+M1/11 1s} Ke2 {+M5/15 0s} 119. Kf4 {+M1/12 2s} Kf1 {+M3/11 0s} 120. Qd3+ {+M1/11 2s} Ke1 {+M4/15 0s} 121. Qc3+ {+M1/12 2s} Ke2 {+M4/15 0s} 122. Qc2+ {+M1/12 2s} Kf1 {+M2/12 0s} 123. Ke3 {+M1/12 2s} Kg1 {+M3/11 0s} 124. Qg6+ {+M1/11 2s} Kf1 {+M2/11 0s} 125. Qg4 {+M1/12 2s} Ke1 {book 0s} 126. Qh5 {+M0/13 2s} Kf1 {book 0s} 127. Qe2+ {+M0/14 1s} Kg1 {book 0s} 128. Kf3 {+M0/15 2s} Kh1 {book 0s} 129. Qe1+ {+M0/14 1s} Kh2 {book 0s} 130. Qe3 {+M0/14 2s} Kh1 {+M3/11 0s} 131. Qe4 {+M0/15 2s} Kg1 {+M2/9 0s} 132. Qb1+ {+M0/14 2s} Kh2 {book 0s} 133. Qf5 {+M0/15 2s} Kh1 {+M2/10 0s} 134. Qc2 {+M0/16 1s} Kg1 {book 0s} 135. Qf2+ {+M0/13 1s} Kh1 {book 0s} 136. Qe1+ {+M0/12 1s} Kh2 {book 0s} 137. Qd2+ {+M0/12 1s} Kh3 {+M2/9 0s} 138. Qc3 {+M0/14 2s} Kh4 {+M3/12 0s} 139. Qc5 {+M0/14 1s} Kh3 {book 0s} 140. Qd4 {+M0/13 2s} Kh2 {book 0s} 141. Qe5+ {+M0/12 2s} Kh1 {+M2/10 0s} 142. Qe3 {+M0/14 2s} Kh2 {book 0s} 143. Qf4+ {+M0/12 1s} Kh1 {+M2/11 0s} 144. Qe5 {+M0/15 2s} Kg1 {book 0s} 145. Qa1+ {+M0/14 1s} Kh2 {book 0s} 146. Qe1 {+M0/14 2s} Kh3 {book 0s} 147. Qe7 {+M0/13 2s} Kh2 {book 0s} 148. Qe5+ {+M0/10 2s} Kh1 {+M2/9 0s} 149. Qe3 {+M0/14 2s} Kh2 {book 0s} 150. Qd4 {+M0/13 2s} Kh1 {+M2/10 0s} 151. Qd3 {+M0/13 1s} Kg1 {+M2/10 0s} 152. Qg6+ {+M0/13 1s} Kh2 {+M1/7 0s} 153. Qb1 {+M0/13 2s} Kh3 {book 0s} 154. Qh7# {+M1/13 2s} 1-0[/pgn]

[Frank Quisinsky]

With Mida 2.3 NN and Kuma 1.2 NN I made also a small test (like Drofa 4.1.0 NN).
But it seems both are also not stronger as 3150 Elo.
For that tournament at the moment not strong enough.
But I am watching if newer versions are available.

Enough for today ...
Great king attacks in the first Velvet 6.0.0 NN games.
This engines is incredible!

[Frank Quisinsky]

I wrote ...
Wasp 6.50 NN in danger.

Never I should thinking such things.
Now Wasp 6.50 NN have to give revanche!

I wrote ...
To 95% sure the new record in average of short wins we will see with around 73-74 moves.

Move average wins for Velvet 6.0.0 NN after around 200 games = 73 moves / 34 won games.

[pgn] [Event "4 Minutes/Game + 2 Seconds/Move"] [Site "test-18, WASP-1"] [Date "2024.01.17"] [Round "11.1"] [White "Velvet 6.0.0 NN"] [Black "Wasp 6.50 NN"] [Result "1-0"] 1. c4 {book 0s} e5 {book 0s} 2. Nf3 {book 0s} d6 {book 0s} 3. Nc3 {book 0s} f5 {book 0s} 4. d4 {book 0s} e4 {book 0s} 5. Ng5 {book 0s} Nf6 {book 0s} 6. f3 {book 0s} exf3 {book 0s} 7. Nxf3 {+0.33/18 7s} Nc6 {+0.20/19 8s} 8. Qc2 {+0.33/19 8s (g3)} g6 {+0.17/20 11s (Be7)} 9. Bf4 {+0.34/19 11s (Bg5)} Bg7 {0.00/20 8s (Bd7)} 10. O-O-O {+0.29/19 9s (e3)} h6 {-0.26/19 8s} 11. h4 {+0.24/20 9s} O-O {-0.28/19 8s} 12. e3 {+0.12/19 8s} Ng4 {-0.07/18 15s (Ne7)} 13. c5 {+0.32/20 31s (a3)} Bd7 {0.00/19 8s (Kh8)} 14. Bc4+ {+0.45/18 5s (Nh2)} Kh8 {-0.04/19 7s} 15. a3 {+0.34/18 8s (Kb1)} Rc8 {-0.19/19 7s} 16. Rhe1 {+0.49/18 6s} Qe7 {-0.04/18 7s} 17. Qd2 {+0.34/19 26s (Ba2)} Rfe8 {+0.09/19 9s} 18. Nd5 {+1.09/20 11s (Ba2)} Qd8 {0.00/20 7s (Qe4)} 19. Bg5 {+1.32/21 17s (Nc3)} hxg5 {0.00/19 8s (Ne7)} 20. hxg5 {+1.41/22 16s} Be6 {+0.47/19 11s} 21. e4 {+1.00/21 15s} f4 {+0.46/19 17s (dxc5)} 22. Qxf4 {+3.07/19 10s} Na5 {+0.33/18 6s} 23. Rh1+ {+3.08/19 14s (e5)} Kg8 {+1.70/20 7s} 24. Qxg4 {+2.88/19 14s (Ba2)} Nxc4 {+1.73/17 5s} 25. Qh4 {+3.70/19 12s} Re7 {+1.78/18 12s (Kf8)} 26. Nf6+ {+3.80/15 1s (Nf4)} Kf8 {+1.69/18 8s} 27. d5 {+3.62/19 8s (b3)} Bf7 {+1.95/17 5s} 28. Nd4 {+4.03/19 3s (Qf4)} Ne5 {+1.79/17 10s (Nxb2)} 29. Rhf1 {+5.83/19 8s (Kb1)} Ra8 {+3.14/16 10s (Rb8)} 30. Qh7 {+8.45/17 1s} Qc8 {+5.23/18 8s} 31. Ng8 {+9.79/19 6s (Kb1)} dxc5 {+5.85/19 3s} 32. Nxe7 {+12.87/22 9s} Kxe7 {+8.29/21 9s} 33. Qxg7 {+13.70/22 5s} Qh8 {+8.69/20 3s} 34. Rxf7+ {+14.59/20 1s} Nxf7 {+9.23/19 5s} 35. Qxg6 {+14.67/21 2s} cxd4 {+9.75/19 5s (Re8)} 36. Qe6+ {+16.47/20 4s} Kd8 {+10.01/18 3s} 37. Qxf7 {+16.96/20 2s} a5 {+10.20/18 3s (Qe8)} 38. g6 {+41.01/23 5s} Kc8 {+11.14/20 6s (Qe8)} 39. g7 {+47.07/21 1s} Qh6+ {+12.01/20 3s} 40. Kb1 {+57.09/23 6s} Kb8 {+12.30/20 3s} 41. g8=Q+ {+64.05/23 6s} Ka7 {+12.56/19 3s} 42. Qgg7 {+63.40/23 2s} Qxg7 {+12.90/17 3s} 43. Qxg7 {+65.20/19 1s} Kb6 {+13.43/16 2s (Rc8)} 44. Rc1 {+M9/22 1s (e5)} c6 {+M9/17 1s} 45. dxc6 {+M8/25 1s} Rb8 {+M8/15 0s} 46. cxb7 {+M6/25 1s} Ka7 {+M7/13 0s} 47. Rc8 {+M5/40 2s (Rc6)} Rxb7 {+M5/10 0s} 48. Qxd4+ {+M4/83 2s} Rb6 {+M4/7 0s} 49. Rc6 {+M3/127 0s} Ka8 {+M3/5 0s (Kb7)} 50. Qxb6 {+M2/127 0s} a4 {+M2/3 0s} 51. Rc8# {+M1/127 0s} 1-0 [/pgn]

Shit, the nimbus is gone.
After a very long time ... I saw a lost game of Wasp 6.50 NN less than 55 moves to mate.
A bad day in computer chess history ... no, no!

Velvet is a king attack monster and need a new logo ... monster logo.
OK, let us wait of the revanche.

[Frank Quisinsky]

Code: Select all

test-19

Obsidian 10.0 NN - Wasp 6.50 NN            2.5	-	0.5		83.33%		
Obsidian 10.0 NN - Wasp 6.63 NN dev        1.5	-	1.5		50.00%		
Obsidian 10.0 NN - Akimbo 0.8.0 NN         2.5	-	0.5		83.33%		
Obsidian 10.0 NN - Starzix 3.0 NN          1.5	-	1.5		50.00%		
Obsidian 10.0 NN - Clarity 4.1.0 NN        2.5	-	0.5		83.33%		
Obsidian 10.0 NN - Caissa 1.16 NN          1.5	-	1.5		50.00%		
Obsidian 10.0 NN - Counter 5.5 NN          2.5	-	0.5		83.33%		
Obsidian 10.0 NN - Pawn 3.0 NN             2.5	-	0.5		83.33%		
Obsidian 10.0 NN - Texel 1.11 NN           2.0	-	1.0		66.67%		
Obsidian 10.0 NN - RubiChess 20240112 NN   1.5	-	1.5		50.00%		
Obsidian 10.0 NN - Avalanche 2.1.0 NN      2.5	-	0.5		83.33%		
Obsidian 10.0 NN - Lizard 10.1 NN          2.0	-	1.0		66.67%		
Obsidian 10.0 NN - Minic 3.40 NN           1.5	-	0.5		75.00%		
Obsidian 10.0 NN - Renegade 1.0.0 NN       1.0	-	1.0		50.00%		
Obsidian 10.0 NN - Arasan 24.1 NN          1.0	-	1.0		50.00%		
Obsidian 10.0 NN - Seer 2.8.0 NN           1.5	-	0.5		75.00%		
Obsidian 10.0 NN - Velvet 6.0.0 NN         1.0	-	1.0		50.00%
	
31.0	-	15.0		67.39%

Code: Select all

test-18

Velvet 6.0.0 NN - Wasp 6.50 NN             6.5	-	6.5		50.00%		
Velvet 6.0.0 NN - Wasp 6.63 NN dev         6.0	-	7.0		46.15%		
Velvet 6.0.0 NN - Akimbo 0.8.0 NN          7.5	-	5.5		57.69%		
Velvet 6.0.0 NN - Starzix 3.0 NN           5.5	-	7.5		42.31%		
Velvet 6.0.0 NN - Clarity 4.1.0 NN         8.0	-	5.0		61.54%		
Velvet 6.0.0 NN - Caissa 1.16 NN           4.0	-	9.0		30.77%		
Velvet 6.0.0 NN - Counter 5.5 NN           8.0	-	5.0		61.54%		
Velvet 6.0.0 NN - Pawn 3.0 NN              6.0	-	7.0		46.15%		
Velvet 6.0.0 NN - Texel 1.11 NN            6.0	-	7.0		46.15%		
Velvet 6.0.0 NN - RubiChess 20240112 NN    3.5	-	8.5		29.17%		
Velvet 6.0.0 NN - Avalanche 2.1.0 NN       9.5	-	2.5		79.17%		
Velvet 6.0.0 NN - Lizard 10.1 NN           5.5	-	6.5		45.83%		
Velvet 6.0.0 NN - Minic 3.40 NN            7.0	-	5.0		58.33%		
Velvet 6.0.0 NN - Renegade 1.0.0 NN        8.0	-	4.0		66.67%		
Velvet 6.0.0 NN - Arasan 24.1 NN           4.5	-	7.5		37.50%		
Velvet 6.0.0 NN - Seer 2.8.0 NN            4.0	-	8.0		33.33%

99.5	-	101.5		49.50%

Both are running on 5 cores ... this need a while!

[Frank Quisinsky]

20. Velvet 6.0.0 NN
21. Obsidian 10.0 NN ... still running
22. Not sure what I should test as next ... maybe a current dev version from Stockfish?

Code: Select all

              Name                 Games     Win     Draw     Lose       Pts         S-B         %    wins-55m  lost-55m  AV-wins  AV-draws  AV-all

01. RubiChess 20240112 NN       :   800  :  372+  :  423=  :    5-  :   583.5  : 220113.50  :  72.94%     42        0        80       78       79
02. Caissa 1.16 NN              :   800  :  352+  :  444=  :    4-  :   574.0  : 215833.50  :  71.75%     20        0        86       78       82
03. Seer 2.8.0 NN               :   800  :  327+  :  454=  :   19-  :   554.0  : 207659.25  :  69.25%     19        0        86       80       83
04. Arasan 24.1 NN              :   800  :  211+  :  518=  :   71-  :   470.0  : 176787.75  :  58.75%     24        3        85       82       83
05. Minic 3.40 NN               :   800  :  144+  :  537=  :  119-  :   412.5  : 155358.00  :  51.56%      5       16        96       92       90
06. Pawn 3.0 NN                 :   800  :  145+  :  530=  :  125-  :   410.0  : 154104.50  :  51.25%      5        8        83       87       87
07. Akimbo 0.8.0 NN             :   800  :  159+  :  502=  :  139-  :   410.0  : 152435.25  :  51.25%      5       15        95       91       90
08. Velvet 6.0.0 NN             :   800  :  156+  :  507=  :  137-  :   409.5  : 153755.50  :  51.19%     17        0        75       84       85
09. Starzix 3.0 NN              :   800  :  135+  :  520=  :  145-  :   395.0  : 146430.75  :  49.38%      3       17        94       90       88
10. Texel 1.11 NN               :   800  :  125+  :  505=  :  170-  :   377.5  : 140435.00  :  47.19%     19        1        77       90       88
---------------------------------------------------------------------------------------------------------------------------------------------------
11. Wasp 6.50 NN                :   800  :  123+  :  508=  :  169-  :   377.0  : 139520.75  :  47.13%     27        1        78       81       83
12. Wasp 6.63 NN dev            :   800  :  121+  :  510=  :  169-  :   376.0  : 140851.00  :  47.00%     15        4        81       80       83
13. Lizard 10.1 NN              :   800  :  119+  :  511=  :  170-  :   374.5  : 139922.25  :  46.81%      7       16        87       96       92
14. Renegade 1.0.0 NN           :   800  :   78+  :  477=  :  245-  :   316.5  : 118456.25  :  39.56%      2       27        91       99       93
15. Clarity 4.1.0 NN            :   800  :   55+  :  490=  :  255-  :   300.0  : 113636.25  :  37.50%      0       19        97       83       84
16. Avalanche 2.1.0 NN          :   800  :   46+  :  373=  :  381-  :   232.5  :  87590.00  :  29.06%      0       60        92       86       83
17. Counter 5.5 NN              :   800  :   36+  :  383=  :  381-  :   227.5  :  87028.50  :  28.44%      3       26        84      101       94


White Wins =  1.796 ( 26.41% )
Draws      =  4.096 ( 60.24% )
Black Wins =    908 ( 13.35% )
Average    = 171.98 ( 85,99 moves )

Code: Select all

   # Player                   :      Elo  Games  Score%  won  draw  lost  Points  Draw%  Error   OppAvg   OppE   OppD
   1 RubiChess 20240112 NN    :  3496.39    800    72.9  372   423     5   583.5   52.9  15.70  3310.20  13.56   16.0
   2 Caissa 1.16 NN           :  3486.13    800    71.8  352   444     4   574.0   55.5  15.32  3310.84  13.59   16.0
   3 Seer 2.8.0 NN            :  3465.06    800    69.3  327   454    19   554.0   56.8  15.13  3312.16  13.60   16.0
   4 Arasan 24.1 NN           :  3384.06    800    58.8  211   518    71   470.0   64.8  13.13  3317.22  13.72   16.0
   5 Minic 3.40 NN            :  3331.88    800    51.6  144   537   119   412.5   67.1  13.27  3320.48  13.72   16.0
   6 Akimbo 0.8.0 NN          :  3329.63    800    51.3  159   502   139   410.0   62.8  12.48  3320.62  13.77   16.0
   7 Pawn 3.0 NN              :  3329.63    800    51.3  145   530   125   410.0   66.3  12.78  3320.62  13.75   16.0
   8 Velvet 6.0.0 NN          :  3329.19    800    51.2  156   507   137   409.5   63.4  13.16  3320.65  13.72   16.0
   9 Starzix 3.0 NN           :  3316.28    800    49.4  135   520   145   395.0   65.0  12.85  3321.46  13.74   16.0
  10 Texel 1.11 NN            :  3300.45    800    47.2  125   505   170   377.5   63.1  12.80  3322.45  13.75   16.0
  11 Wasp 6.50 NN             :  3300.00    800    47.1  123   508   169   377.0   63.5  12.56  3322.48  13.76   16.0
  12 Wasp 6.63 NN dev         :  3299.10    800    47.0  121   510   169   376.0   63.8  12.91  3322.53  13.74   16.0
  13 Lizard 10.1 NN           :  3297.75    800    46.8  119   511   170   374.5   63.9  13.06  3322.62  13.73   16.0
  14 Renegade 1.0.0 NN        :  3244.85    800    39.6   78   477   245   316.5   59.6  13.63  3325.92  13.69   16.0
  15 Clarity 4.1.0 NN         :  3229.39    800    37.5   55   490   255   300.0   61.3  13.87  3326.89  13.68   16.0
  16 Avalanche 2.1.0 NN       :  3162.54    800    29.1   46   373   381   232.5   46.6  15.42  3331.07  13.58   16.0
  17 Counter 5.5 NN           :  3157.27    800    28.4   36   383   381   227.5   47.9  14.63  3331.40  13.63   16.0

White advantage = 52.36 +/- 2.62
Draw rate (equal opponents) = 75.08 % +/- 0.73

I await more short wins from Velvet but the "move average wins" is just great with 75!!
Tomorrow I have the Obsidian 10.0 NN results.
System working with full power at first on the Velvet 6.0.0 results.

Here the current results from Obsidian 10.0 NN ...

Code: Select all

Obsidian 10.0 NN - Wasp 6.50 NN                 11.0	-	4.0		73.33%		
Obsidian 10.0 NN - Wasp 6.63 NN dev              9.0	-	6.0		60.00%		
Obsidian 10.0 NN - Akimbo 0.8.0 NN              12.0	-	3.0		80.00%		
Obsidian 10.0 NN - Starzix 3.0 NN               10.0	-	5.0		66.67%		
Obsidian 10.0 NN - Clarity 4.1.0 NN             11.5	-	3.5		76.67%		
Obsidian 10.0 NN - Caissa 1.16 NN                8.0	-	7.0		53.33%		
Obsidian 10.0 NN - Counter 5.5 NN               12.0	-	3.0		80.00%		
Obsidian 10.0 NN - Pawn 3.0 NN                   9.5	-	4.5		67.86%		
Obsidian 10.0 NN - Texel 1.11 NN                10.0	-	4.0		71.43%		
Obsidian 10.0 NN - RubiChess 20240112 NN         7.0	-	8.0		46.67%		
Obsidian 10.0 NN - Avalanche 2.1.0 NN           11.0	-	3.0		78.57%		
Obsidian 10.0 NN - Lizard 10.1 NN               10.5	-	3.5		75.00%		
Obsidian 10.0 NN - Minic 3.40 NN                11.5	-	2.5		82.14%		
Obsidian 10.0 NN - Renegade 1.0.0 NN             8.0	-	6.0		57.14%		
Obsidian 10.0 NN - Arasan 24.1 NN                8.5	-	5.5		60.71%		
Obsidian 10.0 NN - Seer 2.8.0 NN                 9.5	-	4.5		67.86%		
Obsidian 10.0 NN - Velvet 6.0.0 NN               9.5	-	4.5		67.86%

168.5	-	77.5		68.50%
246 out of 850 games played

[Frank Quisinsky]

Conditions changed:

Code: Select all

Conditions:

GUI:              Shredder 12 by Stefan Meyer-Kahlen

System Wasp-1:    Intel® Core™ i9-10900k = 10 cores, 4.4Ghz overclocked
Time control:     4 Minutes/Game + 2 Seconds/Move
Operating system: Windows 10

System Wasp-2:    Intel® Core™ i7-1185g7 =  4 cores, 3.0Ghz overclocked
Time control:     6 Minutes/Game + 3 Seconds/Move
Operating system: Windows 11
                  Dell Ultrabook is rather rarely used

Elo:              Wasp 6.50 NN is playing with ~3300 Elo (reference)
Tournament type:  Everyone against everyone, 50 games per match
                  Max. 25 engines = 1.200 games per engine
                  If more as 25 engine = A new engine replaces the last place.
                  An engine should reach approx. 3150 Elo

Endgame 5-man:    syzygy
Opening book:     feobos-6m-v2.1.bkt for Shredder GUI, based on 3.698 lines (balanced)
Cores:            1 core, Hyperthreading = off, Large-Pages = off
Hashtables:       128Mb
Contempt:         Default engine settings changed to contempt = 0 (if possible)
Other settings:   resign = off, ponder = off, learning = off

Replay games:     Every draw less than 20 moves



Frank Quisinsky, Germany / Gutweiler

If I have 25 engines in a group I can produced 1200 games for a test-run.
A test-run with 1200 games and 10 cores need around 2 days with game in 4 + 2 seconds.
= Max. 15 engines per months can be test (should be enough games).

If more as 25 engines, a new engine replaces the last place.

Example:
I have 30 different engines.
First 24 + the new engine for the test-run and all results can go in rating list.

Again, for me is important:
Which new engines I can add later in a new tournament with longer time-controls (Iooking on move-average and engine should not lose to many games fast). With the final results that I can make the current group for my 66+6 tournaments better and better.

Best
Frank

[Frank Quisinsky]

Obsidian 10.0 NN ... produced problems!
Engine support syzygy first time, but produced some bugs and lost points!

- does not give mate with KR vs. K ... games ended with draw, 50 moves-rule
- does not give mate with KNB vs. K ... games ended with draw, 50 moves-rule
- does not give mate with KQ vs. K ... games ended with draw, 50-moves-rule
- does not give mate with KQ vs. KR ... games ended with draw, 50-moves-rule

- engine avoid draw in clear draw position ... games ended with 50-moves rule (very high move-average)

But I let the engine play all of the games:
Current results:

Code: Select all

Obsidian 10.0 NN - Wasp 6.50 NN              29.5	-	10.5		73.75%		
Obsidian 10.0 NN - Wasp 6.63 NN dev          28.0	-	12.0		70.00%		
Obsidian 10.0 NN - Akimbo 0.8.0 NN           29.0	-	11.0		72.50%		
Obsidian 10.0 NN - Starzix 3.0 NN            27.0	-	13.0		67.50%		
Obsidian 10.0 NN - Clarity 4.1.0 NN          30.0	-	10.0		75.00%		
Obsidian 10.0 NN - Caissa 1.16 NN            19.0	-	20.0		48.72%		
Obsidian 10.0 NN - Counter 5.5 NN            31.5	-	 7.5		80.77%		
Obsidian 10.0 NN - Pawn 3.0 NN               27.0	-	12.0		69.23%		
Obsidian 10.0 NN - Texel 1.11 NN             26.0	-	13.0		66.67%		
Obsidian 10.0 NN - RubiChess 20240112 NN     18.0	-	21.0		46.15%		
Obsidian 10.0 NN - Avalanche 2.1.0 NN        27.5	-	11.5		70.51%		
Obsidian 10.0 NN - Lizard 10.1 NN            29.0	-	10.0		74.36%		
Obsidian 10.0 NN - Minic 3.40 NN             28.0	-	11.0		71.79%		
Obsidian 10.0 NN - Renegade 1.0.0 NN         27.0	-	12.0		69.23%		
Obsidian 10.0 NN - Arasan 24.1 NN            22.5	-	15.5		59.21%		
Obsidian 10.0 NN - Seer 2.8.0 NN             22.0	-	17.0		56.41%		
Obsidian 10.0 NN - Velvet 6.0.0 NN           28.0	-	11.0		71.79%		

449.0	-	218.0		67.32%		
667 out of 850 games played

Games will be available in the late afternoon on ...
https://www.amateurschach.de/fling/index.html