New engine releases 2024 ... 6+3

Discussion of computer chess matches and engine tournaments.

Moderator: Ras

Frank Quisinsky
Posts: 6888
Joined: Wed Nov 18, 2009 7:16 pm
Location: Gutweiler, Germany
Full name: Frank Quisinsky

Re: New engine releases 2024 ... 6+3

Post by Frank Quisinsky »

21. Obsidian 10.0 NN ... lost around 30 points (without the bug rank 03)
22. Stromphrax 4.0.0 NN ... still running
23. RukChess 3.0.18 NN ... soon

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       :   850  :  375+  :  469=  :    6-  :   609.5  : 243855.25  :  71.71%     42        0       80       81       81
02. Caissa 1.16 NN              :   850  :  355+  :  489=  :    6-  :   599.5  : 239144.00  :  70.53%     21        0       86       79       82
03. Seer 2.8.0 NN               :   850  :  330+  :  493=  :   27-  :   576.5  : 228970.25  :  67.82%     19        0       86       83       84
04. Obsidian 10.0 NN            :   850  :  313+  :  526=  :   11-  :   576.0  : 232168.00  :  67.76%     14        1       90      101       97
05. Arasan 24.1 NN              :   850  :  211+  :  560=  :   79-  :   491.0  : 195924.00  :  57.76%     24        3       85       83       84
06. Minic 3.40 NN               :   850  :  144+  :  564=  :  142-  :   426.0  : 169459.75  :  50.12%      5       18       96       92       90
07. Pawn 3.0 NN                 :   850  :  145+  :  560=  :  145-  :   425.0  : 168972.50  :  50.00%      5        8       83       89       88
08. Velvet 6.0.0 NN             :   850  :  156+  :  538=  :  156-  :   425.0  : 168876.25  :  50.00%     17        0       75       86       87
09. Akimbo 0.8.0 NN             :   850  :  159+  :  531=  :  160-  :   424.5  : 166950.75  :  49.94%      5       15       95       91       90   
10. Starzix 3.0 NN              :   850  :  135+  :  552=  :  163-  :   411.0  : 161519.50  :  48.35%      3       18       94       90       83
---------------------------------------------------------------------------------------------------------------------------------------------------
11. Texel 1.11 NN               :   850  :  126+  :  536=  :  188-  :   394.0  : 155562.25  :  46.35%     19        1       77       90       88
12. Wasp 6.63 NN dev            :   850  :  121+  :  541=  :  188-  :   391.5  : 155411.75  :  46.06%     15        4       81       81       83
13. Wasp 6.50 NN                :   850  :  124+  :  531=  :  195-  :   389.5  : 152391.50  :  45.82%     27        1       78       81       84
14. Lizard 10.1 NN              :   850  :  119+  :  533=  :  198-  :   385.5  : 151984.75  :  45.35%      7       17       87       96       92
15. Renegade 1.0.0 NN           :   850  :   78+  :  507=  :  265-  :   331.5  : 131796.50  :  39.00%      2       30       91      100       94
16. Clarity 4.1.0 NN            :   850  :   55+  :  513=  :  282-  :   311.5  : 124875.25  :  36.65%      0       24       97       83       84
17. Avalanche 2.1.0 NN          :   850  :   46+  :  398=  :  406-  :   245.0  :  98250.25  :  28.82%      0       61       92       88       83
18. Counter 5.5 NN              :   850  :   36+  :  403=  :  411-  :   237.5  :  96256.00  :  27.94%      3       27       84      102       95


White Wins =  2.015 ( 26.34% )
Draws      =  4.622 ( 60.42% )
Black Wins =  1.013 ( 13.24% )
Average    = 174.33 ( 87,16 moves )

Code: Select all

  # Player                   :      Elo  Games  Score%  won  draw  lost  Points  Draw%  Error   OppAvg   OppE   OppD
   1 RubiChess 20240112 NN    :  3496.35    850    71.7  375   469     6   609.5   55.2  14.82  3320.85  13.11   17.0
   2 Caissa 1.16 NN           :  3486.33    850    70.5  355   489     6   599.5   57.5  14.04  3321.44  13.15   17.0
   3 Seer 2.8.0 NN            :  3463.87    850    67.8  330   493    27   576.5   58.0  14.54  3322.76  13.12   17.0
   4 Obsidian 10.0 NN         :  3463.52    850    67.8  313   526    11   576.0   61.9  13.78  3322.78  13.17   17.0
   5 Arasan 24.1 NN           :  3386.74    850    57.8  211   560    79   491.0   65.9  12.58  3327.29  13.24   17.0
   6 Minic 3.40 NN            :  3331.08    850    50.1  144   564   142   426.0   66.4  12.82  3330.57  13.22   17.0
   7 Velvet 6.0.0 NN          :  3330.23    850    50.0  156   538   156   425.0   63.3  12.22  3330.62  13.26   17.0
   7 Pawn 3.0 NN              :  3330.23    850    50.0  145   560   145   425.0   65.9  12.17  3330.62  13.26   17.0
   9 Akimbo 0.8.0 NN          :  3329.80    850    49.9  159   531   160   424.5   62.5  12.46  3330.64  13.24   17.0
  10 Starzix 3.0 NN           :  3318.43    850    48.4  135   552   163   411.0   64.9  12.69  3331.31  13.23   17.0
  11 Texel 1.11 NN            :  3303.84    850    46.4  126   536   188   394.0   63.1  12.68  3332.17  13.23   17.0
  12 Wasp 6.63 NN dev         :  3301.71    850    46.1  121   541   188   391.5   63.6  12.14  3332.30  13.26   17.0
  13 Wasp 6.50 NN             :  3300.00    850    45.8  124   531   195   389.5   62.5  12.40  3332.40  13.25   17.0
  14 Lizard 10.1 NN           :  3296.58    850    45.4  119   533   198   385.5   62.7  12.51  3332.60  13.24   17.0
  15 Renegade 1.0.0 NN        :  3249.71    850    39.0   78   507   265   331.5   59.6  13.14  3335.36  13.20   17.0
  16 Clarity 4.1.0 NN         :  3231.84    850    36.6   55   513   282   311.5   60.4  13.49  3336.41  13.18   17.0
  17 Avalanche 2.1.0 NN       :  3169.01    850    28.8   46   398   406   245.0   46.8  14.49  3340.10  13.13   17.0
  18 Counter 5.5 NN           :  3161.47    850    27.9   36   403   411   237.5   47.4  14.66  3340.55  13.12   17.0

White advantage = 52.75 +/- 2.43
Draw rate (equal opponents) = 76.12 % +/- 0.68
Games can be found in replay-zone from my 66+6 tournament:
https://www.amateurschach.de/fling/index.html ... Point 08!

Here the first Stromphrax 4.0.0 NN results:

Code: Select all

Stormphrax 4.0.0 NN - Wasp 6.50 NN            13.5	-	 5.5		71.05%		
Stormphrax 4.0.0 NN - Wasp 6.63 NN dev        10.0	-	 8.0		55.56%		
Stormphrax 4.0.0 NN - Akimbo 0.8.0 NN         10.0	-	 8.0		55.56%		
Stormphrax 4.0.0 NN - Starzix 3.0 NN          11.5	-	 6.5		63.89%		
Stormphrax 4.0.0 NN - Clarity 4.1.0 NN        14.0	-	 4.0		77.78%		
Stormphrax 4.0.0 NN - Caissa 1.16 NN           4.0	-	14.0		22.22%		
Stormphrax 4.0.0 NN - Counter 5.5 NN          14.0	-	 4.0		77.78%		
Stormphrax 4.0.0 NN - Pawn 3.0 NN             10.5	-	 7.5		58.33%		
Stormphrax 4.0.0 NN - Texel 1.11 NN           10.0	-	 8.0		55.56%		
Stormphrax 4.0.0 NN - RubiChess 20240112 NN    4.5	-	13.5		25.00%		
Stormphrax 4.0.0 NN - Avalanche 2.1.0 NN      13.5	-	 4.5		75.00%		
Stormphrax 4.0.0 NN - Lizard 10.1 NN          10.0	-	 8.0		55.56%		
Stormphrax 4.0.0 NN - Minic 3.40 NN            8.0	-	10.0		44.44%		
Stormphrax 4.0.0 NN - Renegade 1.0.0 NN       10.5	-	 7.5		58.33%		
Stormphrax 4.0.0 NN - Arasan 24.1 NN           9.0	-	 9.0		50.00%		
Stormphrax 4.0.0 NN - Seer 2.8.0 NN            5.0	-	13.0		27.78%		
Stormphrax 4.0.0 NN - Velvet 6.0.0 NN         10.0	-	 7.0		58.82%		
Stormphrax 4.0.0 NN - Obsidian 10.0 NN         6.5	-	10.5		38.24%

174.5	-	148.5		54.02%
323 out of 900 games played
Best
Frank
Frank Quisinsky
Posts: 6888
Joined: Wed Nov 18, 2009 7:16 pm
Location: Gutweiler, Germany
Full name: Frank Quisinsky

Re: New engine releases 2024 ... 6+3

Post by Frank Quisinsky »

Here the tournament log file:

Bugs so far in three engines:

1. Avalanche 2.1.0 NN
2. Peacekeeper 2.20 NN
3. Obsidian 10.0 NN

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

System Wasp-1 = Intel® Core™ i9-10900k = 10 cores, 4.4Ghz overclocked, game in 4 minutes + 2 seconds
System Wasp-2 = Intel® Core™ i7-1185g7 =  4 cores, 3.0Ghz overclocked, game in 6 minutes + 3 seconds, rather rarely used

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

Start point December 17th, 2023 - until ongoing / continuously

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

23. RukChess 3.0.18 NN                                   executable = December 17th, 2023    Test on Wasp-1 = soon
    Ilya Rukavishnikov (RUS)
    https://github.com/Ilya-Ruk/RukChess

22. Stormphrax 4.0.0 NN                                  executable = December 17th, 2023    Test on Wasp-1 = still running
    Conor Anstey (GBR)
    https://github.com/Ciekce/Stormphrax

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

21. Obsidian 10.0 NN                                     released   = January 17th, 2024     Test on Wasp=1 = v21
    Gabriele Lombardo (ITA)
    https://github.com/gab8192/Obsidian

    Engine support syzygy first time but produced some bugs and lost around 30 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
    - does not give mate in some other constellations!
    - engine avoid draw in clear draw position ... games ended with 50-moves rule (very high move-average)

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

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

15. 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

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

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

12. 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

11. 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 or engine produced bugs / problems.

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

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

19. Drofa 4.1.0 NN                                       executable = December 24th, 2023     no test for the moment
    Alexander Litov & Rhys Rustad-Elliott (RUS, CAN)
    https://github.com/justNo4b/Drofa

17. Peacekeeper 2.20 NN                                  executable = December 24th, 2023     no test for the moment
    Kyle Zhang (USA)
    https://github.com/Sazgr/peacekeeper

    Test not possible!
    - 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)

10. Kuma 1.2 NN                                          released   = January 12th, 2024      no test for the moment
    Kato Daichi, JPN
    https://github.com/kato-daichi/kuma
Frank Quisinsky
Posts: 6888
Joined: Wed Nov 18, 2009 7:16 pm
Location: Gutweiler, Germany
Full name: Frank Quisinsky

Re: New engine releases 2024 ... 6+3

Post by Frank Quisinsky »

22. Stormphrax 4.0.0 NN
23. Starzix 4.0 NN ... still running
24. RukChess 3.0.18 NN ... soon

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       :   900  :  402+  :  492=  :    6-  :   648.0  : 275427.00  :  72.00%     47        0       79       82       81
02. Caissa 1.16 NN              :   900  :  378+  :  515=  :    7-  :   635.5  : 269293.50  :  70.61%     22        0       86       79       82
03. Seer 2.8.0 NN               :   900  :  349+  :  524=  :   27-  :   611.0  : 257826.25  :  67.89%     22        0       86       84       85
04. Obsidian 10.0 NN            :   900  :  329+  :  560=  :   11-  :   609.0  : 260519.00  :  67.67%     17        1       90      102       97
05. Arasan 24.1 NN              :   900  :  223+  :  595=  :   82-  :   520.5  : 220587.00  :  57.83%     26        3       84       84       85
06. Stormphrax 4.0.0 NN         :   900  :  227+  :  530=  :  143-  :   492.0  : 203994.50  :  54.67%      3       18       90      110      100
07. Minic 3.40 NN               :   900  :  152+  :  593=  :  155-  :   448.5  : 189362.75  :  49.83%      5       18       97       93       91
08. Akimbo 0.8.0 NN             :   900  :  162+  :  570=  :  168-  :   447.0  : 186675.25  :  49.67%      6       15       94       93       91
09. Pawn 3.0 NN                 :   900  :  147+  :  597=  :  156-  :   445.5  : 187917.00  :  49.50%      5        9       83       90       89
10. Velvet 6.0.0 NN             :   900  :  161+  :  569=  :  170-  :   445.5  : 187808.25  :  49.50%     17        0       75       87       88
---------------------------------------------------------------------------------------------------------------------------------------------------
11. Starzix 3.0 NN              :   900  :  137+  :  585=  :  178-  :   429.5  : 179064.25  :  47.72%      3       18       94       92       90
12. Texel 1.11 NN               :   900  :  134+  :  562=  :  204-  :   415.0  : 173937.75  :  46.11%     20        1       77       91       89
13. Wasp 6.63 NN dev            :   900  :  126+  :  568=  :  206-  :   410.0  : 172626.25  :  45.56%     16        4       81       82       84
14. Wasp 6.50 NN                :   900  :  128+  :  558=  :  214-  :   407.0  : 168940.25  :  45.22%     28        1       78       82       84
15. Lizard 10.1 NN              :   900  :  121+  :  567=  :  212-  :   404.5  : 169247.75  :  44.94%      7       17       87       98       93
16. Renegade 1.0.0 NN           :   900  :   80+  :  541=  :  279-  :   350.5  : 147920.25  :  38.94%      2       30       91      101       95
17. Clarity 4.1.0 NN            :   900  :   58+  :  536=  :  306-  :   326.0  : 138555.75  :  36.22%      0       25       97       84       85
18. Avalanche 2.1.0 NN          :   900  :   46+  :  423=  :  431-  :   257.5  : 109493.00  :  28.61%      0       61       92       90       84
19. Counter 5.5 NN              :   900  :   38+  :  419=  :  443-  :   247.5  : 106198.75  :  27.50%      3       28       84      104       96


White Wins =  2.255 ( 26.37% )
Draws      =  5.152 ( 60.26% )
Black Wins =  1.143 ( 13.37% )
Average    = 177.06 ( 88,53 moves )

Code: Select all

   # Player                   :      Elo  Games  Score%  won  draw  lost  Points  Draw%  Error   OppAvg   OppE   OppD
   1 RubiChess 20240112 NN    :  3503.63    900    72.0  402   492     6   648.0   54.7  14.13  3325.60  13.01   18.0
   2 Caissa 1.16 NN           :  3491.74    900    70.6  378   515     7   635.5   57.2  14.47  3326.26  12.99   18.0
   3 Seer 2.8.0 NN            :  3469.08    900    67.9  349   524    27   611.0   58.2  14.11  3327.52  13.01   18.0
   4 Obsidian 10.0 NN         :  3467.39    900    67.7  329   560    11   609.0   62.2  13.93  3327.61  13.02   18.0
   5 Arasan 24.1 NN           :  3391.76    900    57.8  223   595    82   520.5   66.1  12.83  3331.81  13.08   18.0
   6 Stormphrax 4.0.0 NN      :  3368.52    900    54.7  227   530   143   492.0   58.9  12.58  3333.10  13.09   18.0
   7 Minic 3.40 NN            :  3333.45    900    49.8  152   593   155   448.5   65.9  12.57  3335.05  13.10   18.0
   8 Akimbo 0.8.0 NN          :  3332.25    900    49.7  162   570   168   447.0   63.3  11.92  3335.12  13.13   18.0
   9 Velvet 6.0.0 NN          :  3331.04    900    49.5  161   569   170   445.5   63.2  12.52  3335.19  13.10   18.0
   9 Pawn 3.0 NN              :  3331.04    900    49.5  147   597   156   445.5   66.3  11.96  3335.19  13.13   18.0
  11 Starzix 3.0 NN           :  3318.27    900    47.7  137   585   178   429.5   65.0  12.47  3335.90  13.10   18.0
  12 Texel 1.11 NN            :  3306.47    900    46.1  134   562   204   415.0   62.4  12.31  3336.55  13.11   18.0
  13 Wasp 6.63 NN dev         :  3302.43    900    45.6  126   568   206   410.0   63.1  12.38  3336.78  13.11   18.0
  14 Wasp 6.50 NN             :  3300.00    900    45.2  128   558   214   407.0   62.0  12.74  3336.91  13.09   18.0
  15 Lizard 10.1 NN           :  3297.97    900    44.9  121   567   212   404.5   63.0  12.23  3337.02  13.11   18.0
  16 Renegade 1.0.0 NN        :  3253.57    900    38.9   80   541   279   350.5   60.1  13.25  3339.49  13.06   18.0
  17 Clarity 4.1.0 NN         :  3232.82    900    36.2   58   536   306   326.0   59.6  13.09  3340.64  13.07   18.0
  18 Avalanche 2.1.0 NN       :  3171.27    900    28.6   46   423   431   257.5   47.0  14.25  3344.06  13.00   18.0
  19 Counter 5.5 NN           :  3161.69    900    27.5   38   419   443   247.5   46.6  14.53  3344.59  12.99   18.0

White advantage = 52.25 +/- 2.28
Draw rate (equal opponents) = 75.40 % +/- 0.65
Frank Quisinsky
Posts: 6888
Joined: Wed Nov 18, 2009 7:16 pm
Location: Gutweiler, Germany
Full name: Frank Quisinsky

Re: New engine releases 2024 ... 6+3

Post by Frank Quisinsky »

1.
I saw that I had the wrong link in the replay-zone. The link to this test can be found in the replay-zone to my 66+6 tourney.
https://www.amateurschach.de/fling/index.html

2.
log file updated on some positions.
The tourney is new and I produced here some mistakes ... now it's better.

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

System Wasp-1 = Intel® Core™ i9-10900k = 10 cores, 4.4Ghz overclocked, game in 4 minutes + 2 seconds
System Wasp-2 = Intel® Core™ i7-1185g7 =  4 cores, 3.0Ghz overclocked, game in 6 minutes + 3 seconds, rather rarely used

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

Start point December 17th, 2023 - until ongoing / continuously
Games can be donload under: https://www.amateurschach.de/fling/index.html

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

// todo

24. RukChess 3.0.18 NN                                   executable = December 17th, 2023    Test on Wasp-1 = soon
    Ilya Rukavishnikov, RUS
    https://github.com/Ilya-Ruk/RukChess

23. Starzix 4.0 NN                                       executable = January 20th, 2024     Test on Wasp-1 = still running
(2) Starzix 4.0 NN replaces Starzix 3.0 NN
    Ricardo Pinto, POR
    https://github.com/zzzzz151/Starzix

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

// ready

22. Stormphrax 4.0.0 NN                                  executable = December 17th, 2023    Test on Wasp-1  // v.22
    Conor Anstey, GBR
    https://github.com/Ciekce/Stormphrax

21. Obsidian 10.0 NN                                     released   = January 17th, 2024     Test on Wasp=1
    Gabriele Lombardo, ITA
    https://github.com/gab8192/Obsidian

    Engine support syzygy first time but produced some bugs and lost around 30 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
    - does not give mate in some other endgame constellations!
    - 50-moves rules will be avoided more times, very long draw games

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

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

15. 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

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

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

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

11. 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
    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 ... maybe later this year I will open a 2nd League
I believe not more as ~3150 Elo or engine produced bugs / problems.


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

18. Drofa 4.1.0 NN                                       executable = December 24th, 2023     no test for the moment
    Alexander Litov & Rhys Rustad-Elliott, RUS & CAN
    https://github.com/justNo4b/Drofa

17. Peacekeeper 2.20 NN                                  executable = December 24th, 2023     no test for the moment
    Kyle Zhang, USA
    https://github.com/Sazgr/peacekeeper

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

10. Kuma 1.2 NN                                          released   = January 12th, 2024      no test for the moment
    Kato Daichi, JPN
    https://github.com/kato-daichi/kuma
Frank Quisinsky
Posts: 6888
Joined: Wed Nov 18, 2009 7:16 pm
Location: Gutweiler, Germany
Full name: Frank Quisinsky

Re: New engine releases 2024 ... 6+3

Post by Frank Quisinsky »

Code: Select all

Starzix 4.0 NN - Wasp 6.50 NN                5.5	-	2.5		68.75%		
Starzix 4.0 NN - Wasp 6.63 NN dev            4.0	-	4.0		50.00%		
Starzix 4.0 NN - Akimbo 0.8.0 NN             3.5	-	4.5		43.75%		
Starzix 4.0 NN - Clarity 4.1.0 NN            5.0	-	4.0		55.56%		
Starzix 4.0 NN - Caissa 1.16 NN              2.0	-	7.0		22.22%		
Starzix 4.0 NN - Counter 5.5 NN              6.0	-	2.0		75.00%		
Starzix 4.0 NN - Pawn 3.0 NN                 4.5	-	3.5		56.25%		
Starzix 4.0 NN - Texel 1.11 NN               6.5	-	1.5		81.25%		
Starzix 4.0 NN - RubiChess 20240112 NN       3.0	-	5.0		37.50%		
Starzix 4.0 NN - Avalanche 2.1.0 NN          5.5	-	2.5		68.75%		
Starzix 4.0 NN - Lizard 10.1 NN              4.0	-	4.0		50.00%		
Starzix 4.0 NN - Minic 3.40 NN               5.0	-	3.0		62.50%		
Starzix 4.0 NN - Renegade 1.0.0 NN           5.5	-	2.5		68.75%		
Starzix 4.0 NN - Arasan 24.1 NN              3.5	-	4.5		43.75%		
Starzix 4.0 NN - Seer 2.8.0 NN               2.5	-	5.5		31.25%		
Starzix 4.0 NN - Velvet 6.0.0 NN             4.0	-	4.0		50.00%		
Starzix 4.0 NN - Obsidian 10.0 NN            2.5	-	5.5		31.25%		
Starzix 4.0 NN - Stormphrax 4.0.0 NN         4.5	-	3.5		56.25%

77.0	-	69.0		52.74%		
146 out of 900 games played
A good start ...
Frank Quisinsky
Posts: 6888
Joined: Wed Nov 18, 2009 7:16 pm
Location: Gutweiler, Germany
Full name: Frank Quisinsky

Re: New engine releases 2024 ... 6+3

Post by Frank Quisinsky »

In the maintime I made an important update in my *.zip file.
- PGN database besser sorted now.

Can be found here:
https://www.amateurschach.de/fling/index.html

Starzix 4.0 NN results are very stabil ...
The final results is tomorrow in the morning available.

Code: Select all

Starzix 4.0 NN - Wasp 6.50 NN              18.0	-	 9.0		66.67%		
Starzix 4.0 NN - Wasp 6.63 NN dev          15.0	-	12.0		55.56%		
Starzix 4.0 NN - Akimbo 0.8.0 NN           13.5	-	13.5		50.00%		
Starzix 4.0 NN - Clarity 4.1.0 NN          18.5	-	 7.5		71.15%		
Starzix 4.0 NN - Caissa 1.16 NN             9.0	-	17.0		34.62%		
Starzix 4.0 NN - Counter 5.5 NN            19.5	-	 6.5		75.00%		
Starzix 4.0 NN - Pawn 3.0 NN               12.5	-	13.5		48.08%		
Starzix 4.0 NN - Texel 1.11 NN             15.0	-	11.0		57.69%		
Starzix 4.0 NN - RubiChess 20240112 NN      8.5	-	17.5		32.69%		
Starzix 4.0 NN - Avalanche 2.1.0 NN        17.0	-	 8.0		68.00%		
Starzix 4.0 NN - Lizard 10.1 NN            15.0	-	11.0		57.69%		
Starzix 4.0 NN - Minic 3.40 NN             14.5	-	11.5		55.77%		
Starzix 4.0 NN - Renegade 1.0.0 NN         18.0	-	 8.0		69.23%		
Starzix 4.0 NN - Arasan 24.1 NN            12.0	-	14.0		46.15%		
Starzix 4.0 NN - Seer 2.8.0 NN              9.0	-	17.0		34.62%		
Starzix 4.0 NN - Velvet 6.0.0 NN           13.5	-	11.5		54.00%		
Starzix 4.0 NN - Obsidian 10.0 NN           9.5	-	16.5		36.54%		
Starzix 4.0 NN - Stormphrax 4.0.0 NN        8.5	-	16.5		34.00%

246.5	-	221.5		52.67%
468 out of 900 games played
Frank Quisinsky
Posts: 6888
Joined: Wed Nov 18, 2009 7:16 pm
Location: Gutweiler, Germany
Full name: Frank Quisinsky

Re: New engine releases 2024 ... 6+3

Post by Frank Quisinsky »

23. Starzix 4.0 NN ... ready
24. RukChess 3.0.18 NN ... stopped, I believe not over 3150 Elo
25. Clover 6.1 NN ... still running (release from December 16th, 2023)
26. Altair 6.0.0 NN ... soon (release from December 05th, 2023)

if ready all stronger releases (more as 3150 Elo) since start of December 2023 inside!

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       :   900  :  394+  :  499=  :    7-  :   643.5  : 273569.75  :  71.50%     43        0        80       82       81
02. Caissa 1.16 NN              :   900  :  375+  :  518=  :    7-  :   634.0  : 268697.00  :  70.44%     23        0        86       79       82
03. Obsidian 10.0 NN            :   900  :  324+  :  565=  :   11-  :   606.5  : 259501.75  :  67.39%     16        1        90      102       97
04. Seer 2.8.0 NN               :   900  :  339+  :  533=  :   28-  :   605.5  : 255635.25  :  67.28%     20        0        86       84       85
05. Arasan 24.1 NN              :   900  :  220+  :  591=  :   89-  :   515.5  : 218519.75  :  57.28%     26        3        84       85       85
06. Stormphrax 4.0.0 NN         :   900  :  225+  :  527=  :  148-  :   488.5  : 202763.50  :  54.28%      3       19        92      109      100
07. Starzix 4.0 NN              :   900  :  178+  :  595=  :  127-  :   475.5  : 200507.00  :  52.83%      6        9        89       93       92
08. Akimbo 0.8.0 NN             :   900  :  159+  :  577=  :  164-  :   447.5  : 187005.75  :  49.72%      6       15        95       93       92
09. Minic 3.40 NN               :   900  :  148+  :  598=  :  154-  :   447.0  : 188878.75  :  49.67%      3       17        99       93       91
10. Velvet 6.0.0 NN             :   900  :  155+  :  582=  :  163-  :   446.0  : 188075.50  :  49.56%     18        0        75       89       88
---------------------------------------------------------------------------------------------------------------------------------------------------
11. Pawn 3.0 NN                 :   900  :  146+  :  595=  :  159-  :   443.5  : 187101.00  :  49.28%      4        9        83       91       89
12. Texel 1.11 NN               :   900  :  133+  :  553=  :  214-  :   409.5  : 171798.25  :  45.50%     20        1        77       91       89
13. Wasp 6.63 NN dev            :   900  :  123+  :  570=  :  207-  :   408.0  : 171958.00  :  45.33%     16        4        81       82       84
14. Lizard 10.1 NN              :   900  :  119+  :  567=  :  214-  :   402.5  : 168384.00  :  44.72%      7       17        87       98       92
15. Wasp 6.50 NN                :   900  :  124+  :  556=  :  220-  :   402.0  : 166936.25  :  44.67%     27        1        78       81       84
16. Renegade 1.0.0 NN           :   900  :   81+  :  526=  :  293-  :   344.0  : 145096.00  :  38.22%      2       32        90      101       94
17. Clarity 4.1.0 NN            :   900  :   61+  :  525=  :  314-  :   323.5  : 137480.75  :  35.94%      0       25        96       84       85
18. Avalanche 2.1.0 NN          :   900  :   45+  :  428=  :  427-  :   259.0  : 110163.00  :  28.78%      0       60        92       90       85
19. Counter 5.5 NN              :   900  :   39+  :  419=  :  442-  :   248.5  : 106675.75  :  27.61%      3       30        85      103       95


White Wins =  2.246 ( 26.27% )
Draws      =  5.162 ( 60.37% )
Black Wins =  1.142 ( 13.36% )
Average    = 177.45 ( 88,72 moves )

Code: Select all

   # Player                   :      Elo  Games  Score%  won  draw  lost  Points  Draw%  Error   OppAvg   OppE   OppD
   1 RubiChess 20240112 NN    :  3503.05    900    71.5  394   499     7   643.5   55.4  14.06  3329.74  12.82   18.0
   2 Caissa 1.16 NN           :  3494.08    900    70.4  375   518     7   634.0   57.6  14.16  3330.24  12.81   18.0
   3 Obsidian 10.0 NN         :  3468.94    900    67.4  324   565    11   606.5   62.8  13.63  3331.63  12.84   18.0
   4 Seer 2.8.0 NN            :  3468.04    900    67.3  339   533    28   605.5   59.2  13.50  3331.68  12.85   18.0
   5 Arasan 24.1 NN           :  3391.59    900    57.3  220   591    89   515.5   65.7  12.81  3335.93  12.89   18.0
   6 Stormphrax 4.0.0 NN      :  3369.65    900    54.3  225   527   148   488.5   58.6  12.64  3337.15  12.89   18.0
   7 Starzix 4.0 NN           :  3359.16    900    52.8  178   595   127   475.5   66.1  12.37  3337.73  12.91   18.0
   8 Akimbo 0.8.0 NN          :  3336.66    900    49.7  159   577   164   447.5   64.1  12.38  3338.98  12.91   18.0
   9 Minic 3.40 NN            :  3336.25    900    49.7  148   598   154   447.0   66.4  11.96  3339.00  12.93   18.0
  10 Velvet 6.0.0 NN          :  3335.45    900    49.6  155   582   163   446.0   64.7  12.17  3339.05  12.92   18.0
  11 Pawn 3.0 NN              :  3333.44    900    49.3  146   595   159   443.5   66.1  12.36  3339.16  12.91   18.0
  12 Texel 1.11 NN            :  3306.07    900    45.5  133   553   214   409.5   61.4  11.74  3340.68  12.94   18.0
  13 Wasp 6.63 NN dev         :  3304.86    900    45.3  123   570   207   408.0   63.3  12.34  3340.75  12.91   18.0
  14 Lizard 10.1 NN           :  3300.41    900    44.7  119   567   214   402.5   63.0  12.22  3341.00  12.92   18.0
  15 Wasp 6.50 NN             :  3300.00    900    44.7  124   556   220   402.0   61.8  12.19  3341.02  12.92   18.0
  16 Renegade 1.0.0 NN        :  3252.22    900    38.2   81   526   293   344.0   58.4  12.38  3343.67  12.91   18.0
  17 Clarity 4.1.0 NN         :  3234.79    900    35.9   61   525   314   323.5   58.3  12.96  3344.64  12.88   18.0
  18 Avalanche 2.1.0 NN       :  3176.86    900    28.8   45   428   427   259.0   47.6  14.21  3347.86  12.81   18.0
  19 Counter 5.5 NN           :  3166.82    900    27.6   39   419   442   248.5   46.6  14.65  3348.42  12.78   18.0

White advantage = 51.80 +/- 2.32
Draw rate (equal opponents) = 75.32 % +/- 0.64
Starzix 4.0 NN replaces Starzix 3.0 NN = Elo gain = 40,73

And here the current Clover 6.1 NN results ...

Code: Select all

Clover 6.1 NN - Wasp 6.50 NN              4.0	-	2.0		66.67%		
Clover 6.1 NN - Wasp 6.63 NN dev          4.0	-	2.0		66.67%		
Clover 6.1 NN - Akimbo 0.8.0 NN           3.5	-	2.5		58.33%		
Clover 6.1 NN - Clarity 4.1.0 NN          4.0	-	2.0		66.67%		
Clover 6.1 NN - Caissa 1.16 NN            2.5	-	3.5		41.67%		
Clover 6.1 NN - Counter 5.5 NN            6.0	-	0.0		100.00%		
Clover 6.1 NN - Pawn 3.0 NN               3.5	-	2.5		58.33%		
Clover 6.1 NN - Texel 1.11 NN             4.0	-	2.0		66.67%		
Clover 6.1 NN - RubiChess 20240112 NN     2.0	-	4.0		33.33%		
Clover 6.1 NN - Avalanche 2.1.0 NN        5.5	-	0.5		91.67%		
Clover 6.1 NN - Lizard 10.1 NN            3.0	-	3.0		50.00%		
Clover 6.1 NN - Minic 3.40 NN             4.0	-	2.0		66.67%		
Clover 6.1 NN - Renegade 1.0.0 NN         4.5	-	1.5		75.00%		
Clover 6.1 NN - Arasan 24.1 NN            3.5	-	2.5		58.33%		
Clover 6.1 NN - Seer 2.8.0 NN             3.0	-	3.0		50.00%		
Clover 6.1 NN - Velvet 6.0.0 NN           5.0	-	1.0		83.33%		
Clover 6.1 NN - Obsidian 10.0 NN          3.0	-	3.0		50.00%		
Clover 6.1 NN - Stormphrax 4.0.0 NN       2.0	-	3.0		40.00%		
Clover 6.1 NN - Starzix 4.0 NN            4.0	-	2.0		66.67%		

71.0	-	42.0		62.83%
113 out of 950 games played
If you like you can download the results in replay-zone to my 66+6 tournament:
https://www.amateurschach.de/fling/index.html

Best
Frank
Frank Quisinsky
Posts: 6888
Joined: Wed Nov 18, 2009 7:16 pm
Location: Gutweiler, Germany
Full name: Frank Quisinsky

Re: New engine releases 2024 ... 6+3

Post by Frank Quisinsky »

Code: Select all

Clover 6.1 NN - Wasp 6.50 NN              21.5	-	 8.5		71.67%		
Clover 6.1 NN - Wasp 6.63 NN dev          19.5	-	 9.5		67.24%		
Clover 6.1 NN - Akimbo 0.8.0 NN           19.5	-	10.5		65.00%		
Clover 6.1 NN - Clarity 4.1.0 NN          20.0	-	 9.0		68.97%		
Clover 6.1 NN - Caissa 1.16 NN            12.0	-	17.0		41.38%		
Clover 6.1 NN - Counter 5.5 NN            26.0	-	 3.0		89.66%		
Clover 6.1 NN - Pawn 3.0 NN               18.5	-	10.5		63.79%		
Clover 6.1 NN - Texel 1.11 NN             21.0	-	 8.0		72.41%		
Clover 6.1 NN - RubiChess 20240112 NN     12.5	-	16.5		43.10%		
Clover 6.1 NN - Avalanche 2.1.0 NN        24.5	-	 4.5		84.48%		
Clover 6.1 NN - Lizard 10.1 NN            20.0	-	 9.0		68.97%		
Clover 6.1 NN - Minic 3.40 NN             17.5	-	11.5		60.34%		
Clover 6.1 NN - Renegade 1.0.0 NN         23.0	-	 6.0		79.31%		
Clover 6.1 NN - Arasan 24.1 NN            15.5	-	13.5		53.45%		
Clover 6.1 NN - Seer 2.8.0 NN             13.0	-	16.0		44.83%		
Clover 6.1 NN - Velvet 6.0.0 NN           20.0	-	 8.0		71.43%		
Clover 6.1 NN - Obsidian 10.0 NN          15.5	-	12.5		55.36%		
Clover 6.1 NN - Stormphrax 4.0.0 NN       18.5	-	10.5		63.79%		
Clover 6.1 NN - Starzix 4.0 NN            19.5	-	 8.5		69.64%		

357.5	-	192.5		65.00%
550 out of 950 games played
Looks normal for the good known strength Clover 6.1 NN have.
The final results are tomorrow available.
After I start direct Altair 6.0.0 NN and must thinking what do as next.
Maybe all the new releases from Nov. 2023 or a dev from Stockfish ... I don't know.

Best
Frank
Frank Quisinsky
Posts: 6888
Joined: Wed Nov 18, 2009 7:16 pm
Location: Gutweiler, Germany
Full name: Frank Quisinsky

Re: New engine releases 2024 ... 6+3

Post by Frank Quisinsky »

25. Clover 6.1 NN
26. Altair 6.0.0 NN ... still running
27. Stockifish dev ... soon

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       :   950  :  401+  :  542=  :    7-  :   672.0  : 301479.75  :  70.74%     45         0       80       84       82
02. Caissa 1.16 NN              :   950  :  381+  :  561=  :    8-  :   661.5  : 295807.50  :  69.63%     23         0       86       80       83
03. Seer 2.8.0 NN               :   950  :  345+  :  576=  :   29-  :   633.0  : 282215.50  :  66.63%     20         0       86       85       86
04. Obsidian 10.0 NN            :   950  :  326+  :  607=  :   17-  :   629.5  : 283554.50  :  66.26%     16         1       90      102       98
05. Clover 6.1 NN               :   950  :  336+  :  587=  :   27-  :   629.5  : 280812.00  :  66.26%     18         2       88      103       97
06. Arasan 24.1 NN              :   950  :  223+  :  629=  :   98-  :   537.5  : 240313.00  :  56.58%     26         3       84       84       87
07. Stormphrax 4.0.0 NN         :   950  :  226+  :  560=  :  164-  :   506.0  : 221136.25  :  53.26%      3        20       91      110      101
08. Starzix 4.0 NN              :   950  :  178+  :  624=  :  148-  :   490.0  : 217063.25  :  51.58%      6         9       89       94       92
09. Minic 3.40 NN               :   950  :  148+  :  632=  :  170-  :   464.0  : 206493.75  :  48.84%      3        17       99       93       91
10. Akimbo 0.8.0 NN             :   950  :  159+  :  608=  :  183-  :   463.0  : 203589.00  :  48.74%      6        17       95       93       92
---------------------------------------------------------------------------------------------------------------------------------------------------
11. Pawn 3.0 NN                 :   950  :  146+  :  628=  :  176-  :   460.0  : 204346.25  :  48.42%      4        10       83       91       89
12. Velvet 6.0.0 NN             :   950  :  156+  :  607=  :  187-  :   459.5  : 203537.25  :  48.37%     18         0       75       90       89
13. Texel 1.11 NN               :   950  :  133+  :  582=  :  235-  :   424.0  : 187176.00  :  44.63%     20         1       77       91       89
14. Wasp 6.63 NN dev            :   950  :  124+  :  598=  :  228-  :   423.0  : 187659.25  :  44.53%     16         4       81       81       84
15. Lizard 10.1 NN              :   950  :  119+  :  595=  :  236-  :   416.5  : 183358.25  :  43.84%      7        18       87       98       93
16. Wasp 6.50 NN                :   950  :  124+  :  585=  :  241-  :   416.5  : 182116.25  :  43.84%     27         1       78       81       84
17. Renegade 1.0.0 NN           :   950  :   81+  :  550=  :  319-  :   356.0  : 157940.25  :  37.47%      2        37       90      101       94
18. Clarity 4.1.0 NN            :   950  :   61+  :  552=  :  337-  :   337.0  : 150968.25  :  35.47%      0        25       96       85       85
19. Avalanche 2.1.0 NN          :   950  :   45+  :  443=  :  462-  :   266.5  : 118895.75  :  28.05%      0        66       92       91       85   
20. Counter 5.5 NN              :   950  :   39+  :  432=  :  479-  :   255.0  : 114693.00  :  26.84%      3        32       85      104       95


White Wins =  2.471 ( 26.01% )
Draws      =  5.749 ( 60.52% )
Black Wins =  1.280 ( 13.47% )
Average    = 179.05 ( 89,52 moves )

Code: Select all

   # Player                   :      Elo  Games  Score%  won  draw  lost  Points  Draw%  Error   OppAvg   OppE   OppD
   1 RubiChess 20240112 NN    :  3503.48    950    70.7  401   542     7   672.0   57.1  13.81  3336.59  12.33   19.0
   2 Caissa 1.16 NN           :  3494.17    950    69.6  381   561     8   661.5   59.1  13.01  3337.08  12.37   19.0
   3 Seer 2.8.0 NN            :  3469.66    950    66.6  345   576    29   633.0   60.6  12.87  3338.37  12.38   19.0
   4 Clover 6.1 NN            :  3466.71    950    66.3  336   587    27   629.5   61.8  12.95  3338.53  12.38   19.0
   5 Obsidian 10.0 NN         :  3466.71    950    66.3  326   607    17   629.5   63.9  12.43  3338.53  12.41   19.0
   6 Arasan 24.1 NN           :  3392.91    950    56.6  223   629    98   537.5   66.2  11.43  3342.41  12.46   19.0
   7 Stormphrax 4.0.0 NN      :  3368.63    950    53.3  226   560   164   506.0   58.9  11.31  3343.69  12.46   19.0
   8 Starzix 4.0 NN           :  3356.37    950    51.6  178   624   148   490.0   65.7  11.56  3344.33  12.45   19.0
   9 Minic 3.40 NN            :  3336.49    950    48.8  148   632   170   464.0   66.5  11.88  3345.38  12.43   19.0
  10 Akimbo 0.8.0 NN          :  3335.73    950    48.7  159   608   183   463.0   64.0  11.66  3345.42  12.45   19.0
  11 Pawn 3.0 NN              :  3333.43    950    48.4  146   628   176   460.0   66.1  12.22  3345.54  12.42   19.0
  12 Velvet 6.0.0 NN          :  3333.05    950    48.4  156   607   187   459.5   63.9  11.32  3345.56  12.46   19.0
  13 Texel 1.11 NN            :  3305.80    950    44.6  133   582   235   424.0   61.3  12.07  3346.99  12.42   19.0
  14 Wasp 6.63 NN dev         :  3305.02    950    44.5  124   598   228   423.0   62.9  11.99  3347.04  12.43   19.0
  15 Lizard 10.1 NN           :  3300.00    950    43.8  119   595   236   416.5   62.6  12.02  3347.30  12.43   19.0
  16 Wasp 6.50 NN             :  3300.00    950    43.8  124   585   241   416.5   61.6  12.19  3347.30  12.42   19.0
  17 Renegade 1.0.0 NN        :  3252.31    950    37.5   81   550   319   356.0   57.9  12.48  3349.81  12.40   19.0
  18 Clarity 4.1.0 NN         :  3236.84    950    35.5   61   552   337   337.0   58.1  12.53  3350.62  12.40   19.0
  19 Avalanche 2.1.0 NN       :  3175.99    950    28.1   45   443   462   266.5   46.6  13.53  3353.83  12.35   19.0
  20 Counter 5.5 NN           :  3165.37    950    26.8   39   432   479   255.0   45.5  14.88  3354.39  12.28   19.0

White advantage = 50.52 +/- 2.16
Draw rate (equal opponents) = 76.18 % +/- 0.63
Frank Quisinsky
Posts: 6888
Joined: Wed Nov 18, 2009 7:16 pm
Location: Gutweiler, Germany
Full name: Frank Quisinsky

Re: New engine releases 2024 ... 6+3

Post by Frank Quisinsky »

26. Altair 6.0.0 NN
27. Stockfish 20240121 NN dev ... replaced Obsidian 1.0 NN (syzygy bug) ... still running
28. Midnight 9.0 NN ... replaced Counter 5.5 NN ... soon

I go higher, from min. 3150 Elo to min 3200 Elo.
For that reason Counter 5.5 NN and after Avalanche 2.1.0 NN will be replaced as next

In my download file can be found the conditions, games and log files to the tourney.

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

Here the Altair 6.0.0 NN results ...

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       :  1000  :  425+  :  568=  :    7-  :   709.0  : 335646.75  :  70.90%     49         0       79       84       82
02. Caissa 1.16 NN              :  1000  :  408+  :  583=  :    9-  :   699.5  : 330026.00  :  69.95%     28         0       86       80       84
03. Seer 2.8.0 NN               :  1000  :  370+  :  601=  :   29-  :   670.5  : 315388.00  :  67.05%     23         0       86       85       86
04. Obsidian 10.0 NN            :  1000  :  350+  :  633=  :   17-  :   666.5  : 316629.75  :  66.65%     17         1       89      102       97
05. Clover 6.1 NN               :  1000  :  358+  :  615=  :   27-  :   665.5  : 313266.00  :  66.55%     25         2       87      102       96
06. Arasan 24.1 NN              :  1000  :  237+  :  662=  :  101-  :   568.0  : 267971.00  :  56.80%     29         3       84       87       87
07. Stormphrax 4.0.0 NN         :  1000  :  239+  :  596=  :  165-  :   537.0  : 247858.50  :  53.70%      3        20       91      110      101
08. Starzix 4.0 NN              :  1000  :  183+  :  660=  :  157-  :   513.0  : 240082.00  :  51.30%      6         9       89       94       92
09. Minic 3.40 NN               :  1000  :  161+  :  665=  :  174-  :   493.5  : 231611.75  :  49.35%      3        17       99       93       92
10. Velvet 6.0.0 NN             :  1000  :  172+  :  636=  :  192-  :   490.0  : 228954.00  :  49.00%     22         0       75       90       89
---------------------------------------------------------------------------------------------------------------------------------------------------
11. Pawn 3.0 NN                 :  1000  :  157+  :  665=  :  178-  :   489.5  : 229343.00  :  48.95%      4        10       83       90       89
12. Akimbo 0.8.0 NN             :  1000  :  165+  :  648=  :  187-  :   489.0  : 227081.50  :  48.90%      6        17       95       93       91
13. Wasp 6.63 NN dev            :  1000  :  136+  :  631=  :  233-  :   451.5  : 211225.25  :  45.15%     17         4       82       81       84
14. Altair 6.0.0 NN             :  1000  :  142+  :  614=  :  244-  :   449.0  : 207659.00  :  44.90%      1        32       97       91       89
15. Texel 1.11 NN               :  1000  :  139+  :  616=  :  245-  :   447.0  : 208413.25  :  44.70%     22         1       77       91       89
16. Wasp 6.50 NN                :  1000  :  131+  :  623=  :  246-  :   442.5  : 204301.00  :  44.25%     28         1       78       82       85
17. Lizard 10.1 NN              :  1000  :  128+  :  621=  :  251-  :   438.5  : 203906.25  :  43.85%      8        19       87       98       92
18. Renegade 1.0.0 NN           :  1000  :   85+  :  579=  :  336-  :   374.5  : 175463.00  :  37.45%      2        37       90      101       94
19. Clarity 4.1.0 NN            :  1000  :   63+  :  587=  :  350-  :   356.5  : 168480.00  :  35.65%      0        25       95       85       85
20. Avalanche 2.1.0 NN          :  1000  :   45+  :  473=  :  482-  :   281.5  : 132502.00  :  28.15%      0        66       92       91       85
---------------------------------------------------------------------------------------------------------------------------------------------------
21. Counter 5.5 NN              :  1000  :   43+  :  450=  :  507-  :   268.0  : 127233.00  :  26.80%      3        32       85      104       96


White Wins =  2.727 ( 25.97% )
Draws      =  6.363 ( 60.60% )
Black Wins =  1.410 ( 13.43% )
Average    = 178.92 ( 89,46 moves )

Code: Select all

   # Player                   :      Elo  Games  Score%  won  draw  lost  Points  Draw%  Error   OppAvg   OppE   OppD
   1 RubiChess 20240112 NN    :  3502.09   1000    70.9  425   568     7   709.0   56.8  13.44  3334.04  12.27   20.0
   2 Caissa 1.16 NN           :  3494.04   1000    70.0  408   583     9   699.5   58.3  13.62  3334.45  12.27   20.0
   3 Seer 2.8.0 NN            :  3470.17   1000    67.0  370   601    29   670.5   60.1  12.69  3335.64  12.31   20.0
   4 Obsidian 10.0 NN         :  3466.95   1000    66.7  350   633    17   666.5   63.3  13.24  3335.80  12.28   20.0
   5 Clover 6.1 NN            :  3466.15   1000    66.5  358   615    27   665.5   61.5  12.69  3335.84  12.31   20.0
   6 Arasan 24.1 NN           :  3391.60   1000    56.8  237   662   101   568.0   66.2  11.97  3339.57  12.35   20.0
   7 Stormphrax 4.0.0 NN      :  3368.85   1000    53.7  239   596   165   537.0   59.6  11.77  3340.70  12.36   20.0
   8 Starzix 4.0 NN           :  3351.37   1000    51.3  183   660   157   513.0   66.0  11.59  3341.58  12.37   20.0
   9 Minic 3.40 NN            :  3337.20   1000    49.4  161   665   174   493.5   66.5  11.30  3342.29  12.38   20.0
  10 Velvet 6.0.0 NN          :  3334.66   1000    49.0  172   636   192   490.0   63.6  11.51  3342.41  12.37   20.0
  11 Pawn 3.0 NN              :  3334.30   1000    49.0  157   665   178   489.5   66.5  12.01  3342.43  12.35   20.0
  12 Akimbo 0.8.0 NN          :  3333.93   1000    48.9  165   648   187   489.0   64.8  11.37  3342.45  12.38   20.0
  13 Wasp 6.63 NN dev         :  3306.60   1000    45.1  136   631   233   451.5   63.1  11.70  3343.82  12.36   20.0
  14 Altair 6.0.0 NN          :  3304.77   1000    44.9  142   614   244   449.0   61.4  11.54  3343.91  12.37   20.0
  15 Texel 1.11 NN            :  3303.30   1000    44.7  139   616   245   447.0   61.6  11.56  3343.98  12.37   20.0
  16 Wasp 6.50 NN             :  3300.00   1000    44.3  131   623   246   442.5   62.3  11.92  3344.15  12.35   20.0
  17 Lizard 10.1 NN           :  3297.06   1000    43.9  128   621   251   438.5   62.1  11.79  3344.29  12.36   20.0
  18 Renegade 1.0.0 NN        :  3249.10   1000    37.5   85   579   336   374.5   57.9  12.28  3346.69  12.33   20.0
  19 Clarity 4.1.0 NN         :  3235.18   1000    35.6   63   587   350   356.5   58.7  12.94  3347.39  12.30   20.0
  20 Avalanche 2.1.0 NN       :  3173.74   1000    28.1   45   473   482   281.5   47.3  13.95  3350.46  12.25   20.0
  21 Counter 5.5 NN           :  3161.88   1000    26.8   43   450   507   268.0   45.0  14.02  3351.05  12.25   20.0

White advantage = 50.35 +/- 2.07
Draw rate (equal opponents) = 75.75 % +/- 0.58

Code: Select all

And here the first Stockfish 20240121 NN dev results ...

Stockfish 20240121 NN dev - Wasp 6.50 NN            26.5	-	 5.5		82.81%		
Stockfish 20240121 NN dev - Wasp 6.63 NN dev        27.5	-	 5.5		83.33%		
Stockfish 20240121 NN dev - Akimbo 0.8.0 NN         25.0	-	 7.0		78.13%		
Stockfish 20240121 NN dev - Clarity 4.1.0 NN        28.0	-	 4.0		87.50%		
Stockfish 20240121 NN dev - Caissa 1.16 NN          18.0	-	14.0		56.25%		
Stockfish 20240121 NN dev - Counter 5.5 NN          29.0	-	 3.0		90.63%		
Stockfish 20240121 NN dev - Pawn 3.0 NN             26.5	-	 5.5		82.81%		
Stockfish 20240121 NN dev - Texel 1.11 NN           26.0	-	 6.0		81.25%		
Stockfish 20240121 NN dev - RubiChess 20240112 NN   19.0	-	13.0		59.38%		
Stockfish 20240121 NN dev - Avalanche 2.1.0 NN      27.5	-	 4.5		85.94%		
Stockfish 20240121 NN dev - Lizard 10.1 NN          26.5	-	 5.5		82.81%		
Stockfish 20240121 NN dev - Minic 3.40 NN           25.0	-	 7.0		78.13%		
Stockfish 20240121 NN dev - Renegade 1.0.0 NN       28.0	-	 4.0		87.50%		
Stockfish 20240121 NN dev - Arasan 24.1 NN          23.5	-	 8.5		73.44%		
Stockfish 20240121 NN dev - Seer 2.8.0 NN           21.0	-	11.0		65.63%		
Stockfish 20240121 NN dev - Velvet 6.0.0 NN         26.0	-	 5.0		83.87%		
Stockfish 20240121 NN dev - Stormphrax 4.0.0 NN     26.5	-	 4.5		85.48%		
Stockfish 20240121 NN dev - Starzix 4.0 NN          23.0	-	 8.0		74.19%		
Stockfish 20240121 NN dev - Clover 6.1 NN           19.5	-	11.5		62.90%		
Stockfish 20240121 NN dev - Altair 6.0.0 NN         25.5	-	 5.5		82.26%
		
497.5	-	138.5		78.22%
636 out of 1000 games played
Note: 1% is around 8 Elo.
= at the moment the current Stockfish is around 60-70 Elo stronger as RubiChess 20240112 NN

All looks super for Stockfish, move-average wins and draws, quantity fast wins, no game lost ...
Move-average draws is much better as from v16 in my 40 in 20 tournament.
Move-average wins is at the moment 79, same RubiChess have. Texel and Velvet here the number 1. Wasp have 78 for wins. Maybe Stockfish can beat Wasp in move-average wins. Quantity of fast wins can be a bit better as the result from RubiChess. Tomorrow I have the final results.

At the moment I have not many time for chess.
The final results I can give tomorrow in the evening.

Best
Frank

PS:
A new Starzix 4.0 NN version (I saw newer executable files) but I will not test the engine for that reason again.
Minic 3.41 NN is available. A special version for Graham, I don't know and I will not test it again.
Midnight 9 NN is new and clearly improved and should be go clear over 3200 Elo.
The next after Stockfish here!