Breaking News (30.05.2023):Another new tourney, but Bullet time control, plus a new Unique suite is usedIn other words, this time my goal is to increase the wins..and note that so farThe statistics are very normal e.g White win. percentages: 56% / Blacks. 44%Where the current overall draw ratio is just 83%.. something like NON-NNUE!Here as one of main roles (about appearing less draws) played Unique book!More info about Unique book, it's a tiny suite,e.g based on 264 openings..andIt seems that under these conditions: some Top engines (I mean regarding forRaptora, Littlebeast, Brainlearn etc., which are with learning future) performedSlightly better...and exception the importance of played openings...plus I thinkTheirs 'learning future' can be as other main 'influence' about they are on TopAnd 'learning affect' is playing more role e.g after serious numbers of games!Plus it's clear that the learning is more effective with less number of openings!What I can say more, the engines (with learning future) are played with theirsDefault learning settings., but without using any huge pre-learned exp files etc.I mean they are used to play and learned from the current competition's gamesOk,.finely the tournament is over..many thanks and congrats to all participants!
The Winner: RapTora 2.2  - Congratulations to Antonio Ces, Stockfish team!Btw, Raptora is not so new (released 3 months ago) but a real beast for sure!Note also that exactly the same Raptora ver played in Book vs Experience III
Rank Name               Elo     +    - games score oppo. draws    1 RapTora 2.2        3770    5    5  2200   52%  3754   87%    2 LittleBeast 2 sl   3769    5    5  2200   52%  3754   87%    3 BrainLearn 24      3767    5    5  2200   53%  3751   87%    4 SugaR XPrO 140523  3766    5    5  2200   53%  3751   87%    5 SF-PB 200523       3765    5    5  2200   53%  3751   87%    6 EMAN 8.90          3764    5    5  2200   51%  3756   88%    7 SF-FS Apocalypse   3763    5    5  2200   51%  3757   90%    8 Stockfish 200523   3763    5    5  2200   52%  3752   88%    9 Swordfish 15.4     3762    6    6  1500   50%  3762   91%   10 Polyfish 100523    3762    5    5  2200   52%  3752   87%   11 PREDATOR X5        3762    5    5  2000   51%  3756   88%   12 AIO 1.02           3762    5    5  2200   51%  3756   89%   13 DarkSisTer 5       3761    5    5  2000   51%  3757   88%   14 Corchess 070523    3759    5    5  2100   54%  3740   86%   15 ShashChess 32      3754    5    5  1900   53%  3740   83%   16 Kookaburra 1.03a   3751    5    5  2000   49%  3756   89%   17 SF Polyglot 15.1   3751    5    5  1900   49%  3757   90%   18 DeepBlue 290323    3746    5    5  1900   50%  3746   86%   19 Genko 1.1          3746    5    5  1800   48%  3757   89%   20 Zeus 26.5          3744    5    5  1800   48%  3757   88%   21 AI 28.0            3743    5    5  1800   50%  3742   82%   22 BlueMarlin 15.7    3742    5    5  1900   51%  3733   81%   23 Kayra 1.7          3740    5    5  1800   50%  3742   87%   24 YuliG PrO 5        3739    6    6  1500   47%  3760   86%   25 Swordfish 15.1     3737    5    5  1700   54%  3721   77%   26 FisherovMonk 1.2   3734    5    5  1700   51%  3729   82%   27 Patzer AI 020223   3733    6    6  1600   48%  3744   85%   28 Crystal 6 PMT      3733    6    5  1600   50%  3735   82%   29 Raubfisch X48e     3730    5    5  1500   54%  3711   76%   30 Fat Titz 2         3725    6    6  1400   56%  3698   73%   31 CatroPOLY 060523   3720    6    6  1400   55%  3698   74%   32 BigBallz 010923    3718    6    6  1400   54%  3705   73%   33 KomodoDragon 3.2   3706    5    5  1500   51%  3705   73%   34 CFishNN 240223     3695    5    5  1400   49%  3704   72%   35 Berserk 11.1       3669    5    5  1400   47%  3690   68%   36 FatFritz 3         3651    5    5  1300   44%  3687   68%   37 Vafra 12.4 Cfish   3646    5    5  1400   34%  3716   57%   38 Koivisto 090223    3626    5    5  1200   38%  3685   61%   39 Fire-NN 190523     3552    6    6  1000   19%  3689   35% 
       • Calculated by Bayeselo / Raubfisch is fixed to 3730 Elo / Blue Name = Last Update
     More Details: • Almost all of NNUE engines are AVX2, tested via their default Evalfiles • For more fair conditions: each of the opening line is repeated as twice • Sure where possible: all of engines are played via move overhead: 100  • The overall draw percentage (based on 34700 games) is not high: 83% • All the played tour games are available (in PGN format): NNUE XXXII