Stefan Pohl Computer Chess

Latest Website-News (2020/01/25): AB-testrun of Stockfish 11 finished. Sadly a clear regression to Stockfish 200107 (-7 Elo). Next AB-testrun: Schooner 2.2

I developed 4 new (and better) settings for the Millenium King-Element and King Performance Chesscomputer. Learn more in the "Experiments"- section of my website.

Important news:  The bullet-NN-testing is discontinued... Check out the excellent new bullet-ratinglist of Andreas Strangmueller instead: here 

I decided to discontinue my bullet-NN-testing, because Andreas has a faster hardware and does nearly the same testing, as I did. So, for me, it makes no sense, to continue my testings. Because my hardware is slower (notebooks), it makes much more sense to continue my NN-testings with longer thinking-time. Checkout the new "NN longtime testing"- section of my website...

The "Lc0 / NN testing"- section (bullet NN-testings) will stay online for 1-2 months and then, I will delete it... So, if you want to download the played games, do it right now!

Stay tuned.

Stockfish testing

Playing conditions:

Hardware: i7-6700HQ 2.6GHz Notebook (Skylake CPU), Windows 10 64bit, 8GB RAM

Fritzmark: singlethread: 5.3 / 2521 (all engines running on one thread, only), average meganodes/s displayed by LittleBlitzerGUI: Houdini: 2.6 mn/s, Stockfish: 2.2 mn/s, Komodo: 2.0 mn/s

Hash: 512MB per engine

GUI: Since 19/09/11: Cutechess-cli (GUI ends game, when a 5-piece endgame is on the board), before: LittleBlitzerGUI (draw at 170 moves, resign at -700cp)

Tablebases: None for engines, 5 Syzygy for cutechess-cli

Openings: HERT_500 testset (by Thomas Zipproth) (download the file at the "Download & Links"-section or here)

Ponder, Large Memory Pages & learning: Off

Thinking time: 180''+1000ms (= 3'+1'') per game/engine (average game-duration: around  7.5 minutes). One 5000 games-testrun takes about 7 days.The version-numbers of the Stockfish engines are the date of the latest patch, which was included in the Stockfish sourcecode, not the release-date of the engine-file, written backwards (year,month,day))(example: 170526 = May, 26, 2017). Since July, 2018 I use the abrok-compiles of Stockfish again (http://abrok.eu/stockfish), because they are now much faster than before - now only 1.3% slower than BrainFish-compiles. So, there is no reason anymore to not use these "official" development-compiles.

Download BrainFish (and the Cerebellum-Libraries): here

Each Stockfish-version plays 1000 games versus Komodo 13.1, Houdini 6, Fire 7.1, Xiphos 0.5.6, Ethereal 11.53. All engines are running with default-settings.

To avoid distortions in the Ordo Elo-calculation, from now, only 2x Stockfish (latest official release + the latest version) and 1x asmFish and 1x Brainfish are stored in the gamebase (all older engine-versions games will be deleted, every time, when a new version was tested). Stockfish, asmFish and BrainFish older Elo-results can still be seen in the Elo-diagrams below. BrainFish plays always with the latest Cerebellum-Libraries of course, because otherwise BrainFish = Stockfish.

Latest update: 2020/01/15: Stockfish 11 (-7 Elo to Stockfish 200107)

(Ordo-calculation fixed to Stockfish 11 = 3554 Elo)

See the individual statistics of engine-results here

See the ORDO-rating of the archive-gamebase since 2020 here

Download the current gamebase here

Download the archive-gamebase since 2020 here

     Program                      Elo    +    -   Games   Score   Av.Op.  Draws

   1 BrainFish-2 191007 bmi2    : 3606    9    9  5000    82.1 %   3330   34.2 %
   2 Stockfish 200107 bmi2      : 3561    8    8  5000    77.5 %   3338   40.6 %
   3 Stockfish 11 200118        : 3554    8    8  5000    76.8 %   3338   42.7 %
   4 Stockfish 10 181129        : 3508    4    4 18000    78.0 %   3276   38.3 %
   5 Stockfish 9 180201         : 3457    8    8  5000    74.9 %   3254   41.7 %
   6 Houdini 6 pext             : 3427    3    3 24000    63.9 %   3316   47.8 %
   7 Komodo 13.1 bmi2           : 3408    5    5 14000    57.2 %   3353   48.8 %
   8 Komodo 13.01 bmi2          : 3403    7    7  5000    57.7 %   3345   53.2 %
   9 Komodo 12.3 bmi2           : 3393    7    7  7000    62.7 %   3295   49.4 %
  10 Komodo 13.2 MCTS           : 3314    7    7  6000    44.1 %   3359   54.2 %
  11 Ethereal 11.75 pext        : 3292    6    6  8000    35.1 %   3411   50.0 %
  12 Xiphos 0.6 bmi2            : 3283    7    7  7000    31.3 %   3433   47.8 %
  13 Fire 7.1 popc              : 3278    3    3 24000    44.9 %   3323   51.7 %
  14 Xiphos 0.5.6 bmi2          : 3271    6    6  9000    38.4 %   3366   51.7 %
  15 Ethereal 11.53 pext        : 3264    5    5 11000    40.2 %   3345   50.6 %
  16 Komodo 12.3 MCTS           : 3257    6    6  7000    42.7 %   3315   46.3 %
  17 Ethereal 11.25 pext        : 3250    6    6  7000    36.8 %   3352   50.1 %
  18 Laser 1.7 bmi2             : 3199    7    7  6000    30.8 %   3352   45.8 %
  19 rofChade 2.2 bmi2          : 3199    8    8  5000    27.8 %   3377   43.9 %
  20 Fritz 17                   : 3195    7    7  6000    29.4 %   3359   44.2 %
  21 Fizbo 2 bmi2               : 3195    8    8  5000    36.0 %   3307   39.0 %
  22 Shredder 13 x64            : 3191    7    7  6000    31.9 %   3341   42.6 %
  23 Defenchess 2.2 popc        : 3188    8    8  5000    26.6 %   3377   41.8 %
  24 Booot 6.3.1 popc           : 3181    8    8  5000    34.0 %   3310   44.1 %
  25 Andscacs 0.95 popc         : 3149    8    8  5000    23.1 %   3373   35.4 %

The version-numbers (180622 for example) of the engines are the date of the latest patch, which was included in the Stockfish sourcecode, not the release-date of the engine-file. Especially the asmFish-engines are often released much later!!

Below you find a diagram of the progress of Stockfish in my tests since the end of 2018. 

And below that diagram, the older diagrams.

You can save the diagrams (as a JPG-picture (in originial size)) on your PC with mouseclick (right button) and then choose "save image"...

The Elo-ratings of older Stockfish dev-versions in the Ordo-calculation can be a little different to the Elo-"dots" in the diagram, because the results/games of new Stockfish dev-versions - when getting part of the Ordo-calculation - can change the Elo-ratings of the opponent engines and that can change the Elo-ratings of older Stockfish dev-versions (in the Ordo-calculation / ratinglist, but not in the diagram, where all Elo-"dots" are the rating of one Stockfish dev-version at the moment, when the testrun of that Stockfish dev-version was finished).