Stefan Pohl Computer Chess

Latest Website-News (2026/05/09): Testrun of Reckless 260502 finished: -2 Celo to Reckless 260401.

Testrun of Stockfish 260504 finished: +5 Celo to Stockfish 260426.

(Stockfish now provides just one windows binary (windows universal). This new binary is pretty weird: It is running much faster on my testing-machines, when only 1 instance of Stockfish is calculating. But somewhat slower, if 10 or more instances are running. Additionally, the new dev has a very different search: To reach depth 30 in console-mode, Stockfish 18 evaluates more than 8 MegaNodes, the new dev evaluates only 5.1 MegaNodes - a  huge difference!)

 

Ceres Binary here, Ceres Installation Guide here, Ceres Nets here

 

Dont forget to take a look at my EAS-Ratinglist (The world's first engine-ratinglist not measuring strength of engines, but engines's style of play).

 

 

Stay tuned.

UHO-Top15 Engines Ratinglist (+ regular testing of Stockfish Dev-versions)

 

 

Playing conditions:

 

The 15 strongest engines and the latest StockfishDev version are playing 1000 games vs. each opponent: 15000 games per engine are played. A RoundRobin Tournament with 120000 games overall !

 

Hardware: AMD Ryzen 9 7945HX 16-core (32 threads) notebook with 64GB RAM (since 2024/10/08), 2 identical machines. Windows 11 64bit

Speed: (singlethread, TurboBoost-mode switched off, chess starting position) Stockfish 17: 600 kn/s (when 26 games are running simultaneously)

Hash: 512MB per engine

GUI: Cutechess-cli (GUI ends game, when a 5-piece endgame is on the board, all other games are played until mate or draw by chess-rules (3fold, 50-moves, stalemate))

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

Openings: My UHO_2022_6mvs_+120_+129 openings are used (first 500 lines) (UHO 2022 openings are part of my Anti Draw Openings download-package). 

Ponder, Large Memory Pages & learning: Off

Thinking time: 3min+1sec per game/engine. One testrun (15000 games) takes around 80-85 hours (=3.5 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: 200807 = August, 7, 2020). The used SF compile is the AVX512-compile, which is the fastest on my AMD Ryzen CPU. SF binaries are taken from https://github.com/official-stockfish/Stockfish/releases (except the official SF-release versions, which are taken form the official Stockfish website).

Read the explanation, why I chose my testing-conditions, like I did here

 

Latest update: 2026/05/09: Stockfish 260504 (+5 Celo to Stockfish 260426) & Reckless 260502 (-2 Celo to Reckless 260401)

 

(Ordo-calculation fixed to Stockfish 17.1 = 3854 Celo. Mention, there are no human or "realistic" Elo numbers in enginechess, since the engines reached superhuman strength, so I decided to use "Celo" (=Computer Elo) instead)

 

See the tournament cross-table as a picture (done by Fritz20 GUI) here

See the individual statistics of engine-results here

See the Engines Aggressiveness Score Ratinglist here

See the Gamepair-statistics of the UHO-Top15 Ratinglist here

See the full UHO Ratinglist (with EAS-Ratinglist): here

Download the current gamebase, the full-list gamebase and the archive here

See the old SPCC-Ratinglist (full list) and EAS-List from 2020 until 2023/08/28: here

 

(best Stockfish Celo: Stockfish 260504: 3888 Celo, latest full release (Stockfish 18) has 3873 Celo)

     Program                    Celo    +    - Games    Score   Av.Op. Draws

   1 Stockfish 260504         : 3888    4    4 15000    68.3%   3750   48.6%
   2 Stockfish 18 260131      : 3873    4    4 15000    66.3%   3751   49.1%
   3 Reckless 260502 a512     : 3841    4    4 15000    62.0%   3753   48.7%
   4 Torch 4d a512            : 3828    4    4 15000    60.1%   3754   49.9%
   5 PlentyChess 7.0.37       : 3803    4    4 15000    56.5%   3756   48.5%
   6 Obsidian 250706 a512     : 3785    4    4 15000    54.0%   3757   49.9%
   7 Alexandria 9.0 a512      : 3759    4    4 15000    50.1%   3758   49.0%
   8 Integral 250805 a512     : 3754    4    4 15000    49.4%   3759   50.5%
   9 Berserk 250606 a512      : 3740    4    4 15000    47.3%   3760   50.1%
  10 Viridithas 19.0 a512     : 3724    4    4 15000    45.1%   3761   49.4%
  11 Caissa 1.25 a512         : 3705    4    4 15000    42.3%   3762   49.9%
  12 Clover 9.1 a512          : 3693    4    4 15000    40.5%   3763   50.7%
  13 KomodoDragon 3.3 avx2    : 3692    4    4 15000    40.4%   3763   49.4%
  14 Halogen 16.0 a512        : 3688    4    4 15000    39.9%   3763   49.2%
  15 Tarnished 5.0 a512       : 3684    4    4 15000    39.3%   3763   50.0%
  16 Pawnocchio 1.9 a512      : 3678    4    4 15000    38.5%   3764   50.3%

 

Games        : 120000 (finished)
White Wins   : 59484 (49.6%)
Black Wins   : 1029   (0.9%)
Draws        : 59487 (49.6%)

Below the UHO-Top15 Ratinglist, recalculated with Gamepairs (using my Gamepair Rescoring Tool V1.5), Ordo-calculation fixed to Stockfish 17.1 = 3854 Celo, realizing Vondele's (Stockfish Maintainer) idea: "Thinking uniquely in game pairs makes sense with the biased openings used these days. While pentanomial makes  sense it is a bit complicated so we could simplify and score game pairs only (not games) as W-L-D (a traditional score of 2-0, or 1.5-0.5 is just a W)."

Mention, the errorbar here is +/-5.5 Celo (=11 overall Error), compared to +/-4 Celo in the classical ratinglist above (=1.375 wider errorbar here). But the result-gaps are (at least) 2x up to 2.5x bigger (in average), calculating with gamepairs. So, the statistical reliability of the results is much better in the Gamepair-ratinglist, compared to the classical ratinglist...

CFS means Chance For Superiority in percent, calculated by ORDO.

See the head-to-head gamepair-statistics of the UHO-Top15 Ratinglist here

   # PLAYER                   :    Celo  Error   Pairs     W     D     L   (%)  CFS(%)
   1 Stockfish 260504         :    3923     11    7500  5533  1810   157  85.8     100
   2 Stockfish 18 260131      :    3873   ----    7500  5091  2109   300  81.9     100
   3 Reckless 260502 a512     :    3780     10    7500  4173  2656   671  73.3     100
   4 Torch 4d a512            :    3746     10    7500  3814  2840   846  69.8     100
   5 PlentyChess 7.0.37       :    3680     10    7500  3138  3068  1294  62.3     100
   6 Obsidian 250706 a512     :    3643     10    7500  2768  3137  1595  57.8     100
   7 Alexandria 9.0 a512      :    3581     10    7500  2123  3285  2092  50.2      99
   8 Integral 250805 a512     :    3571     11    7500  2059  3216  2225  48.9     100
   9 Berserk 250606 a512      :    3537     10    7500  1718  3243  2539  44.5     100
  10 Viridithas 19.0 a512     :    3504     10    7500  1417  3223  2860  40.4     100
  11 Caissa 1.25 a512         :    3459     10    7500  1063  3106  3331  34.9     100
  12 KomodoDragon 3.3 avx2    :    3432     10    7500   913  2933  3654  31.7      77
  13 Clover 9.1 a512          :    3429     10    7500   879  2944  3677  31.3      84
  14 Halogen 16.0 a512        :    3425     10    7500   871  2882  3747  30.8     100
  15 Tarnished 5.0 a512       :    3409     11    7500   750  2858  3892  29.1     100
  16 Pawnocchio 1.9 a512      :    3392     10    7500   701  2668  4131  27.1     ---

 

------------------------------------------------------------------- 
--- Number of all Gamepairs          : 60000 
--- Number of drawn Gamepairs overall: 22989 (= 38.31%) 
--- Number of 1:1 drawn Gamepairs    : 11374 (= 18.96%) 
--- Number of 2-draws drawn Gamepairs: 11615 (= 19.36%) 
------------------------------------------------------------------- 

Aborted testruns, because the tested engine was too weak to enter the UHO-Top15 Ratinglist:

Raphael 4.0, Heimdall 1.4.1, Renegade 1.2.0, Igel 3.6.0, Minic 3.41, Booot 7.3, Black Marlin 9.0, Equisetum 1.0, Clarity 8

Below you find a diagram of the progress of Stockfish in my tests since February 2026. 

And below that diagram, the older diagrams.

 

You can save the diagrams on your PC with mouseclick (right button) and then choose "save image"...