Stefan Pohl Computer Chess

Latest Website-News (2026/07/05): Testrun of Pawnocchio 2.0 finished: +48 Celo to Pawnocchio 260527: Incredible progress in just 2 months, now only -25 Celo weaker than Torch 4d...

Testrun of Raphael 4.2.0 finished: +7 Celo to Raphael 260609

 

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). 

 

Next testruns: Stormphrax 8.0 and after that (perhaps strong enough?) trying Cinder 0.5.0, PZChessBot v7.1, Heimdall 1.5.

Because there are so many releases at the same time right now, the next Stockfish dev testrun will start, when all these testruns are finished.

 

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: 2026/07/05: Pawnocchio 2.0 (+48 Celo to Pawnocchio 260527) and Raphael 4.2.0 (+7 Celo to Raphael 260609)

 

(Ordo-calculation fixed to Stockfish 18 = 3873 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 260614: 3891 Celo, latest full release (Stockfish 18) has 3873 Celo)

     Program                     Celo    +    - Games    Score   Av.Op. Draws

   1 Stockfish 260614          : 3891    4    4 15000    67.3%   3762   49.1%
   2 Stockfish 18 260131       : 3873    4    4 15000    64.8%   3764   48.9%
   3 Reckless 260502 a512      : 3838    4    4 15000    59.9%   3766   48.9%
   4 Torch 4d a512             : 3820    4    4 15000    57.3%   3767   50.0%
   5 PlentyChess 8.0.0         : 3816    4    4 15000    56.8%   3767   48.0%
   6 Pawnocchio 2.0 a512       : 3795    4    4 15000    53.6%   3769   49.2%
   7 Obsidian 250706 a512      : 3775    4    4 15000    50.8%   3770   48.9%
   8 Viridithas 20.0 a512      : 3774    4    4 15000    50.5%   3770   50.2%
   9 Alexandria 9.0 a512       : 3749    4    4 15000    46.9%   3772   49.2%
  10 Integral 250805 a512      : 3745    4    4 15000    46.2%   3772   49.8%
  11 Berserk 14.0 a512         : 3731    4    4 15000    44.3%   3773   49.2%
  12 Quanticade 260617 a512    : 3716    4    4 15000    42.1%   3774   49.1%
  13 Raphael 4.2.0 a512        : 3713    4    4 15000    41.7%   3774   48.9%
  14 Halogen 16.5 a512         : 3699    4    4 15000    39.6%   3775   48.5%
  15 Caissa 1.25 a512          : 3696    4    4 15000    39.2%   3775   49.5%
  16 Tarnished 6.0 a512        : 3695    4    4 15000    39.0%   3776   49.1%

 

Games        : 120000 (finished)
White Wins   : 60234 (50.2%)
Black Wins   : 781    (0.7%)
Draws        : 58985 (49.2%)

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 260614          :    3928     10    7500  5265  2069   166  84.0     100
   2 Stockfish 18 260131       :    3873   ----    7500  4734  2411   355  79.2     100
   3 Reckless 260502 a512      :    3777     10    7500  3672  3035   793  69.2     100
   4 Torch 4d a512             :    3736     10    7500  3200  3258  1042  64.4      99
   5 PlentyChess 8.0.0         :    3726     10    7500  3096  3276  1128  63.1     100
   6 Pawnocchio 2.0 a512       :    3678      9    7500  2622  3314  1564  57.1     100
   7 Obsidian 250706 a512      :    3636     10    7500  2161  3405  1934  51.5      84
   8 Viridithas 20.0 a512      :    3631      9    7500  2111  3419  1970  50.9     100
   9 Alexandria 9.0 a512       :    3578      9    7500  1587  3414  2499  43.9      99
  10 Integral 250805 a512      :    3568     10    7500  1485  3422  2593  42.6     100
  11 Berserk 14.0 a512         :    3537      9    7500  1267  3257  2976  38.6     100
  12 Quanticade 260617 a512    :    3505     10    7500   990  3201  3309  34.5      99
  13 Raphael 4.2.0 a512        :    3494     10    7500   913  3162  3425  33.3     100
  14 Halogen 16.5 a512         :    3470     10    7500   745  3053  3702  30.3     100
  15 Caissa 1.25 a512          :    3457     10    7500   691  2935  3874  28.8      63
  16 Tarnished 6.0 a512        :    3455     10    7500   696  2899  3905  28.6     ---

 

------------------------------------------------------------------- 
--- Number of all Gamepairs          : 60000 
--- Number of drawn Gamepairs overall: 24765 (= 41.27%) 
--- Number of 1:1 drawn Gamepairs    : 12619 (= 21.03%) 
--- Number of 2-draws drawn Gamepairs: 12146 (= 20.24%) 
------------------------------------------------------------------- 

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

Icarus 1.1, 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"...