AI Rating List

Discussion of computer chess matches and engine tournaments.

Moderator: Ras

User avatar
Rebel
Posts: 7425
Joined: Thu Aug 18, 2011 12:04 pm
Full name: Ed Schröder

AI Rating List

Post by Rebel »

First Provisional Rating List after just 5 matches.

Code: Select all

   # PLAYER                   :  RATING  ERROR  POINTS  PLAYED   (%)  CFS(%)     W     D     L  D(%)
   1 Caissa-1.23              :  3670.0    6.6  1935.5    3000    65     100  1088  1695   217    57
   2 Rebel-16.4               :  3611.1    5.2  4859.5    9000    54      99  1979  5761  1260    64
   3 Clover.8.0.2             :  3603.1    6.2  4645.5    9000    52      60  1748  5795  1457    64
   4 Chess-System-Tal-2.00    :  3602.2    3.7  4752.0    9000    53     100  1910  5684  1406    63
   5 Fritz-20                 :  3591.3    5.5  4502.0    9000    50      99  1587  5830  1583    65
   6 Rebel-16.3               :  3575.0   13.4  1825.5    3000    61      58   907  1837   256    61
   7 Titan-1.1.0              :  3573.7    8.4  1820.5    3000    61      99   924  1793   283    60
   8 Komodo-Dragon-2.5        :  3560.0    3.8  3876.0    9000    43     100  1135  5482  2383    61
   9 Stockfish-12             :  3550.0    3.8  4157.0    9000    46     100  1508  5298  2194    59
  10 Arasan-25.2              :  3493.7    8.5  3626.5    9000    40     100   966  5321  2713    59
  11 Jet-1.2                  :  3200.0    3.8  1552.5    3000    52     100   690  1725   585    58
  12 Dog-4.10.1               :  3187.1    7.3  5061.0    9000    56     100  3183  3756  2061    42
  13 Simbelmyne-v1.10.0       :  3159.3    8.4  1387.0    3000    46     100   897   980  1123    33
  14 Pingu-5.0.0              :  3059.5    5.8   999.5    3000    33     ---   474  1051  1475    35
Engines are not well connected at the moment, will take some time.

Next matches of engines with book :
. Marvin-6.3.0
. Texel-1.11
. Texel-1.12
. Revenge-3.0
. Revenge-4.0

Details at : https://prodeo.actieforum.com/t1818-ai- ... st-results
90% of coding is debugging, the other 10% is writing bugs.
User avatar
Rebel
Posts: 7425
Joined: Thu Aug 18, 2011 12:04 pm
Full name: Ed Schröder

Re: AI Rating List

Post by Rebel »

Current rating list :



Click on the button right above for full page view.

From now on engines with books has to play each other, to begin with Komodo-Dragon-2.5

. Komodo-Dragon-2.5 - Arasan 25.2
. Komodo-Dragon-2.5 - Chess System Tal 2.0
. Komodo-Dragon-2.5 - Rebel 16.4
. Komodo-Dragon-2.5 - Revenge 4.0

Details at : https://prodeo.actieforum.com/t1818-ai- ... st-results
90% of coding is debugging, the other 10% is writing bugs.
User avatar
Rebel
Posts: 7425
Joined: Thu Aug 18, 2011 12:04 pm
Full name: Ed Schröder

Re: AI Rating List

Post by Rebel »

AI (all inclusive) Rating List 2025-11-06

https://rebel7775.wixsite.com/rebel/ai-rating-list

Observations

. Both Chess System Tal and Rebel gained about 50 elo points.
. Unfortunately Arasan does not handle the 40/10 time control very well, plays too fast.

Code: Select all

Engine                 Depth       Time   Games     Moves  Average Forfeit 
Stockfish-12           22.36   15:37:20    3000    217874    0.26     0    
Rebel-16.3             17.83   14:07:30    3000    210642    0.24     0    
Titan-1.1.0            24.32   13:57:58    3000    219235    0.23     0    
Arasan-25.2            20.14   34:42:57    9000    649678    0.19     0    
0.19 is wasting about 25% of the available time.

Next, the top-10 engines (from a compromise of ccrl and SPCC) will play each other, each 27.000 games.

Currently, as first one, Stockfish 17.1 is running, a screenshot of the match in progress.

Code: Select all

No. Name              Win Draw Loss Unf.  Score Games       %
-------------------------------------------------------------
  1 Stockfish-17.1   +5258 =11141 -1095   *0 10828.5 17494   61.9%
  2 Reckless-0.8     +175 =1316 -453   *0  833.0  1944   42.8%
  3 PlentyChess-7.0  +149 =1357 -438   *0  827.5  1944   42.6%
  4 Obsidian-16      +160 =1301 -483   *0  810.5  1944   41.7%
  5 Stockfish-15     +141 =1313 -490   *0  797.5  1944   41.0%
  6 Alexandria-8.1.2 +116 =1277 -551   *0  754.5  1944   38.8%
  7 Clover-9.1        +94 =1181 -667   *0  684.5  1942   35.2%
  8 Caissa-1.23       +94 =1177 -673   *0  682.5  1944   35.1%
  9 Berserk-13        +84 =1119 -741   *0  643.5  1944   33.1%
 10 Integral-v7       +82 =1100 -762   *0  632.0  1944   32.5%

Total Games:   17494
White Wins:     5389 (30.8%)
Black Wins:      964 (5.5%)
Draws:         11141 (63.7%)
90% of coding is debugging, the other 10% is writing bugs.
User avatar
Rebel
Posts: 7425
Joined: Thu Aug 18, 2011 12:04 pm
Full name: Ed Schröder

Re: AI Rating List

Post by Rebel »

Download games :

Download opening set :

The opening set contains 1500 openings, a mixture of :
1. Jeroen Noomen openings.
2. UHO openings.
3. Frank Quisinsky FEOBOS openings.
90% of coding is debugging, the other 10% is writing bugs.
User avatar
Rebel
Posts: 7425
Joined: Thu Aug 18, 2011 12:04 pm
Full name: Ed Schröder

Re: AI Rating List

Post by Rebel »

First STC (Short Time Control) Rating List

https://rebel7775.wixsite.com/rebel/stc-rating-list

Download games.

Engines with opening books and/or learning facilities are also welcome on the AI-Rating-List.

Next, Leela with BT4-spsa-1740 net, 3000 games.

All match results can be followed on : https://prodeo.actieforum.com/t1818-ai- ... st-results
90% of coding is debugging, the other 10% is writing bugs.
User avatar
Rebel
Posts: 7425
Joined: Thu Aug 18, 2011 12:04 pm
Full name: Ed Schröder

Re: AI Rating List

Post by Rebel »

STC Rating List 2025-11-12

New entry, Leela 0.32 with BT3 net

Code: Select all

No. Name            Win Draw Loss Unf.  Score Games       %
-----------------------------------------------------------
 1 Leela-0.32-BT3 +520 =2126 -354   *0 1583.0  3000   52.8%
 2 Stockfish-17.1 +354 =2126 -520   *0 1417.0  3000   47.2%

Total Games:    3000
White Wins:      772 (25.7%)
Black Wins:      102 (3.4%)
Draws:          2126 (70.9%)
https://rebel7775.wixsite.com/rebel/stc-rating-list

We have a new number one.
90% of coding is debugging, the other 10% is writing bugs.