Chess324 Top 15 Round-Robin - several new engines added

[Modern Times]

This is my ongoing Top15 round-robin tournament of free engines only, using Stefah Pohl's unbalanced 3-move book Chess324_xxl_big_+090_+119.pgn played reversed sides per pairing.

Time Control: game in 2 minutes plus 1 second increment, 2'+1" on an Intel i7-12700 @ 85w
(runs at approx. 4.3Ghz with concurrency 8)
Each engine 512MB hash, 1 thread, ponder off

New versions replace previous versions, and new entrants cause the bottom engine to drop out. It is a revolving tournament rather than a ratings list, however I am showing Ordo Elo as well as the tournament scores because they are of value to some people.

Over the course of a week there were about 8 new engine releases, and I've now updated the tournament with all of them.

New engines in the Top 15: Integral V5 at 10th place, Starzix 6.0 at 11th place, Velvet 8.1.0 at 16th place
(Velvet 8.1.0 didn't quite make the Top 15, but at just 2 Elo lower than Uralochka it is within the margin of error so I left it there.

Updated versions:
Obsidian 14 - goes from 3rd place to 2nd place, but only just. Approx +42 Elo performance
Caissa 1.21 - stays at 4th place. Approx +16 Elo performance
Alexandria 7.1.0 - goes from 9th place to 5th place. Approx +50 Elo performance
Stormphrax 6.0.0 - goes from 15th place to 9th place. Approx +89 Elo performance
Motor 0.8.0 - dropped from 11th place to 13th place due to entry of new engines . Approx +27 Elo performance over 0.7.0

Engines that fell out:
Titan 1.1, Seer 2.8.0 and Clarity 7.2.0

Tournament results:

Code: Select all

 # PLAYER                  :  RATING  ERROR  POINTS  PLAYED   (%)     W     D     L  D(%)  OppAvg  CFS(%)
   1 Stockfish 17          :    3843      8  2159.5    3000  72.0  1374  1571    55  52.4    3674     100
   2 Obsidian 14           :    3762      8  1836.0    3000  61.2   974  1724   302  57.5    3680      66
   3 Berserk 13            :    3759      8  1826.0    3000  60.9   947  1758   295  58.6    3680     100
   4 Caissa 1.21           :    3733      8  1711.5    3000  57.0   812  1799   389  60.0    3682     100
   5 Alexandria 7.1.0      :    3708      7  1603.5    3000  53.5   694  1819   487  60.6    3683      64
   6 Clover 8.0            :    3706      8  1595.0    3000  53.2   687  1816   497  60.5    3683     100
   7 Viridithas 15.0.0     :    3685      7  1503.5    3000  50.1   607  1793   600  59.8    3685      91
   8 PlentyChess 2.1.0     :    3678      7  1471.0    3000  49.0   565  1812   623  60.4    3685      71
   9 Stormphrax 6.0.0      :    3675      7  1458.0    3000  48.6   599  1718   683  57.3    3685      87
  10 Integral V5           :    3669      7  1432.0    3000  47.7   577  1710   713  57.0    3686      75
  11 Starzix 6.0           :    3665      7  1416.5    3000  47.2   551  1731   718  57.7    3686     100
  12 RubiChess 20240817    :    3650      8  1349.5    3000  45.0   472  1755   773  58.5    3687     100
  13 Motor 0.8.0           :    3630      8  1262.0    3000  42.1   369  1786   845  59.5    3688     100
  14 Lizard 10.5           :    3608      8  1166.5    3000  38.9   320  1693   987  56.4    3690      99
  15 Uralochka 3.41        :    3594      8  1109.5    3000  37.0   292  1635  1073  54.5    3691      66
  16 Velvet 8.1.0          :    3592      8  1100.0    3000  36.7   302  1596  1102  53.2    3691     ---

Stats
Total: 24'000 games
played by 16 programs

White wins: 9'600 (40.0%)
Black wins: 542 (2.3%)
Draws: 13'858 (57.7%)
White score: 68.9%

Additional Tournament Info:
200 game pairings for decent number of games and a lower margin of error
CuteChess GUI, Concurrency 8
Adjudication: Syzygy 5-men only, there is no eval based adjudication.
Endgame tablebases: Syzygy 5-men.


Pgn available on request - send me a PM with your email address.

[Modern Times]

Lizard 11.1 has replaced Lizard 10.5 and jumped up a place to 13th. An approx + 40 Elo improvement.

I've temporarily reinstated the 16th to 18th placed engines because they are pretty much within error margins of the 15th placed engine.

Code: Select all

   # PLAYER                :  RATING  ERROR  POINTS  PLAYED   (%)     W     D     L  D(%)  OppAvg  CFS(%)
   1 Stockfish 17          :    3843      8  2465.5    3400  72.5  1589  1753    58  51.6    3669     100
   2 Obsidian 14           :    3763      7  2113.0    3400  62.1  1145  1936   319  56.9    3674      61
   3 Berserk 13            :    3762      7  2106.0    3400  61.9  1125  1962   313  57.7    3674     100
   4 Caissa 1.21           :    3736      7  1983.0    3400  58.3   977  2012   411  59.2    3675     100
   5 Alexandria 7.1.0      :    3712      7  1864.0    3400  54.8   848  2032   520  59.8    3677      62
   6 Clover 8.0            :    3710      7  1856.0    3400  54.6   841  2030   529  59.7    3677     100
   7 Viridithas 15.0.0     :    3687      7  1740.5    3400  51.2   722  2037   641  59.9    3678      84
   8 PlentyChess 2.1.0     :    3682      7  1715.5    3400  50.5   689  2053   658  60.4    3679      73
   9 Stormphrax 6.0.0      :    3678      7  1700.5    3400  50.0   708  1985   707  58.4    3679      88
  10 Integral V5           :    3673      7  1671.0    3400  49.1   705  1932   763  56.8    3679      92
  11 Starzix 6.0           :    3665      7  1635.0    3400  48.1   658  1954   788  57.5    3680      94
  12 RubiChess 20240817    :    3657      7  1594.0    3400  46.9   603  1982   815  58.3    3680      93
  13 Lizard 11.1           :    3649      7  1556.5    3400  45.8   594  1925   881  56.6    3681     100
  14 Motor 0.8.0           :    3633      7  1474.5    3400  43.4   482  1985   933  58.4    3682     100
  15 Uralochka 3.41        :    3596      7  1297.0    3400  38.1   373  1848  1179  54.4    3684      70
  16 Velvet 8.1.0          :    3593      7  1284.0    3400  37.8   376  1816  1208  53.4    3684      59
  17 Titan 1.1             :    3592      7  1278.0    3400  37.6   329  1898  1173  55.8    3684      68
  18 Seer 2.8.0            :    3589      7  1266.0    3400  37.2   338  1856  1206  54.6    3684     ---

[Modern Times]

New versions added:

Clover 8.1 - similar Elo to version 8.0, remains 6th
PlentyChess 3.0.1 - approx +23 Elo, move up a place to 7th
Integral V6 - approx +28 Elo, moves up two places to 8th

15th to 18th places are essentially 15th equal given the error margins, so the extra engines remain for now.

Code: Select all

   # PLAYER                :  RATING  ERROR  POINTS  PLAYED   (%)     W     D     L  D(%)  OppAvg  CFS(%)
   1 Stockfish 17          :    3843      8  2464.0    3400  72.5  1591  1746    63  51.4    3669     100
   2 Berserk 13            :    3761      7  2102.5    3400  61.8  1117  1971   312  58.0    3674      50
   3 Obsidian 14           :    3761      8  2102.5    3400  61.8  1150  1905   345  56.0    3674     100
   4 Caissa 1.21           :    3732      7  1963.5    3400  57.8   957  2013   430  59.2    3676     100
   5 Alexandria 7.1.0      :    3710      7  1856.5    3400  54.6   864  1985   551  58.4    3677      55
   6 Clover 8.1            :    3710      7  1853.5    3400  54.5   851  2005   544  59.0    3677      84
   7 PlentyChess 3.0.1     :    3705      7  1829.0    3400  53.8   868  1922   610  56.5    3677      77
   8 Integral V6           :    3701      8  1809.0    3400  53.2   842  1934   624  56.9    3678     100
   9 Viridithas 15.0.0     :    3683      7  1721.5    3400  50.6   703  2037   660  59.9    3679      95
  10 Stormphrax 6.0.0      :    3674      7  1679.0    3400  49.4   690  1978   732  58.2    3679     100
  11 Starzix 6.0           :    3658      7  1599.0    3400  47.0   637  1924   839  56.6    3680      80
  12 RubiChess 20240817    :    3654      7  1577.5    3400  46.4   593  1969   838  57.9    3680      96
  13 Lizard 11.1           :    3645      7  1532.5    3400  45.1   569  1927   904  56.7    3681     100
  14 Motor 0.8.0           :    3629      7  1458.0    3400  42.9   467  1982   951  58.3    3682     100
  15 Uralochka 3.41        :    3591      7  1274.0    3400  37.5   362  1824  1214  53.6    3684      58
  16 Velvet 8.1.0          :    3590      7  1269.0    3400  37.3   371  1796  1233  52.8    3684      59
  17 Titan 1.1             :    3589      7  1263.0    3400  37.1   319  1888  1193  55.5    3684      75
  18 Seer 2.8.0            :    3585      7  1246.0    3400  36.6   318  1856  1226  54.6    3684     ---

This may be the last update for this Top 15 round-robin - in a sea of tournaments and ratings lists these days there seems little point.
It is modelled on and is as rigorous as SPCC UHO Top15, but instead it uses unbalanced Chess324 openings (subset of DFRC) with a 3-move book by Stefan that really gives the engines a work-out with positions they are not familiar with. It differentiates itself in this respect.

Stats
Total: 30'600 games
played by 18 programs

White wins: 12'399 (40.5%)
Black wins: 870 (2.8%)
Draws: 17'331 (56.6%)
White score: 68.8%

Tournament Info:
Top15 round-robin tournament of free engines only, using Stefah Pohl's unbalanced 3-move book Chess324_xxl_big_+090_+119.pgn played reversed sides per pairing.
Time Control: game in 2 minutes plus 1 second increment, 2'+1"
Each engine 512MB hash, 1 thread, ponder off
200 game pairings for decent number of games and a lower margin of error
CuteChess GUI, Concurrency limited to number of actual cores
Adjudication: Syzygy 5-men only, there is no eval based adjudication.
Endgame tablebases: Syzygy 5-men.

Pgn available on request - send me a PM with your email address.

[Modern Times]

I've added the latest version of Revenge, which is now free. (version 4)

It falls into the group of 5 engines that I regard essentially as 15th equal because of the error margins. Table below (shows as 18th):

Code: Select all

   # PLAYER                :  RATING  ERROR  POINTS  PLAYED   (%)     W     D     L  D(%)  OppAvg  CFS(%)
   1 Stockfish 17          :    3843      8  2616.0    3600  72.7  1696  1840    64  51.1    3667     100
   2 Berserk 13            :    3763      7  2247.5    3600  62.4  1213  2069   318  57.5    3672      59
   3 Obsidian 14           :    3762      7  2241.5    3600  62.3  1234  2015   351  56.0    3672     100
   4 Caissa 1.21           :    3733      7  2097.0    3600  58.3  1034  2126   440  59.1    3673     100
   5 Alexandria 7.1.0      :    3712      6  1989.0    3600  55.3   943  2092   565  58.1    3675      55
   6 Clover 8.1            :    3712      7  1985.5    3600  55.2   932  2107   561  58.5    3675      78
   7 PlentyChess 3.0.1     :    3708      7  1964.5    3600  54.6   956  2017   627  56.0    3675      81
   8 Integral V6           :    3703      7  1940.5    3600  53.9   921  2039   640  56.6    3675     100
   9 Viridithas 15.0.0     :    3686      7  1851.0    3600  51.4   776  2150   674  59.7    3676      97
  10 Stormphrax 6.0.0      :    3677      6  1802.0    3600  50.1   756  2092   752  58.1    3677     100
  11 Starzix 6.0           :    3660      7  1716.5    3600  47.7   697  2039   864  56.6    3677      70
  12 RubiChess 20240817    :    3658      7  1703.0    3600  47.3   662  2082   856  57.8    3678      98
  13 Lizard 11.1           :    3647      7  1649.0    3600  45.8   628  2042   930  56.7    3678     100
  14 Motor 0.8.0           :    3634      7  1578.5    3600  43.8   528  2101   971  58.4    3679     100
  15 Uralochka 3.41        :    3594      7  1377.0    3600  38.3   411  1932  1257  53.7    3681      54
  16 Velvet 8.1.0          :    3593      7  1374.0    3600  38.2   419  1910  1271  53.1    3681      66
  17 Titan 1.1             :    3591      7  1363.5    3600  37.9   359  2009  1232  55.8    3681      61
  18 Revenge 4.0           :    3590      7  1356.0    3600  37.7   367  1978  1255  54.9    3681      62
  19 Seer 2.8.0            :    3588      7  1348.0    3600  37.4   359  1978  1263  54.9    3681     ---

[Modern Times]

This can now be viewed as a clean and simple web page here:

https://e4e6.com/324/

Thanks to Alex Brunetti for offering to do this, unprompted and unexpected. I'll keep this Top 15 round-robin going, it is very low effort and easily manageable.

One day one of those 15th-equal engines will make a 20+ Elo jump, and it will then be back to an actual Top 15.

[Modern Times]

Lizard 11.2 has replaced 11.1 and has jumped two places to 11th as a result of an approx +19 Elo improvement.

https://e4e6.com/324/


Next up is Horsie 1.0 from the same author, so it will replace Lizard if it is stronger, which it is expected to be.

[Modern Times]

Horsie 1.0 a further +20 Elo over Lizard 11.2, and replaces it. And steals 10th place from Stormphrax in the process.

Details here:

https://e4e6.com/324/

[Modern Times]

PlentyChess 4.0.1 replaces the previous version 3.0.1, jumping 3 places to 4th thanks to an approx. +48 Ordo Elo gain under these conditions.

https://e4e6.com/324/

[Modern Times]

Viridithas 16.0.0 replaces the previous version 15.0.0, jumping 1 place to 8th thanks to an approx. +15 Ordo Elo gain under these conditions.

https://e4e6.com/324/