Stormphrax, UCI chess engine implemented NNUE
Thanks to the author: Conor Anstey, Rating Chess Engines Diary CEDR=3662
v.6.0.0:
Stormphrax 6.0.0 is significantly stronger than 5.0.0 - a longstanding bug in datagen has been fixed, and the entire dataset has been regenerated, resulting in a massively improved (and much larger) network. The regular search tweaks have also happened, along with improvements to time management, and Stormphrax now supports go searchmoves.
Many thanks again to everyone who has contributed CPU time to the OpenBench instance that Stormphrax is tested on - now too many to list, but as always especially @JouanDeag for contributing hosting together with @noobpwnftw for a constant presence of workers :3
Results against Stormphrax 5.0.0
STC:
Elo | 87.03 +- 7.22 (95%)
Conf | 8.0+0.08s Threads=1 Hash=32MB
Games | N: 3024 W: 938 L: 196 D: 1890
Penta | [4, 105, 691, 569, 143]
LTC:
Elo | 77.25 +- 8.43 (95%)
Conf | 60.0+0.60s Threads=1 Hash=128MB
Games | N: 1504 W: 373 L: 44 D: 1087
Penta | [0, 24, 414, 275, 39]
FRC STC:
Elo | 115.83 +- 8.10 (95%)
Conf | 8.0+0.08s Threads=1 Hash=32MB
Games | N: 3004 W: 1209 L: 243 D: 1552
Penta | [2, 108, 533, 640, 219]
DFRC STC:
Elo | 136.13 +- 8.64 (95%)
Conf | 8.0+0.08s Threads=1 Hash=32MB
Games | N: 3022 W: 1409 L: 282 D: 1331
Penta | [6, 100, 439, 693, 273]
Elo testing done with OpenBench's 8moves_v3.epd book for standard chess, 3moves_FRC.epd for FRC, and DFRC.epd for DFRC.
Results are obtained from self-play and will likely not scale perfectly to games against other opponents :)
Stormphrax 6.0.0 vs other engines:
| Caissa 1.21 | 51.5/113 | -10 | 113 Games |
| Starzix 6.0 | 59/109 | +9 | 109 Games |
| Alexandria 7.1.0 | 38.5/94 | -17 | 94 Games |
| Viridithas 15.0.0 | 39/83 | -5 | 83 Games |
| RubiChess 20240817 | 40/82 | -2 | 82 Games |
| Dragon 3.3 | 26/70 | -18 | 70 Games |
| Motor 0.8.0 JA | 33/68 | -2 | 68 Games |
| Velvet 8.1.1 JA | 29.5/58 | +1 | 58 Games |
| Chess-System-Tal-2 | 29.5/50 | +9 | 50 Games |
| Berserk 13 | 21.5/48 | -5 | 48 Games |
| PlentyChess 3.0.0 | 20/48 | -8 | 48 Games |
| Velvet 8.1.0 | 26/46 | +6 | 46 Games |
| Devre 6.39 JA | 23/44 | +2 | 44 Games |
| Obsidian 14.01 SE | 18.5/44 | -7 | 44 Games |
| Integral v5 | 18.5/41 | -4 | 41 Games |
| Clover 8.1 JA | 20/38 | +2 | 38 Games |
| Motor 0.8.0 | 17.5/35 | +0 | 35 Games |
| Lizard 11.1 | 17/35 | -1 | 35 Games |
| Lizard 11.2 | 15.5/34 | -3 | 34 Games |
| Uralochka 3.42 dev7 | 16/32 | +0 | 32 Games |
| Integral v6 | 16/32 | +0 | 32 Games |
| Berserk 20241119 | 13/32 | -6 | 32 Games |
| Revenge 4.0 | 15.5/26 | +5 | 26 Games |
| Ethereal 14.25 | 13.5/26 | +1 | 26 Games |
| Stockfish 17 | 10/26 | -6 | 26 Games |
| Titan 1.1.0 | 14/25 | +3 | 25 Games |
| Spaghet BologNNese 1.0 | 13/24 | +2 | 24 Games |
| Viridithas 16.0.0 | 12/24 | +0 | 24 Games |
| PlentyChess 3.0.1 | 10.5/24 | -3 | 24 Games |
| Yuliana 5.50 | 8/22 | -6 | 22 Games |
| Igel 3.6.0 JA | 12.5/20 | +5 | 20 Games |
Windows/Linux & Android profiled builds - Jim Ablett compiles:
Comments
Post a Comment