Motor - UCI chess engine written in C++ 17 with (768 -> 32) x 2 -> 1 NNUE trained with
Currently requires Clang or GCC/MinGW compiler.
Author: Martin Novรกk, Ranking Chess Engines Diary CEDR=3673
We are happy that there is another version of this engine (the previous one was released in August). The previous version plays in the 1st League of the best CEDR chess engines and has a rating of 3673! It will be interesting to see how it fares in new tests - these will be available on our website soon.
Motor 0.9.0 - what changes in the new version?
Progression Tests:
Tests on UHO_Lichess_4852_v1.epd opening book:
Elo | 44.21 +- 3.68 (95%)
Conf | 10.0+0.10s Threads=1 Hash=32MB
Games | N: 10002 W: 3275 L: 2009 D: 4718
Penta | [18, 748, 2317, 1786, 132]
Elo | 50.45 +- 3.36 (95%)
Conf | 60.0+0.60s Threads=1 Hash=64MB
Games | N: 10000 W: 3187 L: 1745 D: 5068
Penta | [5, 599, 2381, 1979, 36]
Tests on 8moves_v3.epd opening book:
Elo | 21.01 +- 2.21 (95%)
Conf | 60.0+0.60s Threads=1 Hash=64MB
Games | N: 10000 W: 979 L: 375 D: 8646
Penta | [0, 232, 3958, 784, 26]
Changes
Most elo comes from correction history experiments
Motor 0.8.0 - vs other engines:
Caissa 1.21 | 16/36 | -4 | 36 Partii |
Alexandria 7.1.0 | 16/36 | -4 | 36 Partii |
Stormphrax 6.0.0 | 17.5/35 | +0 | 35 Partii |
Velvet 8.1.0 | 17.5/32 | +3 | 32 Partii |
Devre 6.39 JA | 13/25 | +1 | 25 Partii |
Starzix 6.0 | 11.5/23 | +0 | 23 Partii |
Viridithas 15.0.0 | 11/22 | +0 | 22 Partii |
RubiChess 20240817 | 10.5/22 | -1 | 22 Partii |
PlentyChess 3.0.0 | 10.5/20 | +1 | 20 Partii |
Berserk 13 | 9.5/20 | -1 | 20 Partii |
Obsidian 14.01 SE | 9.5/20 | -1 | 20 Partii |
Quanticade Electra 0.9 | 6.5/12 | +1 | 12 Partii |
Yuliana 5.50 | 6/12 | +0 | 12 Partii |
Stockfish 20241031 | 5.5/12 | -1 | 12 Partii |
SF-PRO 25.10.2024 | 5/12 | -2 | 12 Partii |
Stockfish 17 | 5/12 | -2 | 12 Partii |
Big Yuli GM 221024 | 5/12 | -2 | 12 Partii |
Arasan 25.0 | 5/10 | +0 | 10 Partii |
Velvet 8.1.1 | 4.5/10 | -1 | 10 Partii |
Igel 3.6.0 | 4.5/9 | +0 | 9 Partii |
Titan 1.1.0 | 4.5/9 | +0 | 9 Partii |
Comments
Post a Comment