Seer - UCI chess engine Author: Connor McMonigle
Rating CEDR=3333
v.2.6.0
This release is estimated to add 100 Elo in self play under UHO testing conditions. This improvement in strength can be attributed to both a number of further optimizations to the network inference implementation and an improved network trained by way of several additional self play training iterations in roughly equal measure. As with the prior release, the embedded network was trained solely on data originating from Seer produced by way of many self play iterations starting with the initial "retrograde learning" network bundled with v2.0.0-v2.1.0 as a base.
Individual statistics: Seer 2.5 - 90 games (+ 12,= 52,- 26), 42.2 %
Raubfisch X48e : 4 (+ 0,= 3,- 1), 37.5 %
SugaR AI 2.50 : 4 (+ 0,= 2,- 2), 25.0 %
Dragon 2.6.1 : 4 (+ 0,= 1,- 3), 12.5 %
Koivisto 8.6 : 2 (+ 0,= 1,- 1), 25.0 %
Stockfish 15 : 4 (+ 0,= 2,- 2), 25.0 %
Fisherov 0.98k : 2 (+ 0,= 1,- 1), 25.0 %
BrainLearn 17 : 2 (+ 0,= 0,- 2), 0.0 %
EMAN 7.93 : 2 (+ 0,= 1,- 1), 25.0 %
ShashChess 22 : 2 (+ 0,= 1,- 1), 25.0 %
Lc0 0.29.0-rc0 : 4 (+ 0,= 2,- 2), 25.0 %
Weiawaga 5.0.0 : 2 (+ 2,= 0,- 0), 100.0 %
Wasp 5.50 : 2 (+ 1,= 1,- 0), 75.0 %
Rebel 15 : 2 (+ 1,= 1,- 0), 75.0 %
Komodo 13.02 : 2 (+ 0,= 2,- 0), 50.0 %
Arasan 23.3 : 6 (+ 2,= 3,- 1), 58.3 %
Avalanche 0.3a : 2 (+ 2,= 0,- 0), 100.0 %
Swordfish 15.1 : 2 (+ 0,= 1,- 1), 25.0 %
Rebel 14.2 : 2 (+ 1,= 1,- 0), 75.0 %
Blue Marlin 15.1 : 2 (+ 0,= 1,- 1), 25.0 %
Berserk 9-dev : 6 (+ 0,= 6,- 0), 50.0 %
Fisherov chess monk 1.1 : 4 (+ 0,= 1,- 3), 12.5 %
Koivisto 8.9 : 4 (+ 0,= 1,- 3), 12.5 %
Igel 3.1.0 : 4 (+ 1,= 3,- 0), 62.5 %
Ethereal 12.75 : 4 (+ 1,= 3,- 0), 62.5 %
Slowchess 2.9 : 4 (+ 0,= 3,- 1), 37.5 %
Minic 3.22 : 4 (+ 0,= 4,- 0), 50.0 %
Pedone 3.1 : 4 (+ 0,= 4,- 0), 50.0 %
Fire 8 NN.MC.3 : 4 (+ 1,= 3,- 0), 62.5 %
Comments
Post a Comment