AndrewGrant wrote: ↑Wed Jun 09, 2021 8:38 pm
Thanks. Its nice to see that Pure lists are back ( At least for FRC ) . That is what I always looked at back in the day. I'de love to see those for each list, and actually have that be the default. Its simply a greater quality list by removing sampling biases
Yes agreed. The pure list never went away on FRC, but there is some sort of bug which means we had to turn them off for the other two lists. I'm trying to fix that for 40/15. One big issue with them is the number of games. Yes you reduce potential bias from testing against multiple versions of the same family, but you increase significantly in some cases the statistical margins of error because of less games.
For example Cfish is Stockfish optimized in C running 30% faster. ShashChess 17.1 implements Alexander Shashin mathematical theory and it's very strong. Corchess is perfect for long time analysis. But also the free (derivatives, NOT clones or they would score exactly like SF) Fisherov 0.98, Zeus 11, Eman 7.22, AI 15.00, KillFish CTR Hybrid v1.4 are deprecated because they do not release source being based on Stockfish GPL 3.0 (exactly like commercials Fat fritz 2 and Houdini 6 before they have been forced to drop the source code in github by the Stockfish team and fans) ...but often they score better than latest Stockfish 13-dev. Not considering them brings to a very partial list of all top engines. [Source] https://chessengines.blogspot.com/p/rating-jcer.html ...For me the most complete and trusted computer rating list *not 3600 Elo but 3000-3200 ELO* for top engines.
Andrew is totally right about SF "Clones". Even I had an clone in 2019, and I don't know how to code at an intermediate level. All I did was a search in fishtest for promising yellow patches (or one green one red) and some copy paste from other engines and some changes in evaluations. It's not a hard job but it has no value whatsoever.
I consider CFish, some Experience engines like Eman and BrainLearn (for also having MCTS) valuable even through I know there isn't a huge elo gain in experience.
I know that is not the same than writing a top level engine from scratch, but having tested all of them, I think that others fine tuned SF derivates deserve my attention and I'll continue to test all their new updates with the most promising NNUEs . I only criticize programmers not sharing their source code to improve all other engines.
PS: Ethereal 13.00 ssse3 (official github build by Andrew) needs his own NNUE, differently from 12.75 NNUE.
Using SF test NNUEs with it has given bad results to me (a lot of loses) also if the evaluation values seem consistent.
I have stopped this engine test on Ryzen 3900x.
AlexChess wrote: ↑Fri Jun 11, 2021 8:53 pm
PS: Ethereal 13.00 ssse3 (official github build by Andrew) needs his own NNUE, differently from 12.75 NNUE.
Using SF test NNUEs with it has given bad results to me (a lot of loses) also if the evaluation values seem consistent.
I have stopped this engine test on Ryzen 3900x.
Yeah Networks are not compatible. My stuff has absolutely nothing to do with Stockfish.
Also, the github builds are not NNUE. Hence it being the "Standard" version.
#WeAreAllDraude #JusticeForDraude #RememberDraude #LeptirBigUltra "Those who can't do, clone instead" - Eduard ( A real life friend, not this forum's Eduard )
AlexChess wrote: ↑Fri Jun 11, 2021 8:53 pm
PS: Ethereal 13.00 ssse3 (official github build by Andrew) needs his own NNUE, differently from 12.75 NNUE.
Using SF test NNUEs with it has given bad results to me (a lot of loses) also if the evaluation values seem consistent.
I have stopped this engine test on Ryzen 3900x.
Yeah Networks are not compatible. My stuff has absolutely nothing to do with Stockfish.
Also, the github builds are not NNUE. Hence it being the "Standard" version.
So what is the estimated Elo of Ethereal 13.00 NNUE ? Is it about 700 Elo weaker than stockfish 12 NNUE or about equal to igel 3.0.5
Chessqueen wrote: ↑Fri Jun 11, 2021 9:27 pmSo what is the estimated Elo of Ethereal 13.00 NNUE ? Is it about 700 Elo weaker than stockfish 12 NNUE or about equal to igel 3.0.5
About 20 Elo stronger than Igel 3.0.5 and 115 Elo weaker than Stockfish 12.
Chessqueen wrote: ↑Fri Jun 11, 2021 9:27 pmSo what is the estimated Elo of Ethereal 13.00 NNUE ? Is it about 700 Elo weaker than stockfish 12 NNUE or about equal to igel 3.0.5
About 20 Elo stronger than Igel 3.0.5 and 115 Elo weaker than Stockfish 12.
Thanks you, So Ethereal 13.00 NNUE 3438 CCRL, compared to igel 3.0.5 3418 3418, therefore most people will download Igel 3.0.5 and test those two in 100 games.
AlexChess wrote: ↑Fri Jun 11, 2021 8:53 pm
PS: Ethereal 13.00 ssse3 (official github build by Andrew) needs his own NNUE, differently from 12.75 NNUE.
Using SF test NNUEs with it has given bad results to me (a lot of loses) also if the evaluation values seem consistent.
I have stopped this engine test on Ryzen 3900x.
Yeah Networks are not compatible. My stuff has absolutely nothing to do with Stockfish.
Also, the github builds are not NNUE. Hence it being the "Standard" version.
Ok I cant wait to have a 13.x build stable on MY system (or a Windows 10, mac M1 or Linux native ARM64 NEON build like Igel 3.0.5)
Meanwhile I will continue to test Ethereal 12.75 NNUE (Etherlito) that is very strong also on modern computers (nehalem) hardware
Chessqueen wrote: ↑Fri Jun 11, 2021 9:27 pmSo what is the estimated Elo of Ethereal 13.00 NNUE ? Is it about 700 Elo weaker than stockfish 12 NNUE or about equal to igel 3.0.5
About 20 Elo stronger than Igel 3.0.5 and 115 Elo weaker than Stockfish 12.
Thanks Graham! I confirm your ranking, but for me ALL the best engines are in the 3000-3400 ELO range on high-end hardware.
AlexChess wrote: ↑Sat Jun 12, 2021 7:18 am
Thanks Graham! I confirm your ranking, but for me ALL the best engines are in the 3000-3400 ELO range on high-end hardware.
Stockfish 13 will be 3600+ for them when they test that.
I just ignore the absolute Elo values and just look at the differences between engines. The absolute values are different on most of the lists depending on how the scale the lists and what ratings tool they use.
It was my reliable source 30 years ago to have the exact ranking of Mephisto, Kasparov, Fidelity and Novag dedicated chess computers... (only one hardware and 1 opening book for each) . Now Stockfish-dev and test NNUEs are updated every day and they are stuck at Stockfish 12.
You are right, the only way is to compare the absolute ranking, not the ELO until you fix an unique standard for both humans and computers.
Regards, AlexChess
Last edited by AlexChess on Sat Jun 12, 2021 8:41 am, edited 5 times in total.