Stockfish derivative engines and registration problem

Questions and comments related to CCRL testing study
Locked
WampServer
Posts: 5
Joined: Thu Mar 21, 2024 6:07 pm
Sign-up code: 10159

Stockfish derivative engines and registration problem

Post by WampServer »

Hi Gabor Szots,
Some time ago I wanted to register my engine as a derivative of Stockfish for CCRL. You said I may announce my engine at one of the public chess fora. Then a CCRL member might pick it up for testing. I thought this was an easy way to introduce a new engine but I realized it's not possible after contacting another CCRL member (Graham Banks). Graham said: "Looks like copy paste Stockfish with a single commit Remove skipEarlyPruning". CCRL has a policy of letting authors register engines for fair competition upon request. I would like to make such a request on my behalf. I am NOT accusing CCRL of anything and in fact wish it well. However, while it is frustrating, there is no way to tell why my engine is not registered. I should add that I am not a maintainer and only one of Lifish contributors.
Thank you.

Discussion: https://github.com/official-stockfish/S ... sions/5129
MyGithub: https://github.com/WampServer
Engine: https://github.com/EarlyEdition/Lifish-HalfKP
Ray
Posts: 22606
Joined: Sun Dec 18, 2005 6:33 pm
Sign-up code: 10159
Location: NZ

Re: Stockfish derivative engines and registration problem

Post by Ray »

WampServer wrote: Sat Mar 23, 2024 12:46 am CCRL has a policy of letting authors register engines for fair competition upon request. I would like to make such a request on my behalf. I am NOT accusing CCRL of anything and in fact wish it well. However, while it is frustrating, there is no way to tell why my engine is not registered. I should add that I am not a maintainer and only one of Lifish contributors.
Thank you.
That is not true. There is no registration process and never has been. Authors can tell CCRL and the entire world that they have an engine, that is it.

Individual testers choose what engines they want to run and in what order of priority, and there are several engines that have never been tested -there are just so many it is not possible to run them all.
WampServer
Posts: 5
Joined: Thu Mar 21, 2024 6:07 pm
Sign-up code: 10159

Re: Stockfish derivative engines and registration problem

Post by WampServer »

Individual testers choose what engines they want to run and in what order of priority, and there are several engines that have never been tested -there are just so many it is not possible to run them all.
Right... But I asked Graham, "Why can't my engine be registered?" Graham said: "Looks like copy paste Stockfish". Graham's answer must be provable. It is not right to ignore the efforts of others.

Unfortunately, my conversation has been locked in the Official_Stockfish.
https://github.com/official-stockfish/S ... sions/5129
I am waiting to hear convincing reasons for not registering my engine here.
User avatar
Graham Banks
Posts: 26919
Joined: Sun Dec 18, 2005 5:47 pm
Sign-up code: 0
Location: Auckland, NZ

Re: Stockfish derivative engines and registration problem

Post by Graham Banks »

WampServer wrote: Sat Mar 23, 2024 1:51 am
Individual testers choose what engines they want to run and in what order of priority, and there are several engines that have never been tested -there are just so many it is not possible to run them all.
Right... But I asked Graham, "Why can't my engine be registered?" Graham said: "Looks like copy paste Stockfish". Graham's answer must be provable. It is not right to ignore the efforts of others.

Unfortunately, my conversation has been locked in the Official_Stockfish.
https://github.com/official-stockfish/S ... sions/5129
I am waiting to hear convincing reasons for not registering my engine here.
1. We can choose which engines we want to test, just as CEGT and other rating list maintainers can.
2. In Discord engine programming channels, some top engine authors including a couple involved with Stockfish, have alluded to the fact that your engine is basically Stockfish, but with an originally trained net. Same as in your github discussions.
3. As you've already been told elsewhere, I have been the only CCRL tester to test Shashchess and SugaR, and I gave you my reasons why I chose those two in particular, and why I had no intention of testing any others.
4. I have also told you that if you announce your engine in the Talkchess forum, you'll find a couple of people over there who love to test Stockfish clones/derivatives.

Now with respect, best to just accept all of that,
Graham.

PS - I think it's really great that you've trained an original net. It would be equally great if you could also write your own engine. You would get plenty of help in these Discord engine programming channels.
https://discord.gg/mxwBj57f
https://discord.gg/ktU37yUT
https://discord.gg/deC9vuBV9R
WampServer
Posts: 5
Joined: Thu Mar 21, 2024 6:07 pm
Sign-up code: 10159

Re: Stockfish derivative engines and registration problem

Post by WampServer »

2. In Discord engine programming channels, some top engine authors including a couple involved with Stockfish, have alluded to the fact that your engine is basically Stockfish, but with an originally trained net. Same as in your github discussions.
My thread was locked in the Official_Stockfish only after 7 hours. Doesn't this prove that they hate competition? They didn't give any logical reason that my engine is just a copy of Stockfish. I would like to know who the top engine authors are including a couple involved with Stockfish.

Code: Select all

# Stockfish, a UCI chess playing engine derived from Glaurung 2.1
Does this comment prove that my engine is copy paste Stockfish?!
3. As you've already been told elsewhere, I have been the only CCRL tester to test Shashchess and SugaR, and I gave you my reasons why I chose those two in particular, and why I had no intention of testing any others.
A large number of commits does not determine the quality of an engine. A large number of features does not determine the quality of the engine. From my opinion, Lifish is better than ShashChess and should be registered for CCRL. It doesn't matter if you don't want to register my engine for personal reasons but know that there will not be a fair competition.
Mathmagician
Posts: 5
Joined: Thu Feb 15, 2024 7:16 am
Sign-up code: 10159

Re: Stockfish derivative engines and registration problem

Post by Mathmagician »

WampServer wrote: Sat Mar 23, 2024 4:29 am Does this comment prove that my engine is copy paste Stockfish?!
How else do you explain Stockfish's copyright notice being present in all of your files?
User avatar
Gabor Szots
Posts: 12849
Joined: Sat Dec 09, 2006 6:30 am
Sign-up code: 10159
Location: Szentendre, Hungary

Re: Stockfish derivative engines and registration problem

Post by Gabor Szots »

Yes, this is what I wrote:

"a CCRL member might pick it up for testing"

Which means each member may decide if he tests it or not. I have decided not to test it. I am simply not interested. Others may decide otherwise.

I can understand your frustration but that's how it is.

My stance might change by time. I'm going to follow the development of your engine and may decide to pick it up for testing when I feel like it.
Ray
Posts: 22606
Joined: Sun Dec 18, 2005 6:33 pm
Sign-up code: 10159
Location: NZ

Re: Stockfish derivative engines and registration problem

Post by Ray »

WampServer wrote: Sat Mar 23, 2024 4:29 am From my opinion, Lifish is better than ShashChess and should be registered for CCRL. It doesn't matter if you don't want to register my engine for personal reasons but know that there will not be a fair competition.
There is no registration process. There are no engines "registered" with CCRL.
WampServer
Posts: 5
Joined: Thu Mar 21, 2024 6:07 pm
Sign-up code: 10159

Re: Stockfish derivative engines and registration problem

Post by WampServer »

Mathmagician wrote: Sat Mar 23, 2024 5:15 am How else do you explain Stockfish's copyright notice being present in all of your files?
My chess engine is derived from Stockfish but not only copy paste! For example, take a look at search.cpp
https://github.com/EarlyEdition/Lifish- ... search.cpp
Gabor Szots wrote: Sat Mar 23, 2024 5:56 am I can understand your frustration but that's how it is.
It′s so hard to explain what I'm feeling but I guess it′s ok, Cause I'll keep believing.
Gabor Szots wrote: Sat Mar 23, 2024 5:56 am I'm going to follow the development of your engine and may decide to pick it up for testing when I feel like it.
I've created a chess match between my engine and Stockfish today. Does this convince you to register my engine?

Code: Select all

[Event Blitz 5.0min+3.0sec"]
[Date "2024.03.23"]
[White "Stockfish 16"]
[Black "Lifish+HalfKP v1.0"]
[Result "1-0"]
[TimeControl "300+3"]
[ECO "D31"]

1. d4 {0.31/25 6 Both last book move} d5 {0.16/28 37} 2. c4 {0.27/25 4} c6 {
0.10/28 26 (e6)} 3. Nc3 {0.32/25 3 (e3)} Nf6 {0.24/28 14} 4. e3 {0.25/25 3 
(cxd5)} e6 {0.14/28 22 (Bf5)} 5. Qc2 {0.26/25 4 (Nf3)} Be7 {0.19/28 56 (Bd6)}
6. b3 {0.32/25 3 (Nf3)} b6 {0.23/29 58 (0-0)} 7. Bb2 {0.32/25 1} O-O {0.26/28
18} 8. Nf3 {0.38/25 5} Bb7 {0.19/28 40 (Nbd7)} 9. Rd1 {0.31/25 6} Na6 {0.20/29
97 (Nbd7)} 10. Be2 {0.48/25 11 (Qb1)} Bd6 {0.21/28 40 (c5)} 11. a3 {0.38/25 6 
(0-0)} Qe7 {0.31/28 53 (Rc8)} 12. O-O {0.34/25 4} Rfc8 {0.31/28 58 (Rac8)} 13.
e4 {0.96/25 3 (b4)} Bxa3 {0.00/28 23 (dxe4)} 14. e5 {1.35/25 3 (Bxa3)} Nd7 {
0.00/28 4} 15. Bxa3 {1.32/25 1} Qxa3 {0.48/28 38} 16. Ng5 {1.35/25 3} Nf8 {
0.90/29 60} 17. Bd3 {1.42/25 2} g6 {1.05/28 76} 18. Qd2 {1.38/25 2} Qe7 {
0.86/28 32} 19. cxd5 {1.27/25 3 (f4)} cxd5 {0.35/28 10} 20. Nb5 {1.36/25 2} Rd8
{0.67/28 15} 21. Nd6 {1.26/25 3} Rxd6 {0.43/29 10} 22. exd6 {1.41/25 2} Qxd6 {
0.70/28 13} 23. Bxa6 {1.42/25 4} Bxa6 {0.55/28 13} 24. Rfe1 {1.39/25 5} Bb5 {
0.72/29 24 (Nd7)} 25. Rc1 {1.32/25 2 (h4)} Bd7 {0.86/28 48 (Nd7)} 26. Nf3 {
1.58/25 6} f6 {1.27/29 58} 27. Rc3 {1.56/25 4 (h4)} a5 {1.39/28 52 (Re8)} 28.
g4 {1.70/25 4} g5 {1.22/29 30 (Re8)} 29. h4 {1.74/25 5} Qf4 {2.11/28 22 (h6)}
30. Qxf4 {2.16/25 4} gxf4 {2.30/28 27} 31. g5 {2.21/25 2 (Rc7)} Ra7 {2.47/28
28 (Kg7)} 32. gxf6 {2.79/25 3} Kf7 {2.25/29 13} 33. Rec1 {2.94/25 2 (Ne5+)}
Kxf6 {2.37/31 6} 34. Rc7 {2.50/25 1} Rxc7 {2.54/28 6} 35. Rxc7 {2.50/25 0} Bb5
{2.58/28 7 (a4)} 36. Ne5 {3.11/25 2} Be8 {2.92/28 3} 37. Rc8 {3.16/25 1 (f3)}
Ke7 {2.94/28 2} 38. Rc7+ {3.04/25 1 (Kg2)} Kd6 {2.79/28 5} 39. Rc8 {2.92/25 0}
Ke7 {2.82/28 2} 40. Kg2 {3.09/25 0} Ng6 {3.18/28 3} 41. Nxg6+ {3.10/25 0} Bxg6
{2.86/28 3} 1-0
Ray wrote: Sat Mar 23, 2024 7:02 am There is no registration process.
I think I have to wait forever! or never?
Ray
Posts: 22606
Joined: Sun Dec 18, 2005 6:33 pm
Sign-up code: 10159
Location: NZ

Re: Stockfish derivative engines and registration problem

Post by Ray »

WampServer wrote: Sat Mar 23, 2024 10:27 pm
Ray wrote: Sat Mar 23, 2024 7:02 am There is no registration process.
I think I have to wait forever! or never?
Given that in CCRL's 18+ year history no engine has ever been registered with CCRL, I'll leave you to draw your own conclusion.
WampServer
Posts: 5
Joined: Thu Mar 21, 2024 6:07 pm
Sign-up code: 10159

Re: Stockfish derivative engines and registration problem

Post by WampServer »

Ray wrote: Sat Mar 23, 2024 7:02 am Given that in CCRL's 18+ year history no engine has ever been registered with CCRL, I'll leave you to draw your own conclusion.
If so, How is it possible for an engine(Torch v1) to be registered in less than a month and be among the top 3 chess rankings?
  • 1-Stockfish 20230613 64-bit 4CPU
  • 2-Torch v1 64-bit 4CPU 3625
  • 3-Dragon by Komodo 3.3 64-bit 4CPU
Meanwhile, the author of this engine(Andrew Grant) commented on my topic in the Official_Stockfish and said my engine should not be registered.
https://github.com/official-stockfish/S ... nt-8872672

Is this fair?
User avatar
Gabor Szots
Posts: 12849
Joined: Sat Dec 09, 2006 6:30 am
Sign-up code: 10159
Location: Szentendre, Hungary

Re: Stockfish derivative engines and registration problem

Post by Gabor Szots »

WampServer wrote: Sun Mar 24, 2024 7:35 pm If so, How is it possible for an engine(Torch v1) to be registered in less than a month and be among the top 3 chess rankings?
I don't understand what you don't understand. There is no registration. Engines are announced or found and we individually decide if we test one or not. Maybe I don't understand what registration means.
Ray
Posts: 22606
Joined: Sun Dec 18, 2005 6:33 pm
Sign-up code: 10159
Location: NZ

Re: Stockfish derivative engines and registration problem

Post by Ray »

Andrew is not CCRL. How many times do we have to tell you that there is no registration process and not a single engine is registered with CCRL?? None, zilch.

There is no further point in continuing this discussion. For the first time in CCRL history, this thread will be locked.
Locked