Page 29 of 47

Re: New engine releases 2019

Posted: Fri Jul 26, 2019 10:03 am
by Roland Chastain
Norbert Raimund Leisner wrote: Thu Jul 25, 2019 6:28 pm By the way: Is Sapeli 1.36-1.37 Windows anywhere available for download?
Hello! I made a compressed file with all versions (I believe) until 1.38.

https://drive.google.com/open?id=1bd1Gt ... UH-cj5lMOn

Re: funny bug in Nawito 1906

Posted: Fri Jul 26, 2019 10:21 am
by Guenther
ernesto wrote: Tue Jul 23, 2019 11:37 pm
Gabor Szots wrote: Tue Jul 23, 2019 7:51 pm
ernesto wrote: Tue Jul 23, 2019 7:44 pm Tmokonen, I'm the Nawito chess engine programmer and the files at https://github.com/etorresf/Nawito1906 constitute
The new version of the engine. I admit that I am a rookie on github, the first time I put something there, so more data is missing,
settings, etc. which will be added progressively.

The https://github.com/etorresf/Nawito-Chess-Engine repository should be deleted as it was created by an error of mine.

Thanks for your comment
Ernesto, for 1812 you provided a compressed file with the necessary things. Could you do it for 1906 as well?

Gabor Szots: all the things necessary for the engine to run are found at https://github.com/etorresf/Nawito1906, however tomorrow compress the 1906 folders to make it easier to download
Nawito 1906 does not comply 100% to the WB protocol and will not play under WB.
(It does under CuteChess though)

It gets into a funny loop with Winboard about 'feature done=1"

Code: Select all

650 <first : feature colors=0 draw=0 ics=1 myname="Nawito1906" ping=1 setboard=1 sigint=0 sigterm=0 variants="normal,nocastle"
650 >first : accepted colors
650 >first : accepted draw
650 >first : accepted ics
650 >first : accepted myname
650 >first : accepted ping
650 >first : accepted setboard
650 >first : accepted sigint
650 >first : accepted sigterm
650 >first : accepted variants
650 <first : feature done=1
650 >first : accepted done
650 <first : feature accepted
650 >first : rejected accepted
650 <first : feature accepted
650 >first : rejected accepted
650 <first : feature accepted
650 >first : rejected accepted
... loop goes on forever - inbetween it makes the move Nc3 though w/o thinking,
but after Blacks first move no search ever is started and the loop goes on until time ends.
Edit:
Ofc it will work adding Protocol V1 to its entries, but I don't think this was intended.

Re: New engine releases 2019

Posted: Fri Jul 26, 2019 10:54 am
by JohnWoe
Norbert Raimund Leisner wrote: Thu Jul 25, 2019 6:28 pm Defenchess 2.1 https://github.com/cetincan0/Defenchess/releases

By the way: Is Sapeli 1.36-1.37 Windows anywhere available for download?
I added Ronald's all: 1.16 -> 1.38 Windows builds are there.
https://github.com/SamuraiDangyo/Sapeli ... naries/bin

1.36 I fixed some memory leaks per valgrind.
Otherwise just very modest bug fixes / updates. As this +0.01 version number increase indicates.

Re: New engine releases 2019

Posted: Fri Jul 26, 2019 1:56 pm
by Roland Chastain
JohnWoe wrote: Fri Jul 26, 2019 10:54 am I added Ronald's all: 1.16 -> 1.38 Windows builds are there.
https://github.com/SamuraiDangyo/Sapeli ... naries/bin
Hello! I saw on GitHub that you released a 1.38-final version, and later you made some other commits to 1.38. Maybe you could delete 1.38 and 1.38-final releases and make a new 1.38 release with the latest code?

Re: New engine releases 2019

Posted: Fri Jul 26, 2019 2:38 pm
by JohnWoe
Roland Chastain wrote: Fri Jul 26, 2019 1:56 pm
JohnWoe wrote: Fri Jul 26, 2019 10:54 am I added Ronald's all: 1.16 -> 1.38 Windows builds are there.
https://github.com/SamuraiDangyo/Sapeli ... naries/bin
Hello! I saw on GitHub that you released a 1.38-final version, and later you made some other commits to 1.38. Maybe you could delete 1.38 and 1.38-final releases and make a new 1.38 release with the latest code?
Thanks!
Couldn't just re-release it as 1.38 so I made this very modest update.
I released this typo fix version as 1.39: https://github.com/SamuraiDangyo/Sapeli ... s/tag/1.39
I removed those older 1.38 releases.
In 1.38 Eval() I used WTM instead of wtm. That caused some issues. When calling Eval(0)... Now it's all good.

Re: New engine releases 2019

Posted: Fri Jul 26, 2019 3:38 pm
by Roland Chastain
JohnWoe wrote: Fri Jul 26, 2019 2:38 pm Couldn't just re-release it as 1.38 so I made this very modest update.
I released this typo fix version as 1.39: https://github.com/SamuraiDangyo/Sapeli ... s/tag/1.39
I removed those older 1.38 releases.
In 1.38 Eval() I used WTM instead of wtm. That caused some issues. When calling Eval(0)... Now it's all good.
Thank you. Updated my ZIP file.

Durandal

Posted: Fri Jul 26, 2019 5:09 pm
by Roland Chastain
Durandal is a UCI chess engine playing traditional chess, Fischer random chess, Capablanca chess and Capablanca random chess.

Discussion here.

Re: Durandal

Posted: Fri Jul 26, 2019 5:46 pm
by JohnWoe
Roland Chastain wrote: Fri Jul 26, 2019 5:09 pm Durandal is a UCI chess engine playing traditional chess, Fischer random chess, Capablanca chess and Capablanca random chess.

Discussion here.
Thanks for the binaries! I added them here. https://github.com/SamuraiDangyo/Sapeli ... naries/bin
Too many Sapeli releases lately I'm gonna focus only on Shuriken now on. :P

Durandal looks very interesting! My own goal is to make Shuriken play as many variants as possible. Next mission is to speed it up.

make tactics
Shuriken can actually solve some simple tactics.

Code: Select all

ruby shuriken.rb -tactics
~~~ antitactics ~~~
1. move j3j4 | ok
2. move d3d4 | ok
3. move f3f4 | ok
4. move a3a4 | ok
= 4 / 4
~~~ tactics ~~~
1. move h2h8 | ok
2. move c5d7 | ok
3. move O-O-O | ok
4. move O-O | ok
5. move d5e3 | ok
6. move d5e3 | ok
7. move d4e6 | ok
8. move c1i7 | error
9. move b1c3 | error
10. move d1b1 | error
11. move g1j4 | ok
= 8 / 11

Re: Durandal

Posted: Fri Jul 26, 2019 7:34 pm
by Roland Chastain
JohnWoe wrote: Fri Jul 26, 2019 5:46 pmThanks for the binaries! I added them here. https://github.com/SamuraiDangyo/Sapeli ... naries/bin
I discovered a problem with some versions. Please see here.
JohnWoe wrote: Fri Jul 26, 2019 5:46 pmDurandal looks very interesting! My own goal is to make Shuriken play as many variants as possible. Next mission is to speed it up.
If you look for an opponent easy to beat, Durandal should do the job. :wink:

Re: funny bug in Nawito 1906

Posted: Sat Jul 27, 2019 3:49 pm
by ernesto
Thanks Guenther

I will work on solving this bug




Guenther wrote: Fri Jul 26, 2019 10:21 am
ernesto wrote: Tue Jul 23, 2019 11:37 pm
Gabor Szots wrote: Tue Jul 23, 2019 7:51 pm
ernesto wrote: Tue Jul 23, 2019 7:44 pm Tmokonen, I'm the Nawito chess engine programmer and the files at https://github.com/etorresf/Nawito1906 constitute
The new version of the engine. I admit that I am a rookie on github, the first time I put something there, so more data is missing,
settings, etc. which will be added progressively.

The https://github.com/etorresf/Nawito-Chess-Engine repository should be deleted as it was created by an error of mine.

Thanks for your comment
Ernesto, for 1812 you provided a compressed file with the necessary things. Could you do it for 1906 as well?

Gabor Szots: all the things necessary for the engine to run are found at https://github.com/etorresf/Nawito1906, however tomorrow compress the 1906 folders to make it easier to download
Nawito 1906 does not comply 100% to the WB protocol and will not play under WB.
(It does under CuteChess though)

It gets into a funny loop with Winboard about 'feature done=1"

Code: Select all

650 <first : feature colors=0 draw=0 ics=1 myname="Nawito1906" ping=1 setboard=1 sigint=0 sigterm=0 variants="normal,nocastle"
650 >first : accepted colors
650 >first : accepted draw
650 >first : accepted ics
650 >first : accepted myname
650 >first : accepted ping
650 >first : accepted setboard
650 >first : accepted sigint
650 >first : accepted sigterm
650 >first : accepted variants
650 <first : feature done=1
650 >first : accepted done
650 <first : feature accepted
650 >first : rejected accepted
650 <first : feature accepted
650 >first : rejected accepted
650 <first : feature accepted
650 >first : rejected accepted
... loop goes on forever - inbetween it makes the move Nc3 though w/o thinking,
but after Blacks first move no search ever is started and the loop goes on until time ends.
Edit:
Ofc it will work adding Protocol V1 to its entries, but I don't think this was intended.