Marek why don't you help current SF? You can get you ideas tested with fishtest. Maybe Gary can write a patch in python so that you can test only certain positions! It would be great publicity! Maybe you can hide your changes so no prying eyes can steal your ideas?
Worth a try!
Congrats and keep up the good work!
Regards
Sting-sf-4 release
Moderators: hgm, Rebel, chrisw
-
- Posts: 101
- Joined: Sat May 28, 2011 11:47 pm
Re: Sting-sf-4 release
There is always a light at end of the tunnel. Just make sure it isn't a train!
-
- Posts: 1136
- Joined: Sun Feb 14, 2010 10:02 pm
Re: Sting-sf-4 release
My project is not an ELO racing engine. Sting is to maintain the parent ELO (SF 211) and it is all in this question.Red Hood wrote:Marek why don't you help current SF? You can get you ideas tested with fishtest. Maybe Gary can write a patch in python so that you can test only certain positions! It would be great publicity! Maybe you can hide your changes so no prying eyes can steal your ideas?
Worth a try!
Congrats and keep up the good work!
Regards
I try to do someting other and I hope it is clear.
I am closer to show that the logical thinking (anything it means ) can be only a choise no synthesis.
I am very happy that I may use the Stockfish's open source code.
Many thanks for the authors and I wish the Stockfish's team the highest ELO.
The lack of the 64-bit compilation is not a problem for me.
Maybe, I can't be friendly, but let me be useful.
-
- Posts: 101
- Joined: Sat May 28, 2011 11:47 pm
Re: Sting-sf-4 release
You misunderstood. You keep doing what you are doing now but i also wanted to suggest that you can try to add some elo to current SF. Maybe also to exploit fishtest for your work as well.
There is always a light at end of the tunnel. Just make sure it isn't a train!
-
- Posts: 2439
- Joined: Sat Sep 03, 2011 7:25 am
- Location: Berlin, Germany
- Full name: Stefan Pohl
Re: Sting-sf-4 release
I did a test of Sting SF 3 for my LS-ratinglist and would like to test Sting SF 4, too, but I would prefer a 64bit-compile, because the tested Sting SF 3 was a 64bit version and for the comparability of the two Sting-versions, a 64bit-version of Sting SF 4 would be much better...perhaps somebody could create a 64bit-compile, please?lech wrote:My project is not an ELO racing engine. Sting is to maintain the parent ELO (SF 211) and it is all in this question.Red Hood wrote:Marek why don't you help current SF? You can get you ideas tested with fishtest. Maybe Gary can write a patch in python so that you can test only certain positions! It would be great publicity! Maybe you can hide your changes so no prying eyes can steal your ideas?
Worth a try!
Congrats and keep up the good work!
Regards
I try to do someting other and I hope it is clear.
I am closer to show that the logical thinking (anything it means ) can be only a choise no synthesis.
I am very happy that I may use the Stockfish's open source code.
Many thanks for the authors and I wish the Stockfish's team the highest ELO.
The lack of the 64-bit compilation is not a problem for me.
Stefan
-
- Posts: 1136
- Joined: Sun Feb 14, 2010 10:02 pm
Re: Sting-sf-4 release
Sorry Stefan, your work for Sting is not that important.pohl4711 wrote:...perhaps somebody could create a 64bit-compile, please?
Stefan
You, Graham and Other are very important for ELO engines, but it is not the reason to try to ask.
Maybe, I can't be friendly, but let me be useful.
-
- Posts: 1136
- Joined: Sun Feb 14, 2010 10:02 pm
Re: Sting-sf-4 release
To avoid in the future some offensive words that Sting is only some devilish trick, I will try to explain my work much deeper.
Both the positions (represent some wide groups) are very easy to solve for chessplayers.
[d]8/5pKN/5Pp1/5kP1/8/Bb6/8/7q b - - bm Qxh7;
[d]5K2/k4p1p/5p1p/1p3p2/bP6/1p6/1P5p/7B w - - bm Ke8(e7);
Also both are extremaly tough for engines. The problem is in so-called "null move search".
By my long and hard work I found a shape for the "null move search" which let us look forward to the solid chess engine.
There was needed to solve the problem of the "value draw" too, which can work like the black hole. I found the solution and still the value draw is ZERO.
Why it is important? At first it is normal or natural one. Moreover I will try to use the huge power (black hole) to evaluate zero (fortress).
I hope it will be in the next version (Sting SF 5) soon.
One condition is important for me too to maintain a similar ELO level of the parent egine (Stockfish 211).
Both the positions (represent some wide groups) are very easy to solve for chessplayers.
[d]8/5pKN/5Pp1/5kP1/8/Bb6/8/7q b - - bm Qxh7;
[d]5K2/k4p1p/5p1p/1p3p2/bP6/1p6/1P5p/7B w - - bm Ke8(e7);
Also both are extremaly tough for engines. The problem is in so-called "null move search".
By my long and hard work I found a shape for the "null move search" which let us look forward to the solid chess engine.
There was needed to solve the problem of the "value draw" too, which can work like the black hole. I found the solution and still the value draw is ZERO.
Why it is important? At first it is normal or natural one. Moreover I will try to use the huge power (black hole) to evaluate zero (fortress).
I hope it will be in the next version (Sting SF 5) soon.
One condition is important for me too to maintain a similar ELO level of the parent egine (Stockfish 211).
Maybe, I can't be friendly, but let me be useful.
-
- Posts: 2439
- Joined: Sat Sep 03, 2011 7:25 am
- Location: Berlin, Germany
- Full name: Stefan Pohl
Re: Sting-sf-4 release
OK. There will be no test of Sting SF4 for the LS-ratinglist (until there is no 64bit-compile)...lech wrote:Sorry Stefan, your work for Sting is not that important.pohl4711 wrote:...perhaps somebody could create a 64bit-compile, please?
Stefan
You, Graham and Other are very important for ELO engines, but it is not the reason to try to ask.
Stefan
-
- Posts: 1010
- Joined: Thu Sep 01, 2011 2:49 pm
Re: Sting-sf-4 release
Hi Marek,
Is it possible to get a 32-bit compile with no DLL dependencies?
I have some problems running your program.
This is the second time I have problems with your sting.
1st prob I remember Sting2 or Sting3 has some bug, buffer overflow, which
is blocked by DEP of windows.
Now this latest version, Sting 4, requires some VC9 DLL.
I think it is also possible to compile with no dependencies and will work
as standalone exe. Just like houdini1.5 or Rybka2.2 or all versions of
stockfish.
Is it possible to get a 32-bit compile with no DLL dependencies?
I have some problems running your program.
This is the second time I have problems with your sting.
1st prob I remember Sting2 or Sting3 has some bug, buffer overflow, which
is blocked by DEP of windows.
Now this latest version, Sting 4, requires some VC9 DLL.
I think it is also possible to compile with no dependencies and will work
as standalone exe. Just like houdini1.5 or Rybka2.2 or all versions of
stockfish.
-
- Posts: 1136
- Joined: Sun Feb 14, 2010 10:02 pm
Re: Sting-sf-4 release
Hi Mike,
Please try to download and install the Microsoft libraries:
http://www.microsoft.com/en-us/download ... aspx?id=29
Marek
Please try to download and install the Microsoft libraries:
http://www.microsoft.com/en-us/download ... aspx?id=29
Marek
Maybe, I can't be friendly, but let me be useful.