My review of Strelka Code

[mclane]

as i said before rolf. stop starting a meta discussion.

do you have test results of strelka to offer to the public ?
if so show them.
if you want to start a meta discussion use CTF.
if you want to complain, do it via the "!" button.
if you want to chat, use the PM or email. my last word about this.
thanks.

[mariaclara]

well, at least this person is honest enough to say he views things in a very biased manner.

I think Strelka author has to understand that this suspicion is there not because of him but because of the lot of frauds who preceeded him.
_________________

meaning, he looks at things not through open, fresh eyes.
but through tainted, colored glasses.


hehehe........
have fun

[Eelco de Groot]

The code that was passed on to me plays the same as the released binary. It is original work and not a clone. If you have questions about the code or how it works I ask that you direct them to author as it is his work.


Bryan

Thanks for doing the analysis Bryan. I realize that you don't want to go into much detail about somebody else's code, but I must say I would be more at ease with a little more detail. For instance if anyone can say whether Strelka uses bitboards or not. Vasik Rajlich did say in his forum that he thinks Strelka is a bitboard program, but I have not seen any other statement about this myself, I believe Yuri Osipov only states in the interview it is a regular alpha-beta program. If Strelka does not use bitboards it would of course be a very strong indication for its originality. Even if it does bitboards, but different ones than Rybka especially in the evaluation side of things, that would still be an important fact to know.

What I am missing up to now is more about the development history of Rybka, how many earlier versions did Yuri Osipov try, how did he test them and if he implemented bitboards or something similar like the big datastructure that seems to be part of the program, is this the reason that Strelka can achieve such high nodes per second. If Strelka uses bitboards, a 64-bit compile would probably be a lot faster and gain some more elo? Hopefully we will hear a bit more about it from Sergei Markov or from Yuri Osipov himself. He does not have to make his sources public by all means.

To be honest with all the facts that we now have I would still guess that Strelka and Rybka must be using similar principles somewhere. I suppose it is inevitable that one day a program would, or will, come along that does work in similar ways as Rybka does, that does not mean this program is or would be a clone or has to be directly based on Rybka's code. Trying to deduce how Rybka works looking at its output, up to a point I think that is perfectly legal, copying code of course is not. So for me Strelka could still be somewhere in a grey area, and the background that its author has given so far I personally find rather vague. But that is my personal opinion, I hope I have the right to say this.

Regards, Eelco

[Dave McClain]

The code that was passed on to me plays the same as the released binary. It is original work and not a clone. If you have questions about the code or how it works I ask that you direct them to author as it is his work.


Bryan

Thanks for doing the analysis Bryan. I realize that you don't want to go into much detail about somebody else's code, but I must say I would be more at ease with a little more detail. For instance if anyone can say whether Strelka uses bitboards or not. Vasik Rajlich did say in his forum that he thinks Strelka is a bitboard program, but I have not seen any other statement about this myself, I believe Yuri Osipov only states in the interview it is a regular alpha-beta program. If Strelka does not use bitboards it would of course be a very strong indication for its originality. Even if it does bitboards, but different ones than Rybka especially in the evaluation side of things, that would still be an important fact to know.

What I am missing up to now is more about the development history of Rybka, how many earlier versions did Yuri Osipov try, how did he test them and if he implemented bitboards or something similar like the big datastructure that seems to be part of the program, is this the reason that Strelka can achieve such high nodes per second. If Strelka uses bitboards, a 64-bit compile would probably be a lot faster and gain some more elo? Hopefully we will hear a bit more about it from Sergei Markov or from Yuri Osipov himself. He does not have to make his sources public by all means.

To be honest with all the facts that we now have I would still guess that Strelka and Rybka must be using similar principles somewhere. I suppose it is inevitable that one day a program would, or will, come along that does work in similar ways as Rybka does, that does not mean this program is or would be a clone or has to be directly based on Rybka's code. Trying to deduce how Rybka works looking at its output, up to a point I think that is perfectly legal, copying code of course is not. So for me Strelka could still be somewhere in a grey area, and the background that its author has given so far I personally find rather vague. But that is my personal opinion, I hope I have the right to say this.

Regards, Eelco

Eelco,

Here is the invitation from the Strelka authour:

http://www.talkchess.com/forum/viewtopi ... ht=#117387

[Philippe]

Ok, if you like, let´s to change the words, it´s the same thing. Some questionings had been very
rudes and, to say the truth, some people had affirmed that Strelka was a clone.
The more important now is that Strelka, as you wrote, are welcome into
the fraternity, but I think the author didn´t deserved to read some words that are write here.

Paulo Soares[/quote]

Strelka is made by a russian author and I suspect some shameful anti-russian spirit of some posters ...

[Eelco de Groot]

Eelco,

Here is the invitation from the Strelka authour:

http://www.talkchess.com/forum/viewtopi ... ht=#117387

Hi Dave, thanks for the reply, yes I think that is positive. Moscow is a bit far away for me regretfully but I am sure that in time Sergei Markoff can tell us more, or we will hear about some more details in a future interview. I agree with you if you say that Yuri Osipow has not been secretive about his work or he would never have handed over his source code to anyone. Hopefully then he will tell us some more details about how Strelka was born in the future, there is no better way to make some publicity for his program if he would want that, than by giving some personal details. So I hope for some more interviews like the last one.

Michael Gurevich by the way has been very helpful by making a Russian to German translation of this first interview - which was held before the visit from Sergei Markoff to Yuri Osipov and the excange of sourcecode between them -, available in the CSS forum. The translation from Michael is much better readable than the automatic Babelfish translations. If you can read German it is certainly worth reading: http://f23.parsimony.net/forum50826/messages/166127.htm

Regards, Eelco

[smirobth]

I have examined the code and Strelka is not a clone. There is a 1 MB data table called 'material' inside Strelka, which is what I thought was some encrypted binary.

It's ironic that I complained about the 'primitive' Winboard parser.
Strelka's winboard parser borrows heavily from Beowulf.

Hi Dan,
I am a little surprised that you shared anything specific about the inner workings of a closed source chess engine.

[Dave McClain]

Eelco,

Here is the invitation from the Strelka authour:

http://www.talkchess.com/forum/viewtopi ... ht=#117387

Hi Dave, thanks for the reply, yes I think that is positive. Moscow is a bit far away for me regretfully but I am sure that in time Sergei Markoff can tell us more, or we will hear about some more details in a future interview. I agree with you if you say that Yuri Osipow has not been secretive about his work or he would never have handed over his source code to anyone. Hopefully then he will tell us some more details about how Strelka was born in the future, there is no better way to make some publicity for his program if he would want that, than by giving some personal details. So I hope for some more interviews like the last one.

Michael Gurevich by the way has been very helpful by making a Russian to German translation of this first interview - which was held before the visit from Sergei Markoff to Yuri Osipov and the excange of sourcecode between them -, available in the CSS forum. The translation from Michael is much better readable than the automatic Babelfish translations. If you can read German it is certainly worth reading: http://f23.parsimony.net/forum50826/messages/166127.htm

Regards, Eelco

Eelco,

I think that is very positive and in time someone will take the invitation. I was skeptical at first myself because of all the clones we have dealt with recently.

A biography from the authour of the program would be most interesting too. It will come in time.........

[smirobth]

I have examined the code and Strelka is not a clone. There is a 1 MB data table called 'material' inside Strelka, which is what I thought was some encrypted binary.

It's ironic that I complained about the 'primitive' Winboard parser.
Strelka's winboard parser borrows heavily from Beowulf.

Have you compared the source code to the source code of Rybka 1.0 beta? If not, than your statement is meaningless.

If so, a simple question about the main evaluation functions.

Don't you see the same bonuses in both programs everywhere?

Of course Bryan may also respond.

Rob,
You already have egg all over your face based on your over the top accusations against Strelka. Isn't it time for you to cut your losses? Perhaps you could take the following advice to heart: "Better to remain silent and be thought a fool than to speak out and remove all doubt." - Abraham Lincoln

[Dirt]

...
Strelka's winboard parser borrows heavily from Beowulf.
;-)

Does "borrows heavily" mean copied? Beowulf appears to be GPL software, so if Strelka uses code from the Beowulf it would be a copyright violation.