Sergei Markow about Strelka

[Werner]

Hi all,
I got these mail from Sergei with the request to post it here:

"Hi Werner!

> and please tell me in english: what do you know about Strelka?

I have examined it's source code today. It's like no one of open source engines and also it is not contains any encrypted binaries. There are no reason to think that it is a clone. I have spoken with Osipov for a long time, he is expirienced enought in computer chess programming topics. Finally we exchanged with our sources: now he has SmarThink src and I have Strelka src. I hope we will cooperate in our research. Also you can see photos of me and Osipov: http://www.sdchess.ru/news.htm

About Dann Corbit's post. Strelka has a table of precalculated bitboards. After start it performs some xor operations on this table. That's why Dann decided that Strelka has some encrypted bynary inside.

The biggest table is precalculated material table which is quite similar to Fruit's or Toga's material hash table.

About behavior with two kings... I don't know why it happens. Possible because both Rybka and Strelka are bitboard engines and may be there are something similar in data representation. BTW this is not a really serious reason to think that Strelka is a Rybka clone.

About Rybka cloning. Of course, I never seen Rybka sources (but Vasik seen my I think there are no any reason to think that Vasik presented Rybka's source to Osipov. Reverse engeneering of Rybka to make structured C-sources is really a Sisyphus work. Theoretically possible, but practically incredible.

Please, post it to talkchess.com forum. My old account was removed from server (probably because mail.ru e-mail now is prohibited on talkchess). I tried to register new account (with my Google mail), but still there are no confirmation from admin.

Regards,
Sergei"

I have had many contacts before when testing Smarthink 1.0 and I am very pleased to post his statement here. Hope we will soon here more from him and his engine.

[PauloSoare]

Werner, I am happy with this words, because Strelka seemed
to be a very strong engine if we consider that is a beta version.

Paulo Soares

[Zlaire]

My main concern is how an 'experienced chess programmer' can't write a decent winboard interface.

I can't shake the bad feeling, it probably goes away with a new improved version or so.

All credit to Osipov if it's not a clone.

[Dann Corbit]

Hi all,
I got these mail from Sergei with the request to post it here:

"Hi Werner!

> and please tell me in english: what do you know about Strelka?

I have examined it's source code today. It's like no one of open source engines and also it is not contains any encrypted binaries. There are no reason to think that it is a clone. I have spoken with Osipov for a long time, he is expirienced enought in computer chess programming topics. Finally we exchanged with our sources: now he has SmarThink src and I have Strelka src. I hope we will cooperate in our research. Also you can see photos of me and Osipov: http://www.sdchess.ru/news.htm

About Dann Corbit's post. Strelka has a table of precalculated bitboards. After start it performs some xor operations on this table. That's why Dann decided that Strelka has some encrypted bynary inside.

It is possible that IdaPro became confused, but it showed almost the entire program as a large data segment (.rdata) and almost no code routines. When I attached a runtime debugger, I saw many other functions that were not revealed by IdaPro.

The biggest table is precalculated material table which is quite similar to Fruit's or Toga's material hash table.

About behavior with two kings... I don't know why it happens. Possible because both Rybka and Strelka are bitboard engines and may be there are something similar in data representation. BTW this is not a really serious reason to think that Strelka is a Rybka clone.

About Rybka cloning. Of course, I never seen Rybka sources (but Vasik seen my I think there are no any reason to think that Vasik presented Rybka's source to Osipov. Reverse engeneering of Rybka to make structured C-sources is really a Sisyphus work. Theoretically possible, but practically incredible.

Please, post it to talkchess.com forum. My old account was removed from server (probably because mail.ru e-mail now is prohibited on talkchess). I tried to register new account (with my Google mail), but still there are no confirmation from admin.

Regards,
Sergei"

I have had many contacts before when testing Smarthink 1.0 and I am very pleased to post his statement here. Hope we will soon here more from him and his engine.

I hope it is true that it is all original. But it still seems very strange to me that such a polished engine should have such a primitive Winboard interface. I suppose it is possible that he wrote some other native interface and only did the Winboard interface in the last minute and has not had time to correct it.

[shiv]

I am now inclined to accept Sergei Markov's words.

However, we can do the following test to see if this is really a clone using QEMU, an open source emulation program (http://fabrice.bellard.free.fr/qemu/).

What we can do is run Rybka (after installing windows and both Rybka and Strelka in a Qemu image) and play a game and enable the logs to save information on which basic blocks were hit (and their order). Then we stop Rybka and run Strelka and get the same log. If a large percentage of the basic block cache is accessed by Strelka too, then there is a chance it is a clone. Of course, as we do not have the source to either Strelka or Rybka, a detailed log and basic block analysis is needed. Navigating thru the logs could become tricky of course.

I guess Dann wanted to something similar with IdaPro but somehow IdaPro became confused.

Anyway, as far as I am concerned, Strelka should have the benefit of the doubt for now.

[Leto]

Thanks very much Werner, from the very first time I've heard of the allegations that Strelka was a Rybka clone I didn't believe any of it. There were just far too many differences, I think people were simply seeing what they wanted to see. I remember hearing many weird nonsense such as "Strelka must be a clone because it has a Russian name" and such.

[Sylwy]

Hi all !
Now all is clear for me: Strelka was first of all a native engine.
Now is the strongest Winboard engine in computer chess history.
I hope a new Strelka will be without major problems (time management & basic endgames).
Be happy !
Sylwy

[Tony Thomas]

If it turns out that I was wrong about Strelka, I will apologize, but it is too early now.