Is Belka a Rybka?

Discussion of anything and everything relating to chess playing software and machines.

Moderators: Harvey Williamson, Dann Corbit, hgm

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
User avatar
slobo
Posts: 2331
Joined: Mon Apr 09, 2007 3:36 pm

Re: Is Belka a Rybka?

Post by slobo » Sun Dec 02, 2007 9:53 pm

GenoM wrote:
matejst wrote:Quote: "The same with you. Serbia isnt a member in the EU, so if I wanted to get a title against you personally, I couldnt get it because of a practical barrier, otherwise you wouldnt talk so lightly and irresponsible against me and about the whole topic. What is also of interest is this: you claim to be this or that, you are a player this or that, but I cant check if you are really authentically this person. Of course the mentioned names might really and authentically exist. Know what I mean? Often people say, but look into the phonebook there in that country. But how could I know that you who is actually writing there is the one out of the phonebook. How this could be researched? Of course it's impossible to do. Unless in a Western country the similar laws are existing and then I could make a formal appeal. But in Russia or Serbia I couldnt."

After such nonsense... I am amazed.

Kind regards!

BS
You simply did not know Rolf. He knows everything and is always right.
Hi Geno,

Sometimes he is right. But in this case I have my doubts. His basic idea is: if you use other people´s ideas, and make the best engine, then you are right, but, if you don´t succeed in making the best engine, then you are wrong, then you are cloner.

A tipical western reasoning: a winner is allowed to make dubious things, others are not.

This is called: double standards.
"Well, I´m just a soul whose intentions are good,
Oh Lord, please don´t let me be misunderstood."

User avatar
Rolf
Posts: 6081
Joined: Fri Mar 10, 2006 10:14 pm
Location: Munster, Nuremberg, Princeton

Re: Is Belka a Rybka?

Post by Rolf » Sun Dec 02, 2007 11:04 pm

slobo wrote:
GenoM wrote:
matejst wrote:Quote: "The same with you. Serbia isnt a member in the EU, so if I wanted to get a title against you personally, I couldnt get it because of a practical barrier, otherwise you wouldnt talk so lightly and irresponsible against me and about the whole topic. What is also of interest is this: you claim to be this or that, you are a player this or that, but I cant check if you are really authentically this person. Of course the mentioned names might really and authentically exist. Know what I mean? Often people say, but look into the phonebook there in that country. But how could I know that you who is actually writing there is the one out of the phonebook. How this could be researched? Of course it's impossible to do. Unless in a Western country the similar laws are existing and then I could make a formal appeal. But in Russia or Serbia I couldnt."

After such nonsense... I am amazed.

Kind regards!

BS
You simply did not know Rolf. He knows everything and is always right.
Hi Geno,

Sometimes he is right. But in this case I have my doubts. His basic idea is: if you use other people´s ideas, and make the best engine, then you are right, but, if you don´t succeed in making the best engine, then you are wrong, then you are cloner.

A tipical western reasoning: a winner is allowed to make dubious things, others are not.

This is called: double standards.
But this is not what I had said. That being weaker means you are a clone and wrong. I said something else.
-Popper and Lakatos are good but I'm stuck on Leibowitz

Uri Blass
Posts: 8828
Joined: Wed Mar 08, 2006 11:37 pm
Location: Tel-Aviv Israel

Re: Is Belka a Rybka?

Post by Uri Blass » Mon Dec 03, 2007 8:39 am

mjlef wrote:
Uri Blass wrote:
mjlef wrote:Here is an interesting experiment. Talke Belka 1.8.13. It comes with two files. One is a header file (EvalConsts.h) describing what some variables do. Although the comments ar ein Russian, I had a Bulgarian friend translate the comments for me. Variable names are very close to what is used in the Piece Square Table (PST.cpp) file that comes with Fruit. For example:

extern int BishopLine[8];

in Fruit this is:

static const int BishopLine[8] = {
-3, -1, +0, +1, +1, +0, -1, -3,
};

In fact, most of the names appear in Fruit and have the same apparent meanings. There are som new terms (not a lot), sveral involving a more detailed passed pawn evaluation.

So, I go to thinking. the person.txt file lets you set new values to be used for any of the terms in person.txt. If you have an empty person.txt, it reverst all values to whatever the defaults are. You can quikly figure out what the default values are by putting numbers in one line like this:

BishopLine 0 0 0 0 0 0 0 0

that would set these values to all 0. You then run a short search to a fixed depth, change the values and run again. When I did this, and used the fact written here that 3399 is a "pawn", so 33.9 would be 1/100 th a pawn (used in Fruit), the values for BishopLine when divided by 33.9 come out to:

-3, -1, 0, 1, 1, 0,-1, -3

Exactly the values used in Fruit 2.1. And this is true for all the "Line" variables I tried, and many of the other constants as well. If Belka is a clone it certainly has a lot of Fruit in it.

This seems to be evidence that the author starte dwith Fruit as a model, which I think was mentioned here.
Note that I looked at the code of strelka1.8 and did not find the string
BishopLine inside of it

I also looked for [8] because it may have a different name and I found no names that are about piece square table and I found only names that are about passed pawns.

Maybe Belka include parts of fruit that strelka does not have.

Uri
I have never seen Strelka code, just the apparent header file for Belka. One possibility is the piece square tables are already populated in Strelka, while Belka, being a development/test version, fills them in using a Fruit-like method of weights along files and columns. You might find tables indexed for pieces in Strelka with something like 64 values (one for each board square) and just already filled in. Things like BishopLine just make it easier to change weights while you experiment.

Mark
Yes

It seems that the code of strelka simply hide formulas that are used to calculate the piece square table when fruit does not hide them.

I did not check if the code does the same thing as fruit but even if it does the same thing it seems to me that starting from fruit and changing the code is not an efficient way to do it in this way and it is probably more simple to understand the code of fruit in english words and later write code that does the same thing with a different data structure.

I continue to believe that no copy and paste was done even if the algorithm is the same as fruit(I do not know if the algorithm of strelka is the same as fruit)

Uri

User avatar
slobo
Posts: 2331
Joined: Mon Apr 09, 2007 3:36 pm

Re: Is Belka a Rybka?

Post by slobo » Mon Dec 03, 2007 8:56 am

Rolf wrote:
slobo wrote:
GenoM wrote:
matejst wrote:Quote: "The same with you. Serbia isnt a member in the EU, so if I wanted to get a title against you personally, I couldnt get it because of a practical barrier, otherwise you wouldnt talk so lightly and irresponsible against me and about the whole topic. What is also of interest is this: you claim to be this or that, you are a player this or that, but I cant check if you are really authentically this person. Of course the mentioned names might really and authentically exist. Know what I mean? Often people say, but look into the phonebook there in that country. But how could I know that you who is actually writing there is the one out of the phonebook. How this could be researched? Of course it's impossible to do. Unless in a Western country the similar laws are existing and then I could make a formal appeal. But in Russia or Serbia I couldnt."

After such nonsense... I am amazed.

Kind regards!

BS
You simply did not know Rolf. He knows everything and is always right.
Hi Geno,

Sometimes he is right. But in this case I have my doubts. His basic idea is: if you use other people´s ideas, and make the best engine, then you are right, but, if you don´t succeed in making the best engine, then you are wrong, then you are cloner.

A tipical western reasoning: a winner is allowed to make dubious things, others are not.

This is called: double standards.
But this is not what I had said. That being weaker means you are a clone and wrong. I said something else.
Rolf, read again your words:

"Alone your claim or wish that the same standards should apply to Rajlich and Osipov is wrong. As up to this very moment Vas is a sober programmer who has created number one! While Osipov is nobody. "
"Well, I´m just a soul whose intentions are good,
Oh Lord, please don´t let me be misunderstood."

Uri Blass
Posts: 8828
Joined: Wed Mar 08, 2006 11:37 pm
Location: Tel-Aviv Israel

Re: Is Belka a Rybka?

Post by Uri Blass » Mon Dec 03, 2007 9:12 am

As far as I see the piece square table for bishops in strelka is only based on the file of the bishop and the rank of the bishop when there is a bonus for bishop that is in one of the long diagnol and I can see the same idea in fruit.

I can also see that the array BishopLine with exactly the same numbers as fruit can be used to build strelka bishop square table.

Uri

mjlef
Posts: 1466
Joined: Thu Mar 30, 2006 12:08 pm
Contact:

Re: Is Belka a Rybka?

Post by mjlef » Mon Dec 03, 2007 11:37 am

Uri Blass wrote:As far as I see the piece square table for bishops in strelka is only based on the file of the bishop and the rank of the bishop when there is a bonus for bishop that is in one of the long diagnol and I can see the same idea in fruit.

I can also see that the array BishopLine with exactly the same numbers as fruit can be used to build strelka bishop square table.

Uri
That is what I expected. I think you will find the same things with RookLine and other things as well. When I tried putting in values in Belka for RookLine, it also seemed to be a perfect match with the Fruit values. If you want, I can do a spreadsheet, and fill in the values from Fruit, and make it easier for you to compare them. It just goes to show a lot of people borrow ideas from other programs.

I suspect some of the other programmers who looked at Strelka would have found a lot of similarities to Fruit, once they get past the hurdle of the forma things like piece square tables are in. Fruit creates them dynamically from just a few numbers, but Strelka probably just has them precomputed. All of this would be a lot easier to see if I actually had the Strelka code to look at, of course! But it is rather fun feeding in numbers and see when they match.

Mark

Uri Blass
Posts: 8828
Joined: Wed Mar 08, 2006 11:37 pm
Location: Tel-Aviv Israel

Re: Is Belka a Rybka?

Post by Uri Blass » Mon Dec 03, 2007 3:40 pm

mjlef wrote:
Uri Blass wrote:As far as I see the piece square table for bishops in strelka is only based on the file of the bishop and the rank of the bishop when there is a bonus for bishop that is in one of the long diagnol and I can see the same idea in fruit.

I can also see that the array BishopLine with exactly the same numbers as fruit can be used to build strelka bishop square table.

Uri
That is what I expected. I think you will find the same things with RookLine and other things as well. When I tried putting in values in Belka for RookLine, it also seemed to be a perfect match with the Fruit values. If you want, I can do a spreadsheet, and fill in the values from Fruit, and make it easier for you to compare them. It just goes to show a lot of people borrow ideas from other programs.

I suspect some of the other programmers who looked at Strelka would have found a lot of similarities to Fruit, once they get past the hurdle of the forma things like piece square tables are in. Fruit creates them dynamically from just a few numbers, but Strelka probably just has them precomputed. All of this would be a lot easier to see if I actually had the Strelka code to look at, of course! But it is rather fun feeding in numbers and see when they match.

Mark
I find that the array are perfect match but the constants that are used in fruit BishopCentreOpening and BishopCentreEndgame have not the same proportion.

BishopCentreOpening=2
BishopCentreEndGame=3
BishopBackRankOpening=10
BishopDiagonalOpening=4 are from fruit2.1

Strelka seems to use different values that have not the same proportions.

I wonder if it is a good idea to have bishop square table that encourages bishop to be in the main diagnol in the opening and not in the endgame.

Did people test it?
Did they find that it is better relative to one piece square table for the bishop for all the game?

Uri

Orlov

Re: Is Belka a Rybka?

Post by Orlov » Mon Dec 03, 2007 8:01 pm

If Strelka/Belka is clone Rybka 1.0 and Strelka/Belka have(based) on fruit's ideas(not code) then Rybka 1.0 also based on fruit's ideas(not 30 points from Fruit source). Very strong argument from Korshunov. Bravo Fabien!!! :D

Uri Blass
Posts: 8828
Joined: Wed Mar 08, 2006 11:37 pm
Location: Tel-Aviv Israel

Re: Is Belka a Rybka?

Post by Uri Blass » Mon Dec 03, 2007 8:14 pm

Orlov wrote:If Strelka/Belka is clone Rybka 1.0 and Strelka/Belka have(based) on fruit's ideas(not code) then Rybka 1.0 also based on fruit's ideas(not 30 points from Fruit source). Very strong argument from Korshunov. Bravo Fabien!!! :D
The piece square table of knights and bishops of strelka is based on the same idea as fruit but with different weights.

I am going to investigate the other parts of the piece square table of strelka later.

I wonder if other programs before fruit had bonus for bishops in the long diagnol in their piece square table or if it is an original idea of fabien.

I never tried it and I was interested in the mobility parts of fruit and not in the piece square table of fruit so only after this discussion I know that fruit has bonus for bishop in the long diagnol.

Note also that the bonus for bishop in one of the long diagnol is only for the opening stage.

I have 2 different piece square tables for the king in the endgame but I do not have 2 different piece square tables for most pieces because I did not think that there is a big difference between values of pieces like bishop between opening and endgame.

Uri

User avatar
Rolf
Posts: 6081
Joined: Fri Mar 10, 2006 10:14 pm
Location: Munster, Nuremberg, Princeton

Re: Is Belka a Rybka?

Post by Rolf » Mon Dec 03, 2007 8:51 pm

slobo wrote:
Rolf wrote:
slobo wrote:
GenoM wrote:
matejst wrote:Quote: "The same with you. Serbia isnt a member in the EU, so if I wanted to get a title against you personally, I couldnt get it because of a practical barrier, otherwise you wouldnt talk so lightly and irresponsible against me and about the whole topic. What is also of interest is this: you claim to be this or that, you are a player this or that, but I cant check if you are really authentically this person. Of course the mentioned names might really and authentically exist. Know what I mean? Often people say, but look into the phonebook there in that country. But how could I know that you who is actually writing there is the one out of the phonebook. How this could be researched? Of course it's impossible to do. Unless in a Western country the similar laws are existing and then I could make a formal appeal. But in Russia or Serbia I couldnt."

After such nonsense... I am amazed.

Kind regards!

BS
You simply did not know Rolf. He knows everything and is always right.
Hi Geno,

Sometimes he is right. But in this case I have my doubts. His basic idea is: if you use other people´s ideas, and make the best engine, then you are right, but, if you don´t succeed in making the best engine, then you are wrong, then you are cloner.

A tipical western reasoning: a winner is allowed to make dubious things, others are not.

This is called: double standards.
But this is not what I had said. That being weaker means you are a clone and wrong. I said something else.
Rolf, read again your words:

"Alone your claim or wish that the same standards should apply to Rajlich and Osipov is wrong. As up to this very moment Vas is a sober programmer who has created number one! While Osipov is nobody. "

Slobo, where is your reading problem coming from so suddenly. I know you as a careful reader. I wrote Osipov is nobody - NOT because he stays below Vas with his program but because for me he's nobody because he never stepped on stage in real life tournaments like Vas. It doesnt help to claim that people could "find" Osipov via email. Because an email address isnt already the authenticity of a real human being just like we all know and like Vas after his numerous appearances in real life. So, for that difference between these twqo names, I conclude that for this actual moment we cant compare the two. I do NOT say that for this or that Osipov is nobody. He's nobody up to now because nobody knows him in real. Well, I've read someone claimed he has phoned Osipov. Fine! But this would be called hearsay in every court trial. Is this clarified for now, Slobo? I do NOT defame someone just because he comes from Russia or elsewhere. I accuse him that he defames Vas on the background of his (Osipov's) own hiding. Under this perspective it's dishonest what he does and his motivation cant be called decent. However if he would step forward not only with his talking about his "program" and if he were real and feasable then this argument of mine would vanish of course. Know what I mean?

Let me finally add something about Vas performance. Let's say he had taken ideas from Fruit. Ok, but his program is stronger than Fruit. Now let's assume Osipov took something from Rybka! But his Strelka isnt stronger than Rybka, it's weaker. Doesnt this tell you something about the qualities of the two programmers?? The same now with the cooperation of two or three authors. And Bjelka. Is it stronger than Rybka? No, of course not. So what is this all about?? Where is the own record of these Russians? If they remain weaker than Rybka? Why should we estimate this sort of copying for good if the result is worse than the original?? By no thinkable interpretations Vas has done the same with Fruit because his engine is way stronger than Fruit. So that Vas could be called a genuine and stronger programmer than Fabien. Osipov however has created nothing new up to this very moment. Period.
-Popper and Lakatos are good but I'm stuck on Leibowitz

Post Reply