the end of the strelka affair

[Orlov]

I should note that your are just making a cliam here, and then repeating it. this does not constitute evidence. Indeed, some of the things that have been said about vas and rybka in this thread border on slander. The claims seem to be unsupported by evidence but mostly meant to hurt him and his business.

It does not matter how many people sarcastly repeat fruit = rybka. This is not evidence.


Vas has been devoted to chess (as an international master) and programming for quite some time, and deserves some benefit of the doubt, until their is evidence. I can say from a chess playing perspective that rybka does not play much like fruit and right from the begining killed fruit 2.2. (much better than the version vas was said to be stealing from). Rybka right from the beginning played like no program i'd ever seen. that is the only evidence i have. It is not iron clad, but it is something. Strelka, in contrast, played just like rybka, even down to the odd errors.

In my view, the current evidence points to strelka being a clone of rybka, and to rybka being something distinctive and new. This is not total proof, but it does fall into the category of evidence i think. It is the best we have.

best

Joseph

I not against Rybka, but I also not against Strelka. And if you speak that Strelka a clone I approve, that it precisely same clone as well as Rybka. They simply have one parent.

No this is untrue. Because otherwise O didnt hide like a crook! That alone is the proof for the evil he has done and he has some support now.

For development computer chess Osipov has made much more than Vas, and it is absolutely free-of-charge.

It is hard to compare because Osipov could not do what he did without Vas.

Uri

http://f23.parsimony.net/forum50826/messages/177397.htm

[Uri Blass]

The author saying it is based on fruit does not in fact make it "fact"


Didn't people already compare fruit and strelka and say that strelka was not a fruit clone. These were top experts in the field. Plus, I think uri has looked at fruit code and he has had access to the strelka code, well before alot of people and is using it to inform the developmennt of his engine. I assume he would be able to say whether or not it is fruit.

Anthony C and others whom I assume are familiar with fruit code have been looking at the strelka code (see rybka forum) and marvelling at some of rybykas/strelks insights and simplicities. why would anybody "marvell" if it was somehtng they had already been familiar with for years.

This is so simple. Unbiased Programing experts can compare the code and say whether it is a clone of fruit (and this, as i recall, has been done)

best
J

I remember history with Fafis, were was 30% code from Crafty. Verdict was - clone. I think Strelka have 30% code from fruit, so what? Not clone

I think strelka has not 30% of fruit.
If my memory is correct the claim was about old DanChess and not about Fafis.

I remember that Fafis was discovered to be a clone but I do not remember people seeing the source of it.

Uri

Strelka have more than 30% fruit's code, some parts are very slightly changed.

some parts are not more than 30%
I think that the structures are the main thing.

If you use the same evaluation with bitboards instead of another structure then I consider the code to be mainly different.

There is some similiarity in the hash code but not enough to get 30%

Uri

[Orlov]

The author saying it is based on fruit does not in fact make it "fact"


Didn't people already compare fruit and strelka and say that strelka was not a fruit clone. These were top experts in the field. Plus, I think uri has looked at fruit code and he has had access to the strelka code, well before alot of people and is using it to inform the developmennt of his engine. I assume he would be able to say whether or not it is fruit.

Anthony C and others whom I assume are familiar with fruit code have been looking at the strelka code (see rybka forum) and marvelling at some of rybykas/strelks insights and simplicities. why would anybody "marvell" if it was somehtng they had already been familiar with for years.

This is so simple. Unbiased Programing experts can compare the code and say whether it is a clone of fruit (and this, as i recall, has been done)

best
J

I remember history with Fafis, were was 30% code from Crafty. Verdict was - clone. I think Strelka have 30% code from fruit, so what? Not clone

I think strelka has not 30% of fruit.
If my memory is correct the claim was about old DanChess and not about Fafis.

I remember that Fafis was discovered to be a clone but I do not remember people seeing the source of it.

Uri

Strelka have more than 30% fruit's code, some parts are very slightly changed.

some parts are not more than 30%
I think that the structures are the main thing.

If you use the same evaluation with bitboards instead of another structure then I consider the code to be mainly different.

There is some similiarity in the hash code but not enough to get 30%

Uri

uci code + hash code + const + defines + variables unless it is not enough of it?

[Uri Blass]

The author saying it is based on fruit does not in fact make it "fact"


Didn't people already compare fruit and strelka and say that strelka was not a fruit clone. These were top experts in the field. Plus, I think uri has looked at fruit code and he has had access to the strelka code, well before alot of people and is using it to inform the developmennt of his engine. I assume he would be able to say whether or not it is fruit.

Anthony C and others whom I assume are familiar with fruit code have been looking at the strelka code (see rybka forum) and marvelling at some of rybykas/strelks insights and simplicities. why would anybody "marvell" if it was somehtng they had already been familiar with for years.

This is so simple. Unbiased Programing experts can compare the code and say whether it is a clone of fruit (and this, as i recall, has been done)

best
J

I remember history with Fafis, were was 30% code from Crafty. Verdict was - clone. I think Strelka have 30% code from fruit, so what? Not clone

I think strelka has not 30% of fruit.
If my memory is correct the claim was about old DanChess and not about Fafis.

I remember that Fafis was discovered to be a clone but I do not remember people seeing the source of it.

Uri

Strelka have more than 30% fruit's code, some parts are very slightly changed.

some parts are not more than 30%
I think that the structures are the main thing.

If you use the same evaluation with bitboards instead of another structure then I consider the code to be mainly different.

There is some similiarity in the hash code but not enough to get 30%

Uri

uci code + hash code + const + defines + variables unless it is not enough of it?

uci code is not about the moves the engine plays so I even did not look at it.
const+some defines+some variables are not the heart of the program and the heart of the program is the move generator and the search code and the evaluation code.

I looked at the question if the functions are basically the same and I think that they usually are not the same(using ideas is allowed so the fact that they are functions that do similiar things prove nothing).

I wonder what is the opinion of Fabien about it.

Uri

[Orlov]

The author saying it is based on fruit does not in fact make it "fact"


Didn't people already compare fruit and strelka and say that strelka was not a fruit clone. These were top experts in the field. Plus, I think uri has looked at fruit code and he has had access to the strelka code, well before alot of people and is using it to inform the developmennt of his engine. I assume he would be able to say whether or not it is fruit.

Anthony C and others whom I assume are familiar with fruit code have been looking at the strelka code (see rybka forum) and marvelling at some of rybykas/strelks insights and simplicities. why would anybody "marvell" if it was somehtng they had already been familiar with for years.

This is so simple. Unbiased Programing experts can compare the code and say whether it is a clone of fruit (and this, as i recall, has been done)

best
J

I remember history with Fafis, were was 30% code from Crafty. Verdict was - clone. I think Strelka have 30% code from fruit, so what? Not clone

I think strelka has not 30% of fruit.
If my memory is correct the claim was about old DanChess and not about Fafis.

I remember that Fafis was discovered to be a clone but I do not remember people seeing the source of it.

Uri

Strelka have more than 30% fruit's code, some parts are very slightly changed.

some parts are not more than 30%
I think that the structures are the main thing.

If you use the same evaluation with bitboards instead of another structure then I consider the code to be mainly different.

There is some similiarity in the hash code but not enough to get 30%

Uri

uci code + hash code + const + defines + variables unless it is not enough of it?

uci code is not about the moves the engine plays so I even did not look at it.
const+some defines+some variables are not the heart of the program and the heart of the program is the move generator and the search code and the evaluation code.

I looked at the question if the functions are basically the same and I think that they usually are not the same(using ideas is allowed so the fact that they are functions that do similiar things prove nothing).

I wonder what is the opinion of Fabien about it.

Uri

every ORIGINAL egine MUST to have uniques const, variables, defines.

[Uri Blass]

The author saying it is based on fruit does not in fact make it "fact"


Didn't people already compare fruit and strelka and say that strelka was not a fruit clone. These were top experts in the field. Plus, I think uri has looked at fruit code and he has had access to the strelka code, well before alot of people and is using it to inform the developmennt of his engine. I assume he would be able to say whether or not it is fruit.

Anthony C and others whom I assume are familiar with fruit code have been looking at the strelka code (see rybka forum) and marvelling at some of rybykas/strelks insights and simplicities. why would anybody "marvell" if it was somehtng they had already been familiar with for years.

This is so simple. Unbiased Programing experts can compare the code and say whether it is a clone of fruit (and this, as i recall, has been done)

best
J

I remember history with Fafis, were was 30% code from Crafty. Verdict was - clone. I think Strelka have 30% code from fruit, so what? Not clone

I think strelka has not 30% of fruit.
If my memory is correct the claim was about old DanChess and not about Fafis.

I remember that Fafis was discovered to be a clone but I do not remember people seeing the source of it.

Uri

Strelka have more than 30% fruit's code, some parts are very slightly changed.

some parts are not more than 30%
I think that the structures are the main thing.

If you use the same evaluation with bitboards instead of another structure then I consider the code to be mainly different.

There is some similiarity in the hash code but not enough to get 30%

Uri

uci code + hash code + const + defines + variables unless it is not enough of it?

uci code is not about the moves the engine plays so I even did not look at it.
const+some defines+some variables are not the heart of the program and the heart of the program is the move generator and the search code and the evaluation code.

I looked at the question if the functions are basically the same and I think that they usually are not the same(using ideas is allowed so the fact that they are functions that do similiar things prove nothing).

I wonder what is the opinion of Fabien about it.

Uri

every ORIGINAL egine MUST to have uniques const, variables, defines.

Original engines can have non original parts.

I do not claim that strelka is original but I think that the main reason is rybka and not fruit.

Uri

[Philippe]

I am more interested in strelka strength. Has anyone alredy made some tournaments vs rybka and if so what are the results ?

[gerold]

I am more interested in strelka strength. Has anyone alredy made some tournaments vs rybka and if so what are the results ?

Ryaka 22/40
Strelka 18/40

[Ovyron]

I am more interested in strelka strength. Has anyone alredy made some tournaments vs rybka and if so what are the results ?

Ryaka 22/40
Strelka 18/40

What's Ryaka? Another clone?

[gerold]

I am more interested in strelka strength. Has anyone alredy made some tournaments vs rybka and if so what are the results ?

Ryaka 22/40
Strelka 18/40

What's Ryaka? Another clone?

Now Rybka V Strelka 24-18