To Jeroen and interested minds, re. Tiger node count

[Rolf]

Well, I am a book expert, so the only thing I hide is strong opening lines.

So, good luck with your personal warfare, backed up by the mods here!

You must be joking. Mods are not backing any personal warfare.

Then what is with Czub? He seems to have just one single fixed idea. The same actually with CT although he does practice what he attacks others for. How can someone with a clone past defame here Vas, a decent gentleman? Arent you agreeing with George who explained in detail what should be done if people had proof and evidence. But a campaign like this during the starting sale of Rybka is wrong IMO.

[tiger]

(2) the rybka/fruit/strelka issue is an interesting question because of GPL issues. He pointed out something I had not considered previously, namely that strelka/fruit have lots of similarities, and then Vas claimed that strelka was so much like rybka 1 that he was claiming it as his own code. Once you digest both of those points, there could be an issue one might have to deal with at some point down the road. The GPL is a legal document that can't be waved away.
.

The Strelka source and the Fruit source has been made public, right? I was just wondering ... Has anyone done detailed comparison, and posted the results? What percentage of Strelka and Fruit are identical, or very similar.

Regards
Dave

It's coming.



// Christophe

[Rolf]

(2) the rybka/fruit/strelka issue is an interesting question because of GPL issues. He pointed out something I had not considered previously, namely that strelka/fruit have lots of similarities, and then Vas claimed that strelka was so much like rybka 1 that he was claiming it as his own code. Once you digest both of those points, there could be an issue one might have to deal with at some point down the road. The GPL is a legal document that can't be waved away.
.

The Strelka source and the Fruit source has been made public, right? I was just wondering ... Has anyone done detailed comparison, and posted the results? What percentage of Strelka and Fruit are identical, or very similar.

Regards
Dave

It's coming.



// Christophe

ChrisW, also here, what is with you verdict that this Fabien' business and if he didnt see a problem then what is the business of Theron in this relation? Isnt it a warfare? As JN said?

[mclane]

can it be that you mix email with posting in a computer chess forum?
why don't you chat with chris on email so that we others are not bored by the content ?

[tiger]

Jeroen Noomen in another thread asked me about the node count displayed by Chess Tiger versions 12 to 15.

The thread in question has been partially deleted by the moderators before I had enough time to answer Jeroen's question, so I just start over here.

Historically, Jeroen has been Chess Tiger's book author for the aforementionned version of Chess Tiger, which is the chess engine I have written.

Jeroen is now working for the Rybka team. I have asked him if he could write another opening book for Chess Tiger some time ago, but as I understand he cannot do so anymore because of his exclusive links with the Rybka team.

Jeroen is a fine person and unlike (unfortunately) most people with whom I have been involved in computer chess, including many people posting here, I have meet him in person several times.

OK, I hope you don't mind all these boring historical details, but in view of the recent events here at CCC I thought it would be good to remind that we are real humans and that there are various kind of links between us. Also it is possible that some people participating in the recent discussions did not know much about Jeroen or about me.

So, I was about to write a very precise explanation about the node count in these versions of Chess Tiger, which include some revelations about an algorithm used in Chess Tiger that I have never seen described anywhere else. Describing this algorithm is absolutely necessary in order to answer Jeroen's question.

However now Jeroen is working for the Rybka team. So I was going to answer a question from the Rybka team, but on the other hand this team has never answered this very same question, which has been repeatedly asked since several months, even years.

I do not want to turn this into a childish game of "I asked you first so you answer first and then I answer".

So here is my proposition: I immediately answer Jeroen's question as precisely as possible (which involves revealing a proprietary algorithm used in Chess Tiger) if Vas answers as honestly and as precisely as possible the very same question, which is about Rybka 1.0 (no need to reveal anything about the more recent versions). As we have most of the source code of Rybka 1.0 already, his explanations can refer to this code, so they can be undiscutable.

I hope this can be considered as acceptable.

// Christophe

Just wondering...

is this your nodecode "obfuscation" related to the paper I sent you a couple of years ago ? Or is it supersoma related ?

Cheers,

Tony

There is no obfuscation. To understand the node count of these versions, one has to understand how Tiger searches the tree and evaluates.

Jeroen does not seem to have the whole story and apparently he thought that there was something fishy about it. I guess the line of thought was that I would immediately keep quiet about the obfuscated node count in Rybka if hard pressed to talk about the node count in Tiger.

Unfortunately for him and the Rybka team I would be very glad to explain how nodes are counted in these versions and I would at the same time reveal a nice search/evaluation algorithm in Chess Tiger. On the other hand Rajlich sees no joy in admitting the obfuscation in Rybka because it is undefendable. The only defense is to admit it has been obfuscated on purpose, and BTW is still obfuscated, as it seems, in the latest versions of Rybka.

You know, I'm not obsessed by the node count. I already said it took me 5 minutes to figure it was fake, so I just stopped considering it had any value. But Jeroen came and fueled that debate again.



// Christophe

[chrisw]

(2) the rybka/fruit/strelka issue is an interesting question because of GPL issues. He pointed out something I had not considered previously, namely that strelka/fruit have lots of similarities, and then Vas claimed that strelka was so much like rybka 1 that he was claiming it as his own code. Once you digest both of those points, there could be an issue one might have to deal with at some point down the road. The GPL is a legal document that can't be waved away.
.

The Strelka source and the Fruit source has been made public, right? I was just wondering ... Has anyone done detailed comparison, and posted the results? What percentage of Strelka and Fruit are identical, or very similar.

Regards
Dave

It's coming.



// Christophe

ChrisW, also here, what is with you verdict that this Fabien' business and if he didnt see a problem then what is the business of Theron in this relation? Isnt it a warfare? As JN said?

Judicially it is entirely Fabien's business it seems to me. Legal situation seems to me that Vas/Rybka is entirely fine and Christophe and others have no say or control over that, in a sense it is none of their business.

In industry terms I suppose all options are open, ultimately it is political. If Christophe and others can persuade the 'industry' whatever and wherever that is, that use of GPL source as part of development even if all original code fragments are removed is not ok (on the hypothetical assumption that is what hapopened), then I guess he will have achieved his aim. You are countering him on political grounds also. Mods will stand to oneside unless and until there are charter breaches.

Incedentally, I never got a reply from anyone to two questions:

1. these alleged code fragments - are they in the UCI handling or the engine?

2. Is Fruit's UCI handling code all (C) Fabien?

I suspect I got no answer because (1) the code fragments are UCI and (2) nobody is sure about the status of the UCI code. If that's the case the Rybka-clone allegation really does begin to fall away, does it not?

[Zach Wegner]

Incedentally, I never got a reply from anyone to two questions:

1. these alleged code fragments - are they in the UCI handling or the engine?

2. Is Fruit's UCI handling code all (C) Fabien?

I suspect I got no answer because (1) the code fragments are UCI and (2) nobody is sure about the status of the UCI code. If that's the case the Rybka-clone allegation really does begin to fall away, does it not?

I did answer these, but they were deleted. It's still there in the big post Swami made quoting some OK-but-deleted-posts. Here are the answers:

Where are these hypothetical non-rewritten code chunks? The UCI handling or the engine AI? Or both?

Both.

Strelka is actually a pretty good place to look for this sort of stuff. It helps to piece together the Rybka code. Did you know that Rybka has the pawn value 3399? Or that Rybka and Fruit use the same easy-move logic with the same margin?

Second question. What is the licence situation for the UCI material? Is that entirely Fabien's work, or did he get all or part of it from someplace else? Where do people get the engine side UCI stuff from?

The UCI protocol is free to use. I suppose that would be considered public domain? I don't know the legal aspect of it. The code to implement it in Fruit however, is copyrighted by Fabien. The engine side of UCI is implemented by reading the protocol specification (at least that's how I did it). I suppose some people might look at another implementation when writing theirs, especially if they are new to chess programming, but the Rybka/Fruit ones are just way too similar IMO.

[bob]

I have only seen two points CT has made.

(1) node counts were obfuscated in Rybka, and of that I (and many others) are absolutely certain. The inane explanations that have been written are simply attempts to get out of a situation one would rather have not gotten into in the first place. The explanations/excuses make no sense, particularly in light of the data that has been presented in the past where the total node count after iteration N+1 is _smaller_ than the node count for iteration N. How can you "unsearch" some nodes? So that is all a crock, and I agree with the conclusions I posted.

(2) the rybka/fruit/strelka issue is an interesting question because of GPL issues. He pointed out something I had not considered previously, namely that strelka/fruit have lots of similarities, and then Vas claimed that strelka was so much like rybka 1 that he was claiming it as his own code. Once you digest both of those points, there could be an issue one might have to deal with at some point down the road. The GPL is a legal document that can't be waved away.

I haven't seen any attempt to 'defame' anyone, nor any personal attacks tossed around. He has asked legitimate questions, and part of the answers are garbage (answers regarding node count / depth / PV obfuscation) and part are simply not provided.

This is most likely an issue that is not going away, until the questions are answered. Which means it is an issue that is just not going away.

Bob, I hope you will agree. If there are people who have watertight proof for evidence then these should begin a court trial. But if you dont do that, if you are not absolutely sure, then you must stop this here in CCC. It cannot be that one CT has here a sort of character assassination for free every Mondays. grolich on Rybka Forum has given enough arguments addressing all arguments also GLP and also George has made nice analyses on the Rybkaforum. Fact is what is going on is sort of bashing with the hope that Vas appeared and clarified. But this wont happen this way. I already made clear what I expect. And it's after EU law, not American to be decided. If the origine of Rybka is doubted then all commercial engines should be doubted. Your science view is clear, but you cant enforce it, in special in case of commercial programs. The presentation of CT is not sound IMO, your conclusion that Vas was caught red-handed isnt conclusive IMO. The campaign should be stopped now. I hope that in Peking/Beijing the events of Graz wont be repeated but I wouldnt trust that organisation with JvdH etc. Also for science reasons. I cant believe that someone gets a PhD title with a report about making love to a computer. If all would come back on ethical norms in computerchess such campaigns here shouldnt be possible in future.

In a different context I discussed the policy in such fora, when someone suddenly surprised me with the statementg that debates with chessplayers are somehow always vicious and never-ending. He said, chessplayers like to quarrel. Blood left my face and I knew at the instant that I was one of that kind myself. I was shockeda that I saw this only at that late stage of my life. But then it was my hobby amd the only possible sport I could have chosen.

I'll state my view one more time. The node count obfuscation has _no_ justification, other than the simple fact that it is used to hide internal details. I don't have a problem with the "hiding". I do have a problem with knowingly producing bogus values. If you want to hide something, hide it. Don't disguise it. That's simple enough.

As far as the "clone issue" my only point was that by claiming strelka, those that use the argument to others of "you haven't seen the source for Rybka so you can't possibly conclude anything about it" are simply wrong. Because since Vas claimed Strelka as _his_ code, a clone of Rybka 1, then he has inadvertantly put "his" source code in a public light, where inspection is possible...

I have not drawn any conclusion about whether Rybka is a derivative of Fruit, gnuchess, or sasquatch. I only commented on the fact that strelka is there for anyone to look at, analyze, and draw conclusions that apply to Rybka 1 as well. I've implied nothing more, nothing less. The node counts are bogus, and intentionally so. As far as the clone issue, I have not looked at any of this and don't intend to.

[bob]

(2) the rybka/fruit/strelka issue is an interesting question because of GPL issues. He pointed out something I had not considered previously, namely that strelka/fruit have lots of similarities, and then Vas claimed that strelka was so much like rybka 1 that he was claiming it as his own code. Once you digest both of those points, there could be an issue one might have to deal with at some point down the road. The GPL is a legal document that can't be waved away.
.

The Strelka source and the Fruit source has been made public, right? I was just wondering ... Has anyone done detailed comparison, and posted the results? What percentage of Strelka and Fruit are identical, or very similar.

Regards
Dave

It's coming.



// Christophe

ChrisW, also here, what is with you verdict that this Fabien' business and if he didnt see a problem then what is the business of Theron in this relation? Isnt it a warfare? As JN said?

Rolf, we _all_ have a dog in this hunt. I don't want to play clone after clone in CCT or WCCC events. Never have wanted to, never will want to. So of course it matters. It matters not that I give someone (say Mike or Tracy on the current team) permission to use the current crafty as a starting point for a new program that they can use to participate. That is not a decision _i_ can make and force others to abide by it.

That is what the issue is about.

[Rolf]

I'll state my view one more time. The node count obfuscation has _no_ justification, other than the simple fact that it is used to hide internal details. I don't have a problem with the "hiding". I do have a problem with knowingly producing bogus values. If you want to hide something, hide it. Don't disguise it. That's simple enough.

As far as the "clone issue" my only point was that by claiming strelka, those that use the argument to others of "you haven't seen the source for Rybka so you can't possibly conclude anything about it" are simply wrong. Because since Vas claimed Strelka as _his_ code, a clone of Rybka 1, then he has inadvertantly put "his" source code in a public light, where inspection is possible...

I have not drawn any conclusion about whether Rybka is a derivative of Fruit, gnuchess, or sasquatch. I only commented on the fact that strelka is there for anyone to look at, analyze, and draw conclusions that apply to Rybka 1 as well. I've implied nothing more, nothing less. The node counts are bogus, and intentionally so. As far as the clone issue, I have not looked at any of this and don't intend to.

Is Dann Corbit for you a nobody? He and two others have analysed the dirty copy. Result, read Corbit for yourself, nothing. That was almost 2 years ago. And now it's kosher if a Theron, who has practically left computerchess Tiger, he is reserved a stage here for such a smear campaign. You say he never insulted with bad wording or such. I say the whole campaign is it. Last time all was discussed, now the same procedure as every year. Not kosher IMO. Ok, I think my position has become clear enough. Technically I cant tell anything for obvious reasons.

It looks as if the sales for Rybka 3 should be damaged. Therefore I see the court case coming. Not this forum involved but Europeans because of the EU law. Let's see what CT has to offer.

Bob, this redhanded from you is a problem for me. Say, Vas did it for fun, then it's still correct how you positioned your critic? It's allowed to insinuate someone did something evil, if he didnt do anything at all? If he just wanted to reveil how people get emotionally aroused? In other words, are you 100% sure that you have a case against Vas? And if not you are allowed to make such strong statements? Or is it allowed because Vas obfuscated for no matter what reasons?