ChessVibes publishes a balanced article re: Rybka

[Rebel]

Riis concludes by comparing UCI parameters, noting that Rybka lacks about
18 of 20 parameters possessed by Fruit (and adds a SearchDirection parameter).
Again this point seems bizarre; a minimally competent programmer should take
no more than 15 minutes to excise such parameters from the Fruit code (making
them their default values, if nothing else).

BB+ uses the term bizarre, but I'd say propaganda, because that's exactly what it is. Given that the ChessBase's opinion piece references C.W. and E.S. as consultants, whom I would like to believe are competent programmers, there's only one interpretation to that UCI parameter comparison -- it is meant to deceive and persuade non-programmers. A picture is worth a thousand words...I dig.

The UCI parameter example is just one of the 20 differences between Fruit and Rybka of which 10+ in EVAL. I would say that's not propaganda especially not if BB+ himself likes to argue this way, see: http://www.top-5000.nl/evidence.htm#C7

[Rebel]

They commented on both sides of the debate, contacted representatives of both sides of the debate for comment and republished unskewed summaries of the arguments (noting their own role). Concluded that it's complicated, implied that they haven't got the technical expertise to sort it out themselves, and resisted the temptation to choose sides. Simple, informative, unsensational, honest.

In short, it's the sort of article which could get a few more people to actually read the evidence, compare the statements and claims of both sides and come to some informed conclusions, rather than relying on the frothy chaos that has characterized the discussion to date. I tend to think that the evidence speaks for itself, and that the more people who take the time to understand it, the better.

jb

+1

[Rebel]

There is a culture, especially in the USA, which values believing your gut over the evidence of your senses. "Better strong and wrong than right and weak". Unfortunately, this culture has collided with the ICGA's academic sensibilities in a pretty ugly way, which can be best seen on the Rybka forum, but has been a general feature of this debate from the start.

I am sorry but I feel the same way about the ICGA hence I left.

This debate for a long time is not about truth-finding any longer but about WINNING the debate only.

Examples, Bob, Harvey, Levy.

The shining ICGA exception is Mark Lefler.

[Jeremy Bernstein]

There is a culture, especially in the USA, which values believing your gut over the evidence of your senses. "Better strong and wrong than right and weak". Unfortunately, this culture has collided with the ICGA's academic sensibilities in a pretty ugly way, which can be best seen on the Rybka forum, but has been a general feature of this debate from the start.

I am sorry but I feel the same way about the ICGA hence I left.

This debate for a long time is not about truth-finding any longer but about WINNING the debate only.

Examples, Bob, Harvey, Levy.

The shining ICGA exception is Mark Lefler.

Ed, any claim you had to desiring to find the truth was nullified by the deliberately false and misleading "data" that you published on your web site. You are a dissembler, preying on a group of technical novices who want so badly to believe that the ICGA's evidence is false (or falsified) that they'll buy whatever you're selling, as long as it confirms their point of view. You should be ashamed of yourself, but, hey, congratulations on getting your name back in the press.

jb

[Rebel]

There is a culture, especially in the USA, which values believing your gut over the evidence of your senses. "Better strong and wrong than right and weak". Unfortunately, this culture has collided with the ICGA's academic sensibilities in a pretty ugly way, which can be best seen on the Rybka forum, but has been a general feature of this debate from the start.

I am sorry but I feel the same way about the ICGA hence I left.

This debate for a long time is not about truth-finding any longer but about WINNING the debate only.

Examples, Bob, Harvey, Levy.

The shining ICGA exception is Mark Lefler.

Ed, any claim you had to desiring to find the truth was nullified by the deliberately false and misleading "data" that you published on your web site. You are a dissembler, preying on a group of technical novices who want so badly to believe that the ICGA's evidence is false (or falsified) that they'll buy whatever you're selling, as long as it confirms their point of view. You should be ashamed of yourself, but, hey, congratulations on getting your name back in the press.

jb

I see you joined the name-calling and mind-reading club.

Bye Jeremy.

[Hood]

Soren Riss is biased and unprofessional in a theme. Why so much rumour?

[Judoka]

Articles are always bias,... always! If anyone feels it is unbiased that is only because it has the same bias as you!

What has interested me in that I am finally hearing from 3rd party programmers that have no axe to grind.
FINALLY people that haven't chosen a side in this civil war.

http://www.chessvibes.com/reports/contr ... ment-68471

Stephen

I am not an interested party in any of this, I am not a chess programmer although I am a commercial programmer.

From what I understand the disqualification was effected several years after the event and retrospectively. Was there or is there no statute of limitations for these tournaments ? If there was then the programmer has no excuse for not keeping his source until that period expired. If there was no such period then it hardly seems reasonable to expect him to keep his source indefinitely unless the rules explicitly say so.

It seems to me that without the actual source code there is no proof of wrongdoing. The best that can be done is to make inferences based on the compiled object code. Although statistically you may prove some similarities I would think that it would be a long way from being actual proof.

As such, in my humble opinion, he may be guilty in fact, but as there is a reasonable doubt i.e. no way to definitely prove it, his innocence must be assumed. Not to do so, is contrary to most people's idea of the way justice works. The fact that he has been found guilty by his peers doesn't change that in my opinion.

It's sad that even something as irrelevant and trivial to the world at large as chess programming can cause such bitter feuding between people that were once friends and colleagues, I know that money is involved etc. but really it does seem rather childish doesn't it ?

Andrew Dalke

I found out about this controversy from a recent posting to the forum site "Hacker News." As a software developer interested in issues related to software copyright, I read the various analyzes with interest.

As I don't know anything about chess evaluation engines, I looked at the parsing code mentioned in http://www.chessvibes.com/plaatjes/rybk ... _Mar11.pdf . The dissassembled code from Rybka is quite clearly different than the equivalent code in Fish, at least from the point of copyright.

The strstr functions take different parameters ("moves" instead of "moves "), the fen parser function takes one parameter instead of two, and in Rybka the fen parser code is alway called with the "startpos" board and then if there is a fen string it's called again, while the Fruit code does one case or the other. Interestingly, this difference is removed from the quoted Fruit code, and its omission has the consequence of highlighting the similarities.

The author says the evidence of copying is not in the actual working code, but is the moves[-1] = '\0' code, which does nothing in Fruit and so should not be needed in Rybka. I verified that in Fruit, but I also know that that idiom is useful in detecting bugs early on in parser development, so it could be a remnant of some protection code.

In any case, the author didn't show that 1) the added NUL does nothing in Rybka, nor 2) that Rybka doesn't have a pattern of using that idiom elsewhere. Interestingly, Fruit does use that idiom in one other spot, so a proper analysis would check to see if Rybka also uses that idiom there, and only there.

In conclusion, this one point of the evidence for copying, which seems to be part of the ancillary evidence to strengthen the main hypothesis, is quite fundamentally flawed. It shows that the two programs use different algorithms, and the analysis does not provide evidence that the method to detect copying is itself valid. If this is characteristic of type of analysis which went into this claim of rule violation, then no wonder there is such an uproar - even if the claim is true, the methods to detect that claim are suspect.

[BB+]

It seems to me that without the actual source code there is no proof of wrongdoing. The best that can be done is to make inferences based on the compiled object code. Although statistically you may prove some similarities I would think that it would be a long way from being actual proof.

As such, in my humble opinion, he may be guilty in fact, but as there is a reasonable doubt i.e. no way to definitely prove it, his innocence must be assumed. Not to do so, is contrary to most people's idea of the way justice works. The fact that he has been found guilty by his peers doesn't change that in my opinion.

All of these aspects were discussed in my rebuttal. As for the idea of the way justice works, when a person chooses not to defend himself against voluminous evidence, he is typically declared guilty. No?

As I don't know anything about chess evaluation engines, I looked at the parsing code mentioned in [RYBKA_FRUIT]. The dissassembled code from Rybka is quite clearly different than the equivalent code in Fish, at least from the point of copyright.

The commenter (Andrew Dalke) is indeed quite adept at disassembly. However, the document to which he refers (RYBKA_FRUIT) was a preliminary (Mar 11) enumeration of (all) possible evidence [therein it indicates that some of this comes from as-of-yet unchecked sources], and one can note the UCI parsing elements that he mentions were dropped in (say) my recent Recapitulation (and appeared nowhere in the Report, was never much discussed in the Panel, etc.). I agree largely with his sentiments regarding the questionable nature of this evidence (it was removed, after all). I do not agree that this is characteristic of the whole.

[BB+]

I agree largely with his sentiments regarding the questionable nature of this evidence (it was removed, after all).

One can expect this essentially to become a Rorschach test. Those on one side will say that it shows that the investigation did indeed take pains to look hard at the evidence, rather than just accept everything that was "thrown together". Those on the other side will say that the fact this "evidence" was even included at a preliminary phase shows that the whole thing was shambolic from the start.

[kingliveson]

Riis concludes by comparing UCI parameters, noting that Rybka lacks about
18 of 20 parameters possessed by Fruit (and adds a SearchDirection parameter).
Again this point seems bizarre; a minimally competent programmer should take
no more than 15 minutes to excise such parameters from the Fruit code (making
them their default values, if nothing else).

BB+ uses the term bizarre, but I'd say propaganda, because that's exactly what it is. Given that the ChessBase's opinion piece references C.W. and E.S. as consultants, whom I would like to believe are competent programmers, there's only one interpretation to that UCI parameter comparison -- it is meant to deceive and persuade non-programmers. A picture is worth a thousand words...I dig.

The UCI parameter example is just one of the 20 differences between Fruit and Rybka of which 10+ in EVAL. I would say that's not propaganda especially not if BB+ himself likes to argue this way, see: http://www.top-5000.nl/evidence.htm#C7

Come on Ed, you know from a technical stand point that this was quite silly. It has already been demonstrated how easy it is to modify Fruit's UCI parameters to mirror that of Rybka's.