FIDE Rules on ICGA - Rybka controversy

[syzygy]

So would the only way for the ICGA to meet any requirement herein be to find a pre-existent outside body to which both sides were agreeable?

I would try to make a clear split between the collection of technical evidence and the evaluation of the evidence. Essentially anyone willing to spit through binaries would be allowed to report his or her findings, but these reports should focus on the facts. Rajlich would be allowed to comment on those. The evaluation of the facts should then be done by a panel of 3 persons with technical expertise, one appointed by the ICGA or whoever wants to play the role of the accuser (Letouzey is indeed an option), one by Rajlich, and the third on agreement of those two.

At that point, it would be hard to construe it as being merely "expert advice" anymore (cf. Rule #2 and EC procedural interpretation therein), rather it would be notably advocatory].

I agree it is not entirely obvious how to follow the "after seeking expert advice" model. The 3-person panel could "advise" the ICGA Board, which would then be practically bound to follow the advice. Or the 3-person panel could play the role of the TD, but then there might not be "expert advice" (and none would be needed, as they are experts themselves).

Advisory bodies in quasi-judicial procedures are usually required to be independent. Same for expert witnesses.

Taken to the absurd, could you then argue that the "seeking expert advice" in Rule #2 should thus mean an independent expert [to the extent considered above]?

I would argue Rule 2 was written for in-tournament application and that there was/is no real provision to deal with the Rybka case.

and don't see why post-tournament this should change to be qualified as "independent expert advice"

I do see why a procedure in 2011 on whether to revise tournament results from 2006, if there must be one at all, should be "fair".

Of course the panel could not help that the ICGA Executive Board even went beyond the facts established by the panel. The panel did not establish any facts regarding Rybka 3, for example.

I think in a Panel discussion it was discussed whether the 2008-10 entries could adequately be modelled by Rybka 3 and 4, and the (only?) opinion on the matter was that this was not yet warranted.

Fair enough. Unfortunately the ICGA Executive Board apparently considered that not deciding on 2008-10 was not an option.

When the ICGA chose to ban Rybka for life due to violating the rules in ICGA events in 2006 and 2007, this required all of Rajlich's programs be removed, since the newer programs would never have been allowed to compete if the violation of earlier versions was discovered sooner and the lifetime ban issued earlier.

I doubt very much that that logic is accepted in analogous contexts. I might come back to this.

I don't remember specifically this refusal being used against him? I don't think the EC complaint specifies anything about this. I could find 3 possible "uses" of his refusal against him.

I meant those in the panel report. (And to some extent the references to Rajlich's refusal to cooperate with the investigation panel made in defense of the ICGA investigation in the aftermath of the whole affair, but those do not matter much for the present discussion.)

If you could show a demonstrative link between the Panel comments (regarding refusal) and anything that later ensued, I would say your view might be more justified. But I would think it more to be a "minor influence" at best.

They are in the panel report, which in my view is enough to make them a problem. As I said, in my view the panel report should not have read like a verdict. I know the Charter sort of asks for a verdict, so I don't think that's correct either.

I didn't know that Ken Thompson was one of his worst enemies.

[/quote]
Did KT spend any significant amount of time on digging into the matter? As far as I know only few were really active.

I would agree finding someone acceptable to both sides and willing to sacrifice a significant amount of his spare time to dig into source code provided by Rajlich could have established the facts (what did Rajlich really do / at what level of abstraction did he copy if at all) to everybody's satisfaction. I don't know if it was up to Rajlich to insist on such a course of action. Maybe it would have been a good idea to involve Rajlich in the discussion on how to shape the investigation procedure.

Btw, when did Rajlich state that he no longer had the source code of a particular version? (If you happen to know this.)

[BB+]

Did KT spend any significant amount of time on digging into the matter?

What I know is that it was KT who suggested asking VR for source code (this suggestion was done privately to the Secretariat soon after he joined the Panel, but became more widely known later, partially by Chinese whispers). [One might say that no one else in the Panel had the cojones (nor the stature) to propose such a direct step]. Though I don't have the correspondence, I think the implication was that he would be willing to examine it if no one else was (back in 2006 Björnsson then Schaeffer independently inspected the LION++ code).

Btw, when did Rajlich state that he no longer had the source code of a particular version? (If you happen to know this.)

Off the top of my head, in his email correspondence to Schüle in June 2010 regarding Rybka 3. Maybe there were prior mentions at RybkaForum in 2008 (regarding Strelka and Rybka 1).

Maybe it would have been a good idea to involve Rajlich in the discussion on how to shape the investigation procedure.

We just need someone to chime in that control of the process is the control of the result... Or to take the antagonistic point of view, why should Rajlich lend legitimacy to the ICGA processes by consulting with them even about the protocol, when "it was clear from the beginning" that whole thing was a sham.

I agree it is not entirely obvious how to follow the "after seeking expert advice" model.

One possible model could be the old adjudication regime. However, my recollection (likely not finding within the post-edit time) is that Levy made a proposal in an old ICGA Journal (circa 1980) that had 3 experts, one for, one against, one neutral, etc., and Thompson pointed out many flaws in it. The CAS has each side pick someone from a predetermined list (which is perhaps ethereally present here), and those two then agree on a third. Other assocations might have a "tribunal" overseen by a past president, who selects the members, etc. Anything to try to make it seem (and hopefully be a reality) that there is independence.

[Rebel]

Was it wise of Levy to write that column?

Allow me to comment only on this part from your excellent post.

Levy had to do something.

There was an email in his inbox signed by 16 (angry) programmers. Doing nothing (against Rybka) would be the end of the yearly WCCC cycle. The CSVN still remembers this when they allowed Rybka version 4 (or later) to enter in their next tournament (after the ICGA verdict) resulting in a massive rebellion of the programmers. The same would have happened to the WCCC.

So Levy did something but not very wise.

------

Secondly there are quite errors in the Levy article on Chessvibes. I only want to highlight one misconception that still is stubbornly alive while false. It's Vas statement about Strelka (Date 2008-01-11 12:26) and people not realizing Vas a few days (and pages) later retracted that in the same thread (for those are willing to look it up).

Strelka is a mix of Fruit and Rybka.
Fabien recognized his own code.
Vas recognized his own code.
The copyright of Strelka belongs to both gents.

Hope this clears the stubborn myth that that particular post of Vas was (a sort of) half confession.

[BB+]

He wrote this just a few days before the case against Rajlich was launched and we know of no other case that he might have had in mind.

There was also Houdini floating around [it is mentioned in the leader of the article, but not by Levy], among other spawnings from IPPOLIT, Robbolito, Fire, Squarknll (which he mentioned), Toby Tal, ... The computer chess world was indeed full of "clones" (and allegations thereof) at the time, and so might give some context for Levy's brutal verbiage. Letouzey's open letter (addressed "to the community") was about Rybka, and Levy's article also goes in this direction. He could of, say, instead spilled more ink concerning Houdini [who was then #1 on rating lists IIRC] -- but then again, there was no Houdini/Rybka analysis readily available (at best there was my IPPOLIT/R3 work) and I doubt that Houdart would have any reason to give Levy the time of day. Additionally, the programmer's open letter to the ICGA came 10 days after Levy's article, and FL's letter had not mentioned the ICGA (and in the TalkChess thread he seems concerned about, if anything, "commercial interest"), so it's not completely clear that Levy could expect this to "land on his desk" (or at least to the degree that it did).

Of the alleged clones most frequently mentioned at the time, only Rybka had competed in ICGA events. Perhaps the article was intended to be "here is what the ICGA knows about the situation", sketching ZW's evidence (which indeed is termed "rather compelling on the surface") and Rajlich's response (this document is horribly bogus.. all that "Rybka code" isn't Rybka code, it's just someone's imagination), along with Rajlich's claims about Strelka (as an enlargement upon the general cloning topic). I'm not quite sure how this is 1+1 to make 2, though with the final paragraph of the forum for investigation one might (rashly IMO) be led in that direction. Probably the best thing to say is that Levy is a publicity man, and not one attuned to judicial-legal matters. He took the road that was most familiar to him, but we can easily second-guess that is was imprudent in the end (and some of us probably could have first-guessed it was imprudent in the beginning).

It's Vas statement about Strelka (Date 2008-01-11 12:26) and people not realizing Vas a few days (and pages) later retracted that in the same thread (for those are willing to look it up).

It seems that this is another ghost reference. I went to the thread (note that the Levy article has a incorrect link), and enumerate the following posts of "Vasik Rajlich":

Page 1: [50 posts per page]
12:26 #39198, initial post
10:10 #39409, thanks for support
10:15 #39412, idea of releasing Strelka source (as his) under GPL
10:20 #39413, IP rights, comment about Shredder and RE
22:09 #39670, continutation of #39412 regarding GPL
22:24 #39671, comment about gray IP areas

Page 2: none
Page 3:
10:21 #39414, clarify point #1 please
10:57 #39425, comment about shooting guns (first)
22:30 #39673, shootings on a 1-parachute airplane

Page 4:
10:24 #39415, thx to Peter Skinner

Page 5:
10:30 #39416, "what goes up, comes down" (to Cozzie)
10:36 #39417, query to Hristo about licensing
22:35 #39674, thx for comments, hold off doing anything

Which one am I missing, where he makes the retraction? I definitely remember him later stating his "claiming Strelka as his own" was illegal (and thus he didn't do it), but I don't think he retracted any of his analysis of similarities.

[And yes, I fully expect to be termed some maniac in that I was actually willing to do what you asked namely read the thread to try to find this supposed retraction].

[Rebel]

Mark,

It starts here [ http://rybkaforum.net/cgi-bin/rybkaforu ... 0#pid39310 ]

Hristo responding on (critizising) Vas' initial post (hasty judgement) the same day.

[BB+]

It starts here http://rybkaforum.net/cgi-bin/rybkaforu ... 0#pid39310
Hristo responding on (critizising) Vas' initial post (hasty judgement) the same day.

Rajilch's comments are:

Thanks. Indeed, there may be some license issues and I'm not very familiar with them.
How exactly would I go about doing what you suggest?

ok, thanks for the comments. I'm going to hold off on doing anything for now, this will all take some time.

I still don't see where he "retracted" anything regarding his Strelka analysis -- he just agrees with Hristo that "that there may be some license issues" with claiming ownership and then "releasing it in the next few days under my own name" (quoting the original post). If the long-standing debate with Rybka/Strelka was whether VR had claimed/retracted ownership then I can see your point -- but any argument has not been about "ownership" but more about Vas's statement that: The picture is quite clear ... Vast sections of these sources started their life as a decompiled Rybka 1.0. Particularly when said Strelka sources were "Fruitified" by Osipov, this was then the starting point of the Rybka/Fruit saga.

One could also mention VR's 2012 comment to an inquiry (of mine) regarding Strelka/IPPOLIT: I doubt that all of that code was typed by hand. If it was, then sure, it's "original at the source code level", and on the first point one must enlarge the meaning via his "super-geek" analysis that original means that the author either typed his own code or typed the code which generated his own code (perhaps recursively). To the best of my knowledge, these are his principal statements regarding Strelka after the initial claims, and that he still holds that "vast parts of [Strelka] sources started their life as a decompiled Rybka 1.0", even though Osipov's end creation could well be "original at the source code level". For extra credit, debate whether VR would have a valid Rule #2 complaint against Strelka, under your/his understanding of it.

[BB+]

Btw, when did Rajlich state that he no longer had the source code of a particular version? (If you happen to know this.)

Off the top of my head, in his email correspondence to Schüle in June 2010 regarding Rybka 3. Maybe there were prior mentions at RybkaForum in 2008 (regarding Strelka and Rybka 1).

After more thought, the first time I could think of would be in (say) Aug-Sep 2008, when various bugs with R3 were beginning to be reported. VR made some comments along the lines that there would be no proper "patch" release, and that he couldn't fix any specific bugs (such as the crash NH found) anyway since the Rybka source had changed too much since the release. [The never-seen R3+ was talked about for the next year, but was always just going to fix problems under the current code base -- indeed, he was talking about toning the ELO down to the R3 level so as not to disadvantage the gap that future R4 buyers would obtain]. [There were plenty of defenders of this "only keep the latest" model of development (especially with a single person), though others inquired whether a source-control system might be warranted]. I think the threads have migrated to some secret part of the Rybka forum (in the R3 wake there was plenty of clone talk). But by this time in 2008 the eagle-eyed could realize that VR likely didn't have any old versions lying around (he actually asked if anyone had the various pre-R2.3.2 executables at one stage, for which there were many due to the subscription model). I don't remember (and certainly am not going to look for) anything he said about the (non)existence of Rybka 1 source code at the time of the Strelka appearance.

What I know is that it was KT who suggested asking VR for source code... [One might say that no one else in the Panel had the cojones (nor the stature) to propose such a direct step].

There is also a (rather important) contextual matter here. There were plenty of rumblings at the time that the whole point of the ICGA exercise was to let his competitors "peek at his source" or something like that. It could well have been a PR disaster and/or an extended IP dispute (complete with lawyers, if the afore-cited 2008 Beijing thread is to be believed, though admittedly the poster therein is not exactly shy of hyperbole: If I was Vas at Beijing, I would have a lawyer available on the other end of a telephone, or present in person, to deal with the icga if they tried this stunt again). Anyway, the situation with commercial entities and source code availability seems to do nothing more than translate the question from "what is original?" to "what is a reasonable complaint?" and again the ICGA could be on-the-hook no matter what it does (which admittedly, is part of being an association, sometimes you are going to have to make decisions and endure disputes). [And the previously mentioned comments of Ban regarding the failed Rybka-Junior match could be useful in this regard].

[Rebel]

Similarly in the 2013 emendation regarding the purported "right to an appeal", Schröder's cited correspondence (as Rajlich's carte blanche representative) tends to the vexatious by the end.

You may find that, I could say the same for your sense for fair procedures to be vexatious as well.

An appeal is a fundamental right.

[Rebel]

Another argument would be that he had known the essence of these allegations for some time, and it's not clear exactly what giving extra time would really gain for him (other than as a delaying tactic), certainly without himself specifying anything in particular. How long should it take to construct a defense to the main claims (ignore Rybka/Crafty for now)?

• Nearly the entire evaluation function (of Rybka) is derived from Fruit
• In summary, the evaluation functions are nearly identical
• Rybka 1.0 Beta and Fruit 2.1 have exactly the same evaluation features (preliminary finding)
• Comparison of evaluation features shows high similarity of Rybka (including 2.3.2a) and Fruit

It took me (while scrutinizing your and Zach's document) 4-6 weeks to see something fundamental was wrong summarized in one word: tunnel-vision. It took me about 3 months to conclude there is barely evidence Rybka started its life as Fruit. And much later my final conclusion ending with the words of Fabien himself. Link to Fabien's full statement.

The two of you created a lot of complicated stuff and in retrospect I say none of us (you and Zach included) were experts in the field of judging the material you produced. I only have to list all the things you missed.

In such a complex scenario it's only reasonable the accused gets a proper time to put up a defence and refute the 2 documents point by point. Which was denied to him.

[Rebel]

It starts here http://rybkaforum.net/cgi-bin/rybkaforu ... 0#pid39310
Hristo responding on (critizising) Vas' initial post (hasty judgement) the same day.

Rajilch's comments are:

Thanks. Indeed, there may be some license issues and I'm not very familiar with them.
How exactly would I go about doing what you suggest?

ok, thanks for the comments. I'm going to hold off on doing anything for now, this will all take some time.

I still don't see where he "retracted" anything regarding his Strelka analysis -- he just agrees with Hristo that "that there may be some license issues"

I see human being horrified, who just became aware his program is cloned making an emotional public outburst, what's new? Then Hristo comes along and puts some sense back into him.

with claiming ownership and then "releasing it in the next few days under my own name" (quoting the original post).

Which he did not do.

Isn't that all what matters?

If the long-standing debate with Rybka/Strelka was whether VR had claimed/retracted ownership then I can see your point -- but any argument has not been about "ownership" but more about Vas's statement that: The picture is quite clear ... Vast sections of these sources started their life as a decompiled Rybka 1.0. Particularly when said Strelka sources were "Fruitified" by Osipov, this was then the starting point of the Rybka/Fruit saga.

Osipov... the guy who wanted to commercialize Strelka and was pissed with Vas because didn't allow him?

And you believed him?