Reports | June 29, 2011 8:42

Rybka disqualified and banned from World Computer Chess Championships

Rybka disqualified and banned from World Computer Chess ChampionshipsThe International Computer Games Association (ICGA) has disqualified and banned Rybka and its programmer Vasik Rajlich from previous and future World Computer Chess Championships. The ICGA accuses Rajlich of plagiarizing two other programs, Crafty and Fruit, and demands that he returns the trophies and prize money of the World Computer Chess Championships in 2007, 2008, 2009 and 2010.

For quite a while there has been some serious turmoil in the computer chess world. When we reported about the Houdini-Rybka match in February of this year, the article triggered lots of comments about the issue of cloning. Was Houdini derived from the Ippolit series? Was it plagiarized from Rybka? And what about Rybka, was it largely based on the code of other engines? Nine days later we published an article by IM David Levy, President of the International Computer Games Association (ICGA), who shared his thoughts about how to tackle the issue. A few days later he announced the establishment of the ICGA Clone and Derivative Investigation Panel. Not long after, on March 1st, we received an open letter about the Rybka-Fruit case signed by fourteen chess programmers. They all supported the claim that Rybka was cloned from Fabien Letouzey’s Fruit. In the last few months all the allegations have been seriously studied by the International Computer Games Association (ICGA). On Tuesday night we received the following text from the ICGA President himself.

Rybka Disqualified and Banned from World Computer Chess Championships The International Computer Games Association (ICGA) has been conducting an investigation into allegations that, in the chess program Rybka, the programmer Vasik Rajlich plagiarized two other programs: Crafty and Fruit. The ICGA has considered and evaluated the evidence presented to the investigation panel and the report prepared by the panel’s Secretariat. (The report and evidence files are attached.) We would like to thank those members of the panel who contributed to this investigation and the Secretariat for the enormous amount of conscientious work they have put in to this matter. By a unanimous 5-0 decision of executive members of the ICGA we find ourselves in agreement with the verdict of the Secretariat’s report. We are convinced that the evidence against Vasik Rajlich is both overwhelming in its volume and beyond reasonable question in its nature. Vasik Rajlich is guilty of plagiarizing the programs Crafty and Fruit, and has violated the ICGA’s tournament rules with respect to the World Computer Chess Championships in the years 2006, 2007, 2008, 2009 and 2010. Specifically, Vasik Rajlich, on all five occasions, violated Tournament Rule 2 which requires that: Each program must be the original work of the entering developers. Programming teams whose code is derived from or including game-playing code written by others must name all other authors, or the source of such code, in their submission details. Programs which are discovered to be close derivatives of others (e.g., by playing nearly all moves the same), may be declared invalid by the Tournament Director after seeking expert advice. For this purpose a listing of all game-related code running on the system must be available on demand to the Tournament Director. By claiming other programmers’ work as his own, and failing to comply with the abovementioned rule, Vasik Rajlich has unfairly been awarded one shared 2nd-3rd place (in 2006) and four World Computer Chess Championship titles (in 2007, 2008, 2009 and 2010). Furthermore, it seems to the ICGA that Vasik Rajlich clearly knew that he was in the wrong in doing so, since he has repeatedly denied plagiarizing the work of other programmers. The ICGA regards Vasik Rajlich’s violation of the abovementioned rule as the most serious offence that a chess programmer and ICGA member can commit with respect to his peers and to the ICGA. During the course of the investigation and upon presentation of the Secretariat’s report Vasik Rajlich did not offer, despite repeated invitations from the ICGA to do so, any kind of defence to the allegations, or to the evidence, or to the Secretariat’s report, other than to claim in an e-mail to myself on May 13th 2011 that: Rybka has does not "include game-playing code written by others", aside from standard exceptions which wouldn't count as 'game-playing'. The vague phrase "derived from game-playing code written by others" also does not in my view apply to Rybka. The ICGA is of the view that such a serious offence deserves to be met with correspondingly serious sanctions against the perpetrator. In deciding on appropriate sanctions the ICGA has borne in mind the approach of the International Olympic Committee for dealing with the most serious cases of the violations of its rules. The ICGA has therefore decided as follows:

  1. Vasik Rajlich is hereby disqualified from the World Computer Chess Championships (WCCC) of 2006, 2007, 2008, 2009 and 2010.
  2. The 2nd-3rd place awarded to the program called “Rajlich” in the 2006 WCCC is hereby annulled, sole 2nd place is awarded to the program Shredder, and 3rd place in that event is awarded to the program Zappa.
  3. The 1st places and World Computer Chess Champion titles awarded to the program Rybka in the 2007, 2008, 2009 and 2010 WCCCs are hereby annulled, and all the other programs that competed in those events are moved up in the final tournament standings by one place. Thus the revised tournament standings and titles for those events will now be as follows. 2007 1st Zappa (World Champion) 2nd Loop =3rd GridChess =3rd Shredder 2008 1st Hiarcs (World Champion) 2nd Junior 3rd Cluster Toga 2009 =1st Junior (Joint World Champion) =1st Shredder (Joint World Champion) =1st Deep Sjeng (Joint World Champion) 2010 =1st Rondo (Joint World Champion) =1st Thinker (Joint World Champion) 3rd Shredder
  4. In due course those programmers whose programs have been elevated to World Champion (or joint World Champion) status will receive from the ICGA replicas of the Shannon trophy for the appropriate years.
  5. The plaques on the Shannon trophy that currently bear the name Rybka (for the years 2007-2010) will be removed from the trophy and new plaques will be engraved with the names of the revised winners of the title.
  6. Similarly, the titles of World Computer Speed (Blitz) Chess Champion that were awarded to Rybka in 2009 and 2010 are hereby annulled. The revised winners of the speed chess title for those years are therefore: 2009 Shredder 2010 Jonny and Shredder (joint champions)
  7. Vasik Rajlich is banned for life from competing in the World Computer Chess Championship or any other event organized by or sanctioned by the ICGA.
  8. The ICGA demands that Vasik Rajlich return to the ICGA the four replicas of the Shannon Trophy presented at the World Computer Chess Championships in 2007, 2008, 2009 and 2010, and to return to the ICGA all prize money awarded for Rybka’s performances in those events.

David Levy [President - ICGA] June 28th 2011

Mr Levy also sent us a big number of documents which according to the ICGA form the evidence to the claim that Rybka was plagiarized from Crafty and Fruit. These documents can be downloaded below for anyone who wishes to dive further into the material. The news is obviously a huge blow for the Rybka team. The impact in the computer chess world must be comparable to arguably the most famous example of doping in athletics: the positive drug testing of Canadian sprinter Ben Johnson in 1988. We've asked Vasik Rajlich for a comment and hope to add this later.

Downloads

Peter Doggers's picture
Author: Peter Doggers

Founder and editor-in-chief of ChessVibes.com, Peter is responsible for most of the chess news and tournament reports. Often visiting top events, he also provides photos and videos for the site. He's a 1.e4 player himself, likes Thai food and the Stones.

Chess.com

Comments

BlunderSuck's picture

Time for revolution...aux armes citoyens!

S3's picture

maybe because Rybka is better than Carlsen Zeb? ;)

Brian Wall's picture

I interviewed Vasik Rajlich on ICC and I like him a lot. However he got started he obviously put an enormous amount of work into improving Rybka. So did IM Lawrence Kaufman.

hatsekidosie's picture

And why were Crafty and Fruit not world champions?

TMM's picture

That's where the other 36% comes in...

Jeremy Bernstein's picture

A commercial Houdini has all the same problems as Rybka. It's a derivative of Robbolito/Ivanhoe/IPPOLIT, with some additional (very good) work done to improve playing strength. However, it's plagiarized (output (eval and pv) of early versions of Houdini corresponds 100% to Robbolito output in many positions). Robbo/Ivan/IPPO are unlicensed open-source, so the code is fair game, but it's not cool to take someone else's code and call it your own.

Not only that, the origins of Robbolito/Ivanhoe/IPPOLIT are still unclear, with a fair amount of circumstantial evidence pointing to an origin based on reverse-engineered code, probably from Rybka 3. Illegal? Likely not, but again, ethically dubious.

Houdini competing at WCCC is most likely out of the question for these reasons. Houdini being sold? It's a risk that Robert Houdart would be taking -- if Vasik Rajlich decided to pursue a copyright case against Houdini and could prove the links mentioned about (in particular the Rybka-IPPOLIT link, which would be difficult). And it's most certainly a risk for larger companies (ChessOK or ChessBase) who would consider distributing it.

Then again, most people don't care about this stuff and are just happy to pay whoever for whatever, as long as it's the strongest. Sure, Houdini is a strong and useful engine, and Robert Houdart deserves some credit for improving the code he started with. But it's unfair to the people who developed that code (be it the IPPOLIT developers, or Vasik Rajlich, if the RE claims are true (unclear)) to claim it as his own -- he added icing to someone else's cake and wants to be known as a master baker, to quote another post in this thread.

Anyhow, ICGA should be lauded for taking the high road with these issues.

MarkWatkins's picture

"Houdini competing at WCCC is most likely out of the question for these reasons."

I think it would have been an interesting question to put to Rajlich as to whether "derived from game-playing code written by others" would apply (in his opinion) to any putative Houdini entry for the WCCC, or even a non-anonymous IPPOLIT/IvanHoe for that matter. If so, what's the crucial difference vis-a-vis Rybka/Fruit?

As for commercialising Houdini, my current thinking is that IPPOLIT *does* violate copyright, but Rajlich has no real monetary claim against it (scuttlebutt is that it didn't hurt his R3 sales, for instance), and getting a cease-and-desist order regarding distribution would be woefully untimely. As for Houdart, I would suspect that there will be no IPPOLIT remnants if/when he goes commercial.

Incidentally, Rybka tried to enter the ICT in Leiden a couple weeks back, but they were denied entry, even prior to the ICGA ruling (I think the basis was, rather Kafka-esquely, that the investigation was still ongoing).

Sergio's picture

This verdict would not stand in a proper court I think. If it would then Rybka already would have been sued for compensation, cause they used open source codes for commercial use.

Another thing I don't understand is why the Rybka programmer had to give evidence to defend himself? Doesn't that go in against the principle: innocent untill proven guilty?

snits's picture

There is nothing against the law in using open source code for commercial use as long as you honor the license of that open source code. The GPL in no way prohibits the selling of GPL software.

I don't think he had to give evidence, he was offered a chance to provide a rebuttal against the evidence submitted.

TG's picture

That is the whole point of a defense: To show that the prosecution's evidence is faulty. Vas chose not to try for whatever reason.

Mike Pacasi's picture

The solution for this mess is to contribute to increase the support to free software chess engines like GNUChess, Crafty and so on. Linux can beat MS, so Free chess can beat closed engines also, because millions of brains will produce better code than a few.

TMM's picture

Yeah, Linux has definitely beaten Microsoft. Everywhere I go, I only see Linux machines! And Microsoft is nearly bankrupt!

(That was sarcasm by the way.)

Anonymous's picture

I get your sarcasm but chances that you really see Linux machines everywhere you go.
Android is based on Linux and most of it is open source. Many embedded devices such as home routers use Linux too. Apple's iOS and OS X don't use linux but at the core is Darwin, an open source BSD-style Unix.

MarkWatkins's picture

Regarding the ChessVibes text: "Not long after, on March 1st, we received an open letter about the Rybka-Fruit case signed by fourteen chess programmers. They all supported the claim that Rybka was cloned from Fabien Letouzey’s Fruit."

First, it ended up being 16 programmers who agreed to be signatories. But more importantly, I think it is a bit misleading to say that they all *necessarily* supported the claim that Rybka was cloned from Fruit. Rather they all (at the very least) considered the evidence to be sufficiently robust (and of high enough importance) that the ICGA should commence a full and proper review of it (allowing, of course, Rajlich the ability to respond to the allegations). Some of these programmers later participated in the evidence gathering/debating phases carried out by the Panel, while others simply chose to allow the ICGA to handle the matter.

Henk de jager's picture

I think now would be a great time to commercialize Houdini. It is clearly the strongest programm ever. Have it participate (and probably win) the next CCWC, build in a database and a cool opening book and every serious player would HAVE to buy it. I know I would.

Johnny's picture

@Bert, I don't necessarily agree you on all points, but i appreciate your thoughtful answer. My first issue is with the phrase "nearly identical" which the panel apparently intends to mean "derivative".

Second, as a legal concept I am curious about the scope of so-called "copyright" protection for what amounts to a function-oriented process. Yeah, patents are harder to get, for good reason! We don't want folks "patenting" useful processes (eg, "chess evaluation function") unless the patent application is sufficiently detailed to enforce any infringements. And that seems a huge problem here: as I understand it, the GPL is not regulated-- you just stick the GPL logo bitmap in your FTP/website folder and boom, you claim copyright protection under GPL. Except perhaps such content is not even copyrightable. In that case the GPL would be enforceable only by the "honor system". :) Again you are correct that this is all very technical-- and not my field of practice in any event (certainly not European law at that!). In any event, I would prefer if the ICGA panel had engaged an intellectual property lawyer to sort out such issues. As presented, the reports strike me as conclusory and rushed which is rather unbelievable considering that the events in question are from 2005-2006.

Thanks for your useful comment, Bert.

Johnny's picture
Crazylocha's picture

As far as the issues of "similarity in eval %", one should remember that in programming terms, it means a lot more than a normal evaluation in say, a car accident trial. When you consider the detailed reports show similar programs show very high percentages against each other, e.g. bitboard 64bit programs vs. same, and low percentages against other types, it begins to come into a clearer focus and understanding. The detailed reports show they extensively tested against several style engines for statistical probabilities. Low side was 24% against non Fruit and Crafty. High side was 72% against them (note two versions og rybka were tested, early vs. Crafty, later vs. Fruit). The margin of variation was 7%, over quite a few runs. This does vary in percentage due to analysis tweaks that are done by programmers, e.g. value of a rook on 7th rank or pawn on the a or h file. That is why it is not a 100% match. They did also show where Fruit changed a.major value for evaluation, and rybka did same, at exact value, not an easy coincidence to overlook. When you understand the intricacies of chess programming, the work the panel did was extremely well done and thorough. There is not ANY squiggle room.

It might behoove "Mr. Lawyer" to read the part of the report where early rybka used Crafty's obsolete undeleted code (20 or so lines of it). Really?? But not guilty??

Fwiw, since rybka was both copyrighted and trademarked, wouldn't that mean the programmers of Fruit and Crafty are legally due to the monies made from selling rybka? I can see the trial judge not taking lightly to refusals of source code. Whom do you suspect the judge would allow for expert testimony? Maybe a jury of peers? I think we have just seen that. NOT ONE PERSON THOUGHT HIM NOT GUILTY!

NOT A SINGLE ONE

Crazylocha's picture

As far as the issues of "similarity in eval %", one should remember that in programming terms, it means a lot more than a normal evaluation in say, a car accident trial. When you consider the detailed reports show similar programs show very high percentages against each other, e.g. bitboard 64bit programs vs. same, and low percentages against other types, it begins to come into a clearer focus and understanding. The detailed reports show they extensively tested against several style engines for statistical probabilities. Low side was 24% against non Fruit and Crafty. High side was 72% against them (note two versions og rybka were tested, early vs. Crafty, later vs. Fruit). The margin of variation was 7%, over quite a few runs. This does vary in percentage due to analysis tweaks that are done by programmers, e.g. value of a rook on 7th rank or pawn on the a or h file. That is why it is not a 100% match. They did also show where Fruit changed a.major value for evaluation, and rybka did same, at exact value, not an easy coincidence to overlook. When you understand the intricacies of chess programming, the work the panel did was extremely well done and thorough. There is not ANY squiggle room.

It might behoove "Mr. Lawyer" to read the part of the report where early rybka used Crafty's obsolete undeleted code (20 or so lines of it). Really?? But not guilty??

Fwiw, since rybka was both copyrighted and trademarked, wouldn't that mean the programmers of Fruit and Crafty are legally due to the monies made from selling rybka? I can see the trial judge not taking lightly to refusals of source code. Whom do you suspect the judge would allow for expert testimony? Maybe a jury of peers? I think we have just seen that. NOT ONE PERSON THOUGHT HIM NOT GUILTY!

NOT A SINGLE ONE

szoker's picture

wow...

what a shame to his name...

Dominic Menard's picture

Oh and also Rajlich never stated that he used a binary system. He should have stated that his system was based on a binary system and thanked Gottfried Leibniz.

Nowadays, you don't code from scratch, you google, you import code you use pre-existing routines. Noone starts with a blank page. Maybe Rajlich made a mistake and was inspired with the source code of other programs, but nowadays any programmer does it and you cannot say whether he was inspired and thought his own ideas were more important than what he replicated. Let Rajlich reply.

TMM's picture

Indeed, the "rules" are rather vague on this. At what point does it start/stop being necessary to give credits to others about work you did? There's a big grey area here, which is claimed to be white by the ICGA

Joeri's picture

Who are the members of the panel? I could not find it.

MarkWatkins's picture

Third link of above, Page 3 (section 1.5): http://www.chessvibes.com/plaatjes/rybkaevidence/RybkaInvestigation.pdf

Secretariat: Hyatt, Lefler, Williamson
Members (some joined quite late, others were mostly bystanders): Silver, Ban, Roberson, Theron, Czechowski, Dailey, Hallsworth, Letouzey, Friedel, Isenberg, Horvath, Bauer, Krabbenbos, Himstedt, Thompson, van Kervinck, Szmit, Watkins, Uniacke, Georgiev, Deville, David Tabibi, Skinner, Schaefer, Vida, Pijl, Meyer-Kahlen, Mayer, Romstad, Pronk, Vuckovic, Garvin, Bjornsson, Wegner.

Zach Wegner and I were the principal investigators, while Bob Hyatt did some (for the Crafty parts). The opinions of various members can be found in section 3 (page 8) in the same link.

steve's picture

on a side note: i do not care.

leigh's picture

My impression:
1. jealous, i think this was jealous from other programmer;
2. illegal, a lot of judges were plaintiffs
3. the evidents not enough

Harvey Williamson's picture

The panel were not the Judges. they prepared a report for the ICGA to Judge.

At least half of the panel members have never competed against Rybka at the WCCC or a similar event.

leigh's picture

but they are competiters, plaintiffs.

Michel83's picture

Just a question to the ones who write about "witch hunt" and "jealousy", no matter what your general opinion about "cloning" might be:

Wasn't it the Rybka-team which spammed the whole Internet publically accusing other engines of having cloned Rybka and talking about ethics?

The irony. What comes around goes around...

I don't know if the accusations are true (just like I never knew if the accusations against other engines were true), but I do know Rybka launched the same "witch hunt" on other engines.
What could be fair enough if they had evidence (which they never presented), I am just saying that the Rybka-team acted in the same way as what is happening here and accused others of cloning, were you there to blame them of "jealousy" and "witch hunt" as well? So let's keep the discussion fair and not have double standards.

...but well, as much as the question of ownership and cloning is interesting, as much it's also a market with money and egos involved...same old story I guess.

leigh's picture

4. all the words used here were used by somebody else, so you can say I plagiarized

Mike Pacasi's picture

I would suggest to the Rybka team to open the code as a Free Software Chess project, under the protection of a social contract.

Ashish's picture

I would like have liked the evidence to include some description of how other programs do it, i.e. that the supposed commonalities between the evaluation functions of Rybka and Fruit are not also common to all chess programs. (I've looked at a couple of the docs, and did not see any such mention.)

Without such a description, the charges are like accusations that some GM cheated in a game because "his moves match Rybka's moves 90% of the time." Well, maybe those were clearly the best/only moves!

Tano-Urayoan's picture

Code vs code was analyzed and found a lot of similarities or even equal routines between both first pre-Rybka-Crafty and then Rybka-Fruit it has nothing to do with evaluating positions. If ad-verbatim code was copied then Rybka code needed to be published as requirement for GPL license agreement.

MarkWatkins's picture

The document you want is: http://www.chessvibes.com/plaatjes/rybkaevidence/EVAL_COMP.pdf
This compares two Rybka versions, Fruit 2.1, and six other pre-2005 open-source engines, using a quantitative method (which is of course subject to critical analysis).

NPS's picture

Such a comparison is included in the investigation.

Vasisacheat's picture

Finally justice is served...

Arvin's picture

I'm glad that I'm only using Fritz, Junior, and Fruit engines.

alx's picture

I don't think this is over.
1- I'm pretty sure the Rybka team is looking into ways to overturn this. Otherwise, this is the end of their business.
2- This looks like enough reason to start a GPL violation investigation since there is strong evidence of "derived" work and distribution in the GPL sense. Even if the original author(s) don't care, the FSF and gpl-violations.org try to enforce GPL compliance and the GPL seems solid so far. I'm not sure if there are less radical ways of compliance than a solution that leads to the availability of Rybka's sources since last time I checked some of the derived parts were all over and essential, certainly not something to easily factor out into a separate program.

That Don Guy's picture

Unless the Rybka owners lose some sort of actual civil case and are forced to pay damages and/or cease selling the software, how is this the end of their business? True, there are "new" world champions for these years, but one fact remains; they were inferior to Rybka. (Pretty much the only thing that would prevent Rybka from being marketed as "Better than the World Computer Chess Champions!" is some sort of trademark on the term "world computer chess champion.")

My first thought was, if Rybka is "primarily" Crafty and Fruit, allow it to keep its titles, but require whoever was paid the money and given the trophies to give them back, and then give them to the Crafty and Fruit developers - but then I realized that Rybka (presumably) did break the rules, and even this wouldn't be fair to the other programs (who knows who could have won if all of the developers were allowed to, er, "borrow" code/algorithms from other engines?).

Michel83's picture

Another question:
The cloning-dicussion has been around for a long time now and even among computer-experts there seems to be very differing opinions (not only "morally", but also purely "technically").
Is there any precedent court decision about copyright concerning programming codes, the evidences that need to be presented, the limits of it etc.?
If not, then if this goes to court it will create an important precedent for the future that will influence computer chess and chess programs a lot, no?

Septimus's picture

What kind of commercial impact will this have on Rybka? Are any of the other programs patented? Will we see a slew of infringement suites against Rybka?

Johnny's picture

your comment shows a lack of sense of humor, but that's another story, too. :D

Dennis M's picture

Two people can each have a sense of humor without necessarily finding the same things amusing. :) I can see why it would be funny if the analogy made sense, but it completely misfires.

leigh's picture

I read the reports.
I am sure that this is a jealous case.
Your programs couldn't beat Rybka, So you guys beats Rajlich. Funny!
All the judges from competiters, from interests related, from plaintiffs, how could the investigation keeps fair!!!!?????
if it was illeagal, it is easy to decide; you don't need to invite the competiters to decide. you should invite some common programmers to decide.
I am an amatuer, I know the name of Rybka, but I never used it. Because it costs money. I only use the free engine.
in our daily life, (1+1)+(1+1)=2+2=4, this is right way. professors and primary school students do it same.

Dave Acevedo's picture

There is no copying. Whats the point? These amazing, wonderful and fear inspiring and hated events were know by all of Jerusalem if not all of Isreal and further. It is not copying, it's unavoidable similarity. We have several viewpoints from men moved to write by God Himself to place side by side to make it easier for many different types of persons to understand God. What is more - Their similarities PLUS their differences are a mark of the truth to what is recorded.

Johnny's picture

I am afraid Vasik Rajlich's next chess engine will be called Skynet and together with its friends Houdini, Rybka, Stockfish, and Critter will take over the world. Get your ass to mars! :)

Al's picture

Shocking and sad news, reminds me of copying other students computer programs at school and changing variables names and the formatting of the code to look a little different not a good thing to do, very sorry to hear this as I have been a big Rybka supporter, even called my fish Rybka!

Can Houdini compete in ICGA tournaments? There will be no competition for Robert for a while!

leigh's picture

Nobody dare to, hahahaaa

Greg's picture

Houdini probably cannot compete in ICGA tournaments, but then again he has never tried to. Houdini appears to be based on an open source program of unclear provenance which, whether legal or not, the ICGA would not allow.

Loch Fairnees's picture

Wouldn't it be fair to let Vasik face the accusation and state his point of view before judgement?

Pages

Latest articles