A short history of the anti-rybka inquisition

[Lusakan]

Vas came on the computer chess scene with a weak Rybka, joined the CCC and asked questions from some of the old timers here and got answers, he looked at the fruit code and then he produced Rybka 1.0 which he claimed was 100 elo stronger than the mighty Shredder. He was almost laughed out of the CCC. Then the testing results started coming in and he was right. Rybka was blowing all the hitherto feared chess programs out of the water. His is a story of he came, he saw, he conquered.

Vas decided to go commercial with Rybka and reduced participation in CCC discussions. He soon had his own Rybka forum and built a following of Rybka fans. The fans become a sort of search group for bugs and weaknesses and they kept Vas’s feet to the fire with a constant barrage of criticism and ideas which he worked on promptly issuing a series of updates. Vas become the most accessible commercial programmer in history, which was interesting to watch because he was a commercial programmer while at the same time being a sort of “people’s programmer” at the head of fan group dedicated towards a vision of producing the strongest chess engine possible or perhaps solving chess. This led to the growth of the strength of Rybka in leaps and bounds. Soon the chessbase engine room was being called by some people as the Rybka room. This has continued to this day when the dominance of Rybka has become as some people have said, “insane”.

Obviously the programmers who represented the old paradigm were not going to enjoy this. Sometime during the time of Rybka 2.0 someone asked a question to programmers on CCC to share how they felt watching their brain child and life’s work being beaten senseless by Rybka, he got a volley of angry retorts along the line of “shut up stupid kid!”, and he did shut up, but for the first time it was clear that some egos that had taken years to build had been busted and there was a festering anger somewhere below the surface and soon or later it would break out.

Soon it become obvious from discussions in the CCC that programmers were shifting the direction of their energies less from implementing their own ideas into their programs and more on trying to figure out what exactly made Rybka stronger than their programs. The easiest route was to ask Vas, but Vas had stopped participating in the CCC. This was annoying. A number of comments and complaints were heard suggesting that it was bad and wrong what Vas did, coming to the CCC asking questions and learning things and after learning going commercial and not sharing with others what he had learned. The logic was: he knows something we don’t know, he must have got it here because he was here, and therefore he is in debt to tell us what he knows. The possibility that he has no obligation to do so or that what he learnt is the core of Rybka as a commercial entity and that he needs to protect that to make a living is considered a non-defense. This was the first skirmish.

Then next skirmish: In the absence of vas telling people the secret of his gprogram, energies shifted to theoretical speculation on how Rybka works. There where endless debates on whether Rybka was strong because it out-searched other engines or because it was more knowledge based. There were theories on how it searched and on what and how its knowledge was structured. To answer these theoretical questions people tried to analyse Rybka’s PV and node output to try and extract inferences of its inner workings. They hit a brick wall. Rybka’s output was not following the normal academic conventions and standard definitions of PVs and nodes and its output was useless for purposes of peering into its inner workings. Discussions about Rybka’s node count “obfuscation” raged for months. How dare Vas use none standard ways of calculating nodes? A node should be a node so that rival programmers have a fair chance of knowing how the engine … , no, so that customers who have paid for the engine are given correct information about how deep the engine searches. Vas is not being professional by doing things his own way rather than how the rest of the chess programming profession do things. One of the fiercest accusers of Vas, who always demands yes or no answers from vas, was challenged by someone who knows his own node count practices to say publicly whether or not his program adheres to the standard he wanted Vas to follow. He failed to answer this question and failed to dispute this public charge of hypocrisy that was put before him. Instead he told the person that what he knows was privileged and he should not have asked about it publicly in the forum. He instead offered to share his way of counting nodes “to the Vas camp” if Vas please could agree to show him how Rybka counts nodes. This deal offer was declined. Another interesting thing with this hue and cry about Vas’s node count from programmers was that ordinary customers are contented and happy that they were getting their full elo’s worth of the money they had paid and don’t feel cheated in any way by how nodes are counted. All international competitions and testing groups are showing that chess programs written by people who adhered to the so called gold standards of coding PVs and node counts were having their butts kicked blue by Rybka.

Third skirmish: And then in comes Strelka. A hacker was claiming to have managed to decompile Rybka and was making the source code available to the general public under the name strelka. At last, programmers could finally look at Rybka’s code and learn its secrets. The holy grail of chess programming would at last be in the open for any rival programmer who wanted to catch up with rybka to analyse. Unfortunately the hacker had mixed the decompiled code with other code from other programs such as fruit. It was not clear which parts ware original Rybka code and which parts were from other programs. Anyway people poured over the Strelka code to discover what made Rybka different. Antony Cozzie, one of the most feared programmers looked at the Strelka code and his conclusion was that to his surprise it was a simple unremarkable program, with little chess knowledge, but optimised for fast search. No one else has given a statement of what makes Rybka strong from an analysis of Strelka code. Looks like not much was leant from that code and this road may have reached a dead end. There are also challenges on the value and utility of what you find. Even if you figure out how Rybka works from Strelka code, if those ideas are dramatically original your engine would be a clone. To make things worse Strelka is said to be based on Rybka 1.0 where as the current Rybka 3.0 has had more than 2 years of work and is vastly more powerful, hence it will take a long time for your engine to catch the strength of Rybka 3.0. To have any chance of being competitive you should at least start with a clean copy of Rybka 3.0 source code.

Fouth skirmish: “Hay wait, didn’t Vas say he learnt from fruit source code when developing Rybka? Isn’t fruit under the GPL licence? Doesn’t the GPL licence say if you get even a single line from fruit code you have to make your whole program open source? Ureka! This means if you can prove that Rybka has even a single line of code from fruit you can make noise and unending accusations of him having violated the GPL licence of fruit. Can’t you see? If Vas got a single line from fruit the entire set of Rybka’s, including version 3.0 consists of clones regardless of what work Vas put into them and we can demand that he should hand over the entire source code to us, at last we can have his source code. What? A hacker has compared Strelka and Rybka binaries and has found similarities with fruit? Good job. Let the shaming game begin. Unlike us patented geniuses who single handedly invented every single idea in our programs, which we have worked on for the past 4 to 5 decades (see how knowledgeable we are?), Vas is a common thief with no originality. At last we can knock him off his pedestal of honor and steal his thunder. We can do what our programs failed to do, checkmate vas! Our programs, even if they are 500 elo weaker than Rybka, will shine again for their sublime originality, who cares about solving chess? Any one with proof to support our objectives is welcome. Original mediocrity and scientific stagnation is better than learned excellence and scientific progress.” This is the current line of attack against Vas. This logic aims at disgracing Vas and gives no recognition to his contribution to chess whatsoever. This logic of protecting property rights is bogus because while on the surface it makes quasi-legal sense most people know that it is a Trojan horse for other motives and it ignores the history of how knowledge makes progress. How much has any chess programmers paid Alan Turing or Claude Shannon? How much are Toyota, BMW, Landrover etc paying to the family of the guy who invented the internal combustion engine? How much have the electronics and computer companies paid to the family of the guys who invented the transistor? How much are aircraft manufacturers paying the family of the guy who invented the Aircraft? Why aren’t all these cloners being dragged through the hot coals? Because there comes a time, even in the academic world, when innovative use of existing knowledge gets recognized as a new contribution to knowledge. Rybka is now qualitatively and quantitatively different from Fruit. Or are we saying Rybka has not been an innovation? Has computer chess regressed, stagnated or progressed with the advent of Rybka?

This whole Anti-Vas crusade has now become a no-holds-barred assault to an extend where it has gone full circle, which is confusing: Self titled defenders of property rights, originality, ethics and truth; respected longstanding CCC members who say they stand on the side of high programming standards and progress in chess programming, are in open alliance, debating in lock step and basing their judgment on the work of hackers, reverse engineers, and self confessed cloners. Whence cometh this new unity? Gentlemen, fruit code is still out there as open source, remember? Let’s go read it and tinker with it, who knows, maybe in less time than it takes monkeys hammering at a type writer to come out with what I have written here, we will come out with a program stronger than Rybka. Nah, forget it, we are too knowledgeable to learn anything from fruit, our sole interest is to investigate GPL infringements and enforce GPL rules as an end in itself. We want vas to make his programs open source code first and then we will prove our case later, see? Sensibly Fabien who is alleged to have been wronged by Vas has not embroiled himself in this campaign. Reasonably so because he has already been publicly honored by Vas for his contribution to chess programming and, from a chess programming advancement stand point, he is probably more proud of Vas as his student than he is of Vas’s vehement critics.

[Graham Banks]

Nice first post and a lot of truths in there I'd imagine.

[Tony Thomas]

Nice first post and a lot of truths in there I'd imagine.

Why not read it?

[Graham Banks]

Nice first post and a lot of truths in there I'd imagine.

Why not read it?

Obviously I read it. What a strange thing for you to say.

[Tony]

And again a new account is created.

I tought some people weren't able to understand this stuff, but it seems it's on purpose.

Tony

[Sylwy]

Agreed ! 100% !

You tiked off very well all sides !

But don't wait for a reply ! Your post isn't welcome here ! CCC is an anti-Rybka territory.

Don't be angry. Vas is an intelligent man.He don't pay any attention to these persons.He knows better how to do. Vas rules !



Regards,
Silvian

[Tony Thomas]

And again a new account is created.

I tought some people weren't able to understand this stuff, but it seems it's on purpose.

Tony

True that it's a new account, but the name is familiar.

[Tony Thomas]

Nice first post and a lot of truths in there I'd imagine.

Why not read it?

Obviously I read it. What a strange thing for you to say.

Nit pickering.

[Tony Thomas]

Nice first post and a lot of truths in there I'd imagine.

Why not read it?

Obviously I read it. What a strange thing for you to say.

I guess the forum is having problems..double post..

[Terry McCracken]

Nice first post and a lot of truths in there I'd imagine.

You're sure? A lot of truths? You imagine?

It's all conjecture and opinion Graham.

I really find it sad the length people go to prevent the truth from coming out!

Maybe court is the only way to settle this issue afterall?

Very Sad Indeed!

There appears to be something fishy going on here, and it stinks to high heaven!