Complaints against the Chess programs LOOP and THINKER

[hgm]

On the ICGA home page ( http://www.icga.org ), about in the middle vertically in the section 'Results', there are links Day1 to Day7, which each lead to a slide show of many pictures. I just selected two pictures (one from Day1, the other from Day2) that gave an overview of the entire game rooms. (Many photographs are close-ups of individual tables.)

[Vinvin]

Last year's ICGA event in Yokohama


And the other room

Thanks for these photos,
know where to get more?
One link?.

Thanks.

http://icga.uvt.nl/?page_id=780

[kranium]

The second appears to derive from IvanHoe (thus partially Rybka 3).

The only Rybka vs Ippolit comparision that exists is BB's excellent analysis here:
http://www.open-chess.org/viewtopic.php?f=3&t=119
and instead of demonstrating that they are in any way related, it clearly show how completely different they are.

Here are reactions from just 'some' of the programmers that have read and understood it:

Chan Rasjid: "with so much differences mentioned, it seems impossible that Ippolite is NOT an original engine."
BT07: "This makes things very clear to me. Took awhile to read but was well worth reading. Ippos are NOT CLONES period."
Matthias Gemuh: "Incredibly detailed proof that Ippolit is not a reverse-engineered Rybka !"

Very disturbing and irresponsible IMO that Mr. Levy sees fit to publicly make such unfounded accusations.

Besides enormous amounts of innuendo, frequent statements such as "anyone with half a brain knows it's true", and a concerted campaign by many here to discredit anything related to 'Ippolit' in any way, as a clone...
not 1 shred of evidence demonstrating a close Rybka/Ippolit relationship has been presented anywhere, (not even 1 small byte of code comparison).

If Mr. Levy has such evidence, I call on him to present it, instead of adding to the injustice by publicly convicting anonymous authors (whom he knows won't defend themselves) 'in absentia', and without evidence of any kind.

Norm

[bob]

If this is a way to show to the world icga is still alive, well, it is a pity way....but as US people say "if all you have is a hammer, everything looks like a nail"

The authors even didn't bother answering icga, I can understand them, I would have not answered myself too: I would have not allowed my name to be used in an instrumental way just to advertise a dead association.

The ONLY reason to "not answer" is having something to hide. The rules are clear.

[bob]

Is it April 1 ??? I really think that David Levy lives in his own little cocoon. Does he not realize that the ICGA is a defunct entity? I have seen strange methods by people trying to get publicity ... for example Farrah Abraham's better known as "teen mom" when she did her porn sex tape!

[Link removed by the moderation team]

Good luck David, try a sex tape ... you might get better results getting publicity.


PS. guys you should be ashamed of yourselves if you went ahead and searched the sex tape link.

You DO realize that there was a formal complaint filed about Loop? You DO realize that Ed has been haranguing the ICGA wanting a decision on Loop?

Or, more likely, you don't have a clue about any of this and are just posting to post...

[Richard Allbert]

HG's pics say it all.

Nice event, nice to see people meeting each other otb, which is a great experience.

The rest is pointless, pontificating and arguing about clones - the Houdini episode shows how much influence the ICGA has.

The main achievement was generating free publicity for it, and now it's doing nicely on the Chessbase.com front page.

[Gerd Isenberg]

Oups, I first confused Harvey's post and link to the ICGA article with the two years old one, reporting the formal complaint by Fabien Letouzey. Now, more than three years after Mark Watkins's post on OpenChess, it was time for the ICGA to force some decision concerning rule 2.

[bob]

The second appears to derive from IvanHoe (thus partially Rybka 3).

The only Rybka vs Ippolit comparision that exists is BB's excellent analysis here:
http://www.open-chess.org/viewtopic.php?f=3&t=119
and instead of demonstrating that they are in any way related, it clearly show how completely different they are.

Here are reactions from just 'some' of the programmers that have read and understood it:

Chan Rasjid: "with so much differences mentioned, it seems impossible that Ippolite is NOT an original engine."
BT07: "This makes things very clear to me. Took awhile to read but was well worth reading. Ippos are NOT CLONES period."
Matthias Gemuh: "Incredibly detailed proof that Ippolit is not a reverse-engineered Rybka !"

Very disturbing and irresponsible IMO that Mr. Levy sees fit to publicly make such unfounded accusations.

Besides enormous amounts of innuendo, frequent statements such as "anyone with half a brain knows it's true", and a concerted campaign by many here to discredit anything related to 'Ippolit' in any way, as a clone...
not 1 shred of evidence demonstrating a close Rybka/Ippolit relationship has been presented anywhere, (not even 1 small byte of code comparison).

If Mr. Levy has such evidence, I call on him to present it, instead of adding to the injustice by publicly convicting anonymous authors (whom he knows won't defend themselves) 'in absentia', and without evidence of any kind.

Norm

I do not believe that ANYONE with software programming experience can look at ippolit source and conclude anything other than "This is clearly a reverse-engineered program based on the coding style that is completely unlike anything a human would use."

That's a completely different issue from "is it a Rybka 3 derivative?" I've not seen any credible evidence to show where it came from. But the source itself is credible evidence that ippolit is NOT "original code."

And with 100% certainty, Houdini is a derivative of ippolit, so it is absolutely not an original work.

More than that is conjecture. But ippolit is NOT "original code". It looks EXACTLY like something produced by hex rays or another binary/asm-to-C translator program. I worked on such a program a LONG time back, and my experience was, it is not so hard to turn asm into the language of your choice (we were doing IBM /360 asm to FORTRAN). But converting it into a form that actually comes pretty close to the original FORTRAN is VERY difficult. Inlining by itself presents a daunting task. Did the programmer replicate the code several times or did he use a procedure the compiler inlined? How can you know? De-inlining is a HARD computational process. And that is just one facet of the problem. Even simple things become hard when you look at all the different ways a single asm loop could have been written in the high-level language. And all you have to look at is the final output from a compiler that inlined where appropriate, simplified wherever it could, moved code around to make it more efficient, folded constants, eliminated common subexpressions, reordered instructions, etc., etc., etc. About the best you can hope for is something like hex rays that will at least take the asm back to functional C source code that actually works. Even though it looks like hell.

And that is exactly what IPPOLIT looks like. Some of the variable initialization looks like it had to have been done using massive macro substitutions. reverse-engineering that is simply impractical and pointless as there is no way to know which of the N different ways it could have been done represents the way it actually was done.

None of those are original, even though the original program is unknown. Any other explanation (original, etc) are simply way too implausible to gain any traction with anyone that writes code.

Don't look at it through the eyes "OK, what was this copied from?". Instead look at it with "could a human actually write and then debug code written like this?"

[kranium]

The second appears to derive from IvanHoe (thus partially Rybka 3).

not 1 shred of evidence demonstrating a close Rybka/Ippolit relationship has been presented anywhere, (not even 1 small byte of code comparison).

I've not seen any credible evidence to show where it (Ippolit) came from.

Bob,
yes, apparently we agree completely...

[kranium]

The second appears to derive from IvanHoe (thus partially Rybka 3).

but it is kinda very odd, don't you think?

Mark Watkins (BB) is the one advising (presenting evidence to) Mr. Levy in these new Loop/Thinker 'clone' matters,
as well as previously, in the Fruit/Rybka case:
http://www.open-chess.org/viewtopic.php?f=5&t=1772

BB is also the one who wrote the Ippolit/Rybka document, which almost all agree, shows no connection between Rybka and Ippolit:
http://www.open-chess.org/viewtopic.php?f=3&t=119

Is Mr. Levy simply unaware of Mark's previous work regarding Ippolit/Rybka?
I hope he's not forming and publishing opinions based simply on what he's heard through the misinformation grapevine...

or does someone else have his ear in these matters?