BB's Rybka/Ippolit comparison

General discussion about computer chess...

Re: BB's Rybka/Ippolit comparison

Postby Sentinel » Sun Jun 13, 2010 7:52 pm

zwegner wrote:It is very impressive, but IMO the most impressive part about it is that BB says he did this without IDA or any other tool, but just by using the output of objdump. Pretty ridiculous.

And I do know who he is (I actually met him in person), but I imagine he wants to remain anonymous. I'll point him towards this forum though so he can maybe participate in this discussion.

Thanks Zach, and thx for providing this report from BB.
Sentinel
 
Posts: 122
Joined: Thu Jun 10, 2010 12:49 am

Re: BB's Rybka/Ippolit comparison

Postby Chris Whittington » Sun Jun 13, 2010 8:00 pm

zwegner wrote:
Chris Whittington wrote:
zwegner wrote:OK, here's what you've all been waiting for: some actual evidence about Rybka and Ippolit. This comes from the man known as BB on the Rybka forum, and he gave me permission to repost here.

It's quite technical and long, but very thorough.


what an astonishing piece of hard and detailed work, congratulations to BB

ok, I read through fairly rapidly, I'm assuming the comparisons and detail are all accurate and honest btw, and the 'fast' conclusion is

1. there's no re-compilation of disassembled code here, too much stuff is different and it looks like the ideas in IPPOLIT are implemented such that the programmer absolutely has to fully understand what he is doing. the many listed differences require far too much skill to leave to a non-understanding re-compilation. there doesn't even seem to be evidence of fragments of re-compiled or stolen material, the closest I read was in piece tables but they're not that close, have substantial differences anyway, and the possible 'close' matches are explainable and probably repeatable in many programs piece tables.

2. there is the possibility that the Rybka code and the IPPOLIT code were written substantially by the same person/team, if I was investigating this, that's were I would put a strong line of questioning and research

3. There is the possibility that the Rybka code and the IPPOLIT code both owe quite a lot to ideas (not code, necessarily) from the same source (using source in its original meaning not source code)

Getting really sneaky and suspicious and off the wall, an explanation could be that Rybka's programmer departed the company and went away and wrote IPPOLIT. That could account for loss of source of R3, accusation of copying and creation of new stronger program. Then again such a scenario could be totally random.


100% agreed. The decompilation story leaves a lot to be desired IMO. I do hope the real story emerges some day, along with the identities of those involved.


More reading, more thoughts ....

assuming BB is honest, this report is enough for a conclusive proof that IPPOLIT is clean w.r.t. Rybka.

A certain moderator left at CCC has some explaining to do to, imo. Perhaps it would be best if he departed computer chess for good given the damage he has wreaked on this and several other fronts through a mix of oafishness, stubbornness and hubris.
User avatar
Chris Whittington
 
Posts: 437
Joined: Wed Jun 09, 2010 6:25 pm

Re: BB's Rybka/Ippolit comparison

Postby DaveD » Sun Jun 13, 2010 8:07 pm

Is there a BB Rybka/Fruit report? :|
User avatar
DaveD
 
Posts: 38
Joined: Thu Jun 10, 2010 3:04 am

Re: BB's Rybka/Ippolit comparison

Postby Sentinel » Sun Jun 13, 2010 8:11 pm

Chris Whittington wrote:assuming BB is honest, this report is enough for a conclusive proof that IPPOLIT is clean w.r.t. Rybka.

If they had anything on him, I'm pretty sure Felix would have started a lynch campaign on him on Rybka forum long time ago.
Just that is enough for me to believe the guy is honest.
Sentinel
 
Posts: 122
Joined: Thu Jun 10, 2010 12:49 am

Re: BB's Rybka/Ippolit comparison

Postby kingliveson » Sun Jun 13, 2010 8:14 pm

Chris Whittington wrote:
More reading, more thoughts ....

assuming BB is honest, this report is enough for a conclusive proof that IPPOLIT is clean w.r.t. Rybka.

A certain moderator left at CCC has some explaining to do to, imo. Perhaps it would be best if he departed computer chess for good given the damage he has wreaked on this and several other fronts through a mix of oafishness, stubbornness and hubris.



From a legal stand point of view, it would seem Vas has no case to make against Ippolit which he himself alluded to earlier in another post. I think that this document, assuming it to be correct on Rybka/Ippolit similarities/differences, it would be within reason to conclude Ippolit author(s) know how Rybka works and used similar ideas. Furthermore, I have an Idea who "BB" might be. Hope to see him here posting soon. We can use his technical expertise.

Franklin
PAWN : Knight >> Bishop >> Rook >>Queen
User avatar
kingliveson
 
Posts: 1388
Joined: Thu Jun 10, 2010 1:22 am
Location: 28°32'1"N 81°22'33"W

Re: BB's Rybka/Ippolit comparison

Postby Chan Rasjid » Sun Jun 13, 2010 8:29 pm

The only thing I can confirm is the article is about chess programming as far as I understand and not voodoo.

The author seem to be talking from the "sources" of Rybka 3; he actually mentioned using " painstaking analysis via debugging tools", whether it means the same as others talk about decompiling.

If it is genuine, then it is a gem of an article for anyone wanting to write an elo 3000 engine.

Rasjid
Chan Rasjid
 
Posts: 33
Joined: Thu Jun 10, 2010 4:41 pm

Re: BB's Rybka/Ippolit comparison

Postby Chan Rasjid » Sun Jun 13, 2010 8:37 pm

Also, with so much differences mentioned, it seems impossible that Ippolite is NOT an original engine. All the while the anti-ippolites like to think only Vasik knows chess programming, which is NOT very true. There are too many experts around. Milos mentioned he could write an elo 2000 engine in a week... and an elo 3000 engine in 3.

Rasjid
Chan Rasjid
 
Posts: 33
Joined: Thu Jun 10, 2010 4:41 pm

Re: BB's Rybka/Ippolit comparison

Postby BTO7 » Sun Jun 13, 2010 8:47 pm

This makes things very clear to me. Took awhile to read but was well worth reading. Ippos are NOT CLONES period. The idea of cut and paste and transferring to a different code always sounded (while possible i guess) dumb. This is obviously not a CLONE. Vas little letter where he down plays the roll of fruit in rybka and says 20% at best looks to be about the only part ippo has the same....the both used some ideas of open source fruit. Fruit is closer to rybka and ippo then ippo and rybka. A clone should be almost identical yet here hardly NOTHING is the same. These two engines do things completely different and just so happen to be close in strength. Vas couple lines about what he thinks he saw ...when he did take a look ..what couple years back? ....says he hasnt looked at the newer versions?? Like he doesnt have time LOL. He should get a academy award for best actor. The differences between the two are substantially different! I think IPPO's should be UNBANNED IMMEDIATELY !!!! Vas has caused them enough damage to their great programs. Take this document and put that in your pipe graham and smoke it. Legal action should be taken by the IPPO team for defamation of character and a Vas apology letter sent immediately with some cash for damages.

Regars
BT
BTO7
 
Posts: 101
Joined: Thu Jun 10, 2010 4:21 am

Re: BB's Rybka/Ippolit comparison

Postby Charles » Sun Jun 13, 2010 8:48 pm

Chan Rasjid wrote:Also, with so much differences mentioned, it seems impossible that Ippolite is NOT an original engine. All the while the anti-ippolites like to think only Vasik knows chess programming, which is NOT very true. There are too many experts around. Milos mentioned he could write an elo 2000 engine in a week... and an elo 3000 engine in 3.

Rasjid



..but couldn't it be an engine developed from studying Rybka 3 code either rev. engineered or from actual source that would account for many of its apparent similarities?
Charles
 
Posts: 71
Joined: Thu Jun 10, 2010 7:41 pm

Re: BB's Rybka/Ippolit comparison

Postby Sentinel » Sun Jun 13, 2010 8:51 pm

Chan Rasjid wrote:and an elo 3000 engine in 3.

Nope that's certainly not true. To get a 3000 elo engine, you would need at least 6 month of tuning with a dedicated cluster... Not to mention all the details of implementation, optimizations and debugging.
For 2000 elo, you almost don't need any tuning, and you can do it even without bitboards which makes things lot easier.
Sentinel
 
Posts: 122
Joined: Thu Jun 10, 2010 12:49 am

PreviousNext

Return to General Topics

Who is online

Users browsing this forum: No registered users and 3 guests