RobboLito vs Rybka 3


RobboLito vs. Rybka 3

Back to forum

Thibault de Vassal    (2009-12-29)
RobboLito vs. Rybka 3

What are RobboLito, Ippolit and Igorrit ? It looks like these names are invading computer chess forums... As you may have read in the discussion mentioned below, the Rybka 3's source code may have been compromised and these engines "may" be clones of Rybka 3 (everyone does not agree on this). Good or really bad news, anyway this open source chess engine may have many consequences on the computer chess world, and correspondence chess as well...

From the wikispace mentioned below :

IPPOLIT : Intellectual Persons Promotion Of Leninist International Tradesunions (!??)

Q. What is RobboLito?
A. RobboLito is the version of IPPOLIT that now contains endgame tablebases, the RobboBases.

A few links on RobboLito 0.085f1a, Ippolit & Igorrit (says it all IMO) :

see also :
-> IPPOLIT banned from chessprogramming wiki!
PlayChess banned IPPOLIT from use online also.

"RobboLito is an open-source UCI chess engine by: Yakov Petrovich Golyadkin, Igor Igorovich Igoronov, and Roberto Pescatore."

RobboLito does not yet support: multiPV, own book, egbbs, tablebases, multiple CPUs/cores, chess variants

Estimated rating: ~ 3300 ELO

Available versions:

RobboLito 0.085g3 w32 (optimized windows 32-bit executable and source code)
RobboLito_0.085g3_w32_no_SSE2 (optimized windows 32-bit executable - for older CPUs that don't support Intel SSE2 instruction set)
RobboLito 0.085g3_x64 (fast windows 64-bit PGO executable compiled by peterpan)
RobboLito 0.085g3l_x86 (optimized linux 64-bit executable and source code ported to linux by unisky) (french)

Wayne Lowrance    (2009-12-29 19:58:13)

Wow, I cannot keep up with all the improved downloadable engines. LobboLite I do not have and was not aware of it, before it was mentioned in response to my earlier post. gracious :) CC chess is the better for it. It gives latitude for Centaur personality/engine team work. So maybe CC chess will not be so determinate in the near future.

Thibault de Vassal    (2009-12-30 00:39:48)
RobboLito NOT a clone of Rybka ?

According to Tony Thomas, RobboLito & Ippolit wouldn't be a clone of Rybka 3... and it would be 50 to 100 points stronger!

William Taylor    (2009-12-30 11:38:30)

There are lots of discussions on talkchess about whether or not RobboLito is a Rybka clone (some of which are made longer by the fact that not everyone sticks to the topic). As far as I can tell there is no clear consensus - perhaps most are saying that until someone provides clear evidence that Robbo is a clone it should be assumed not to be, but some are adamant that it is a clone (either of Rybka 3, or of something closer to Rybka 4 - leaked code). Vas says it is a clone, and that the hackers have even informed him of their progress via e-mail, but he has apparently not provided evidence for his claim as he did for the Strelka case (Strelka was a Rybka clone). Many programmers on the forum say that they cannot see any evidence that Robbo is a Rybka clone by looking at the source code, and it often gives quite different evaluations.

Nick Burrows    (2009-12-30 16:17:12)
Rybka Clones.

It has to be doesn't it?

How can a random programmer out of nowhere produce a program that compete with money invested and the years of highly specialised knowledge that is the Rybka project?

Sebastien Benoit    (2009-12-30 17:22:55)
On reverse enginerring

I'v read some of the arguments for and against the theory that RobboLito is a Rybka clone and I must say that I'm more inclined to beleive it is not. But for those who think that the reverse engineering of a software as complex as Rybka is not possible... sorry but it is possible! A team ( don't know how many guys would be needed) of good computer scientists would certainly be able to reverse engineer enough of it to get a global picture of what kind of optimisations it has. And remember that computer chess is not new and a lot of knowledge is available trough books and of course, the web.

Some are claiming that a reverse engineered clone of Rybka would have the same level of functionality as the original. They say that, because they don't really understand the process. Reverse engineering doesn't necessarily lead to an approximate clone of the origninal source code. It is more a process of understanding how it works. Once you got this understanding you can write from scratch and add incrementally whatever functionnality you desire.

Wayne Lowrance    (2009-12-30 19:10:08)
Rybka clone

Nick I think you have hit the nail on the head. Yes that is suspicious. Commercial programmers of programs like shreddar, Fritz, Junior and Zappa, etc have worked a long time and not come to Rybka playing floor yet. So it would seem unreasonable these recent free programs have made the jump mostly right at the start. A huge compliment to Vas I would say

William Taylor    (2009-12-30 22:17:44)

Stockfish is not a commercial program, and has only been released relatively recently (I think), but it is already at about the level of Rybka. I know it's based on Glaurung, which has had years of work put into it, but perhaps Robbo is also based on a strong open source program. I'm not saying Robbo's not a Rybka clone - very likely it is - I'm just pointing out that it is possible for free, open source programs to approach or surpass the level of commercial ones.

On another note, for anyone interested I ran a quick 12xRR 2 2 tournament today with Rybka 3 32-bit, Stockfish 1.6 JA and RobboLito 0.085g3 w32. Rybka scored 13/24, Stockfish 12/24, and Robbo 11/24. Of course, the time control was very quick and I'm not putting this result forward as a serious test, but it seems that the 3 (or 2) may be close enough to be competitive. An interesting time for computer chess.

Sebastien Benoit    (2009-12-30 22:57:24)
On open source

William, you are right. Open source model has produced very nice software as the linux kernel, just to name one. Why not a chess engine? Crafty despite its age and relative weakness is not a bad example... The knowledge is there to anyone interested.

Rodolfo d Ettorre    (2010-01-01 02:50:25)
Open Source ...

Really I think a commercial program like Rybka will be very likely the strongest not because of the software developers working on ir, but because they can get some tips from open source programs, like Robbolito.

Thibault de Vassal    (2010-02-10 23:33:26)
Download RobboLito

Here is the link to download RobboLito at FICGS, I'll add Igorrit & Ippolit when I'll find some time...

Rodolfo d Ettorre    (2010-02-13 15:04:22)
RobboLito vs. Rybka 3

Some facts or fictions:

Robbolito is a clone of Rybka
Robbolito is open source


Rybka is now open source!

Francesco Romeo    (2010-04-18 07:26:55)
RobboLito vs. Rybka 3

download RobboLito vs. Rybka 3 games in pgn

Benjamin Block    (2010-04-19 16:38:36)
RobboLito vs. Rybka 3

I prefer Rybka 3 in corr chess because i can only show 1 the best move with RobboLito:S. which make it really hard...