Rybka is no longer strongest chess engine

Rybka on 1 cpu has lost in multiple tests/time controls to Robbolito

http://www.chesslogik.com/robbolito.htm

On multicore it loses to Igorrit 0.086v7

 Anywone interested in computer engine vs engine matches will find this fascinating.  ..

Iggorit/Robbolito with SCID  is a great  free alternative to Chessbase/deep Rybka/Aquarium.

Is anyone here interested in engine vs engine matches?

sounds very interesting

I heard this engine was a rykba clone.

Define clone.

Maybe ...but how is it stronger? And I have seen only a few similarities with rybka ...The blind bishop bug, and the "show mate" bug are fixed... fyi rybka does not show all the moves leading to  mate .,.

I think it probably did originate from decompiled sources but the modifications might  be enough to make that a moot point.

According to this website: http://www.chesslogik.com/robbolito.htm the programmers reverse engineered Rybka, but seeing as they provide both binary and source code downloads they must've written their own engine from scratch (i.e they did not just dissassemble the Rybka binary and start modifying it).

So what makes these two engines (Rybka and Robbolito or whatever it is called) clones of eachother when any two other chess engines that

1. Do not share the same source code

2. Are not equally strong

are clearly not clones?

Anyway I think it is cool how strong engines one can 'make on his own' if you just know some programming. I will definately keep an eye on this 'robbolito' engine :)

If you look at my original post I added that they already have multi-cpu support.

It is being actively tested... if you look

http://immortal223.borda.ru/?1-17-0-00000224-000-340-0

w32 version: rapidshare.de/files/48982733/Igorrit_0086v7_w32.exe.html 

x64 version: rapidshare.de/files/48982758/Igorrit_0086v7_x64.exe.html

I'm currently running the w32 version against rybka 3 ponder off 1m+6 seconds (a much longer tc then what most are testing with) and it is ahead by +32 elo so far.

And people only knew about it recently ...It has been in development for quite a while. 

You also seem to imply that they are hackers and not programmers ...thats odd because only very good programmers can rev. eng a product and then improve on it. (which is what Vas from rybka has said they did without providing any proof)

This is all supposition until its playing power can be proved in actual competition. I'll wait for these lists to change before I believe the author's (likely biased) assertion that this engine is a 3300 and the strongest in the world:

http://wbec-ridderkerk.nl/html/BayesianElo_ed16.htm

http://www.husvankempen.de/nunn/40_40%20Rating%20List/40_40%20All%20Versions/rangliste.html

http://www.computerchess.org.uk/ccrl/4040/rating_list_all.html

The Robbolito source code is available. It doesn't look like decompiled source -- in fact, it's full of comments. In Italian. There is no reason to think that it contains any copied chunks at all.

If you charge someone with a crime you have to provide proof. If the code is already out there, what can you lose by showing parts of the code that is already stolen that is identical?  Hell, any experienced programmer can do that without giving away any secrets the "thieves" didnt get.

You apparently know nothing of programming. You think its easy to  just rev eng an engine and make it +60elo in blitz?

I showed that they have multi-cpu because you claimed otherwise:

>>You come up with the "strongest engine in the world," beating hardcore professional work at the highest level the computer chess scene has ever witnessed that has been going on for years in just a couple of weeks and just can't add multi processor sport, or even book sport? Seriously. >>

My point is even if someone can prove that it is rev. eng.. - the programmers still came up with an innovation from a chess programming point of view.

But make no mistake these are a group of very talented programmers.

the engine has dethroned rybka.. not "on par" with rybka. 

Vas himself can be a cloner.   Programmer generally do this sort of thing a lot. Reliable experts have already charged Rybka for having stealing from open source fruit...(Since it is bound by GPL you cannot steal its code and make a profit )

A competing company could have done the same thing not released the source and no one would say anything.

Chess ratings list will eventually include this engine when it becomes apparant that Vas will never back up his accusation.

You have not defined what a clone is..

1. Rybka still loses in slow time control -20-32 elo ...faster time controls as much as -100 elo.

2. Its not a clone then. Rybka is not "genuine" then.  Parts of fruit code were lifted.

3. seriously -- you are the spammer -

Rybka PWNs all

+63 Elo for  this time control  1min 6 second increment is just a tad slower than the 40/4 control used in ccrl.

Loading 30 balanced positions from mfl modified suite repeat with colors switches.

1: Igorrit_0086v7_w32 35.5/60    59% +63 

1=0=1=00=0=10=1=0==111=1=11=1=0===00===1111=10===11111==01==

2: Rybka 3 w32 (all 4 cores)       24.5/60  41% -63

60 games played / Tournament is finished
Name of the tournament: Igorrit_v7 vs Rybka 4T pond off 1+6

Level: Blitz 1/6 -
Hardware: Intel(R) Core(TM)2 Quad CPU    Q6600  @ 2.40GHz4x  2394 MHz with 3,072 MB Memory

With just a 20-32 ELO difference (assuming this all wasn't a complete fabrication), I highly doubt the engines would not draw at every possible opportunity. 20-32 ELO at the 3200+ rating would not make for a significant win/loss difference. But, as I said, this is assuming this information isn't a complete fabrication, which it most likely is.