Arimaa Forum - Print Page


    
      
        Arimaa Forum
        (http://arimaa.com/arimaa/forum/cgi/YaBB.cgi)
      

        Arimaa >> Off Topic Discussion >> Rybka Chess Program [Maybe] Didn't Cheat (Article)
        
(Message started by: SpeedRazor on Jan 3^rd, 2012, 6:50am)

Title: Rybka Chess Program [Maybe] Didn't Cheat (Article)
Post by SpeedRazor on Jan 3^rd, 2012, 6:50am

Rybka might not have been a clone of Fruit afterall; but the jury seems to be still out if Houdini is a clone of Rybka. (As mentioned by Omar in an earlier thread.)

Part 1: http://chessbase.com/newsdetail.asp?newsid=7791

I thought that that was strange: how could the clone be so superior to the original? (Over 100 Elo to start; about 400 a few years later...) If vindicated, I hope the programmer gets his four world ICGA titles back.

Title: Re: Rybka Chess Program Maybe Didn't Cheat (Articl
Post by Fritzlein on Jan 4^th, 2012, 10:52am

Bah, this long-winded and repetitive article provides no evidence of anything. The entire argument is deferred to part II. Of course I, too, would like Rajlich to be vindicated if he is innocent, but I see no purpose whatsoever in releasing a rebuttal like this in two parts.

Title: Re: Rybka Chess Program Maybe Didn't Cheat (Articl
Post by jdb on Jan 4^th, 2012, 2:04pm

on 01/04/12 at 10:52:33, Fritzlein wrote:

Parts two and three are on the chessbase website.

Title: Re: Rybka Chess Program Maybe Didn't Cheat (Articl
Post by SpeedRazor on Jan 4^th, 2012, 6:17pm

These are the final three parts:

Part 2: http://chessbase.com/newsdetail.asp?newsid=7807

Part 3: http://www.chessbase.com/newsdetail.asp?newsid=7811

Part 4: http://www.chessbase.com/newsdetail.asp?newsid=7813

Title: Re: Rybka Chess Program Maybe Didn't Cheat (Articl
Post by Fritzlein on Jan 5^th, 2012, 10:55am

Thanks for the links! I'll check it out...

Title: Re: Rybka Chess Program [Maybe] Didn't Cheat (Arti
Post by omar on Jan 8^th, 2012, 1:04am

Even if someone started with the source code of another strong engine, making software changes that improve the engine by even 50 points is very difficult.

Title: Re: Rybka Chess Program [Maybe] Didn't Cheat (Arti
Post by NickBentley on Jan 8^th, 2012, 5:33pm

What Omar said,

The most interesting programming I've ever seen takes place in the Matlab scripting contests where all code is public and you're encouraged to steal and improve all the other contestants' code. After seeing what happens in those contests I've become a pretty strong advocate for openness.

Title: Re: Rybka Chess Program [Maybe] Didn't Cheat (Arti
Post by rbarreira on Jan 10^th, 2012, 4:54am

And now a retort to the Chessbase article comes out:

http://harveywilliamson.com/ICGA/Riis3.pdf

Title: Re: Rybka Chess Program [Maybe] Didn't Cheat (Arti
Post by ocmiente on Jan 10^th, 2012, 9:41am

16 pages worth. Seems like an emotional response. The ratio of substantial rebuttal to name calling is too low for my tastes.

Interesting situation in any case. Thanks for the link.

Title: Re: Rybka Chess Program [Maybe] Didn't Cheat (Arti
Post by Fritzlein on Jan 10^th, 2012, 9:55am

on 01/10/12 at 09:41:34, ocmiente wrote:

The ratio of substantial rebuttal to name calling is too low for my tastes.

This is true, and it annoys me as well. But what about the Riis article? The ratio of substance to blathering in that piece is also very low.

Title: Re: Rybka Chess Program [Maybe] Didn't Cheat (Arti
Post by ocmiente on Jan 10^th, 2012, 10:12am

on 01/10/12 at 09:55:59, Fritzlein wrote:

...what about the Riis article? The ratio of substance to blathering in that piece is also very low.

Agreed.

The Riis article was more entertaining though. It had the whole conspiracy theory thing going for it, which I enjoyed reading. This latest salvo needs a more compelling script if they want to get the readers on their side.

Also, the rebuttal feeds into the Riis article. At 16 pages, we can see the beginning of the flood Riis described - the beginnings of an overwhelming group dog pile that the author of Rybka could not have overcome had he challenged the ruling against him.

Someone needs to buy the movie rights.

Another link:
"Another Conversation with Vasik Rajlich" By Nelson Hernandez (on Rybka chess) (http://www.youtube.com/watch?v=cQshTNJ4pSM)

Title: Re: Rybka Chess Program [Maybe] Didn't Cheat (Arti
Post by Hippo on Jan 10^th, 2012, 2:26pm

The main argument against Rajlich (for me) was comparision of a part of the Fruit code and corresponding fake of Rybka's code ... let us see how the cause would end.

Title: Re: Rybka Chess Program [Maybe] Didn't Cheat (Arti
Post by jdb on Jan 11^th, 2012, 9:05am

For those still following the saga,

A NEUTRAL summary was posted by someone:

http://www.talkchess.com/forum/viewtopic.php?topic_view=threads&p=442916&t=41913