Not logged inRybka Chess Community Forum
Up Topic Rybka Support & Discussion / Rybka Discussion / Lets get on with it.
- - By Wayne Lowrance (***) Date 2009-10-22 17:56
Here on rybka forum we tend to be too pro Rybka. Lets accept this new engine
as for what it is' a very strong engine, that Vas needs to compete against.
Lets drop all of this Rybka bias stuff here on the forum. Life can be so much
more pleasant for us all. I am a die hard fan  of the "lil girl" but chess life is
more than Rybka, except of course the Vas team.
Wayne
Parent - - By M ANSARI (*****) Date 2009-10-22 19:44
Accept what new engine?  The engine mentioned is made from stolen code of Rybka 3 and patched up together to get it to run.  Of course it is a strong engine because it is made from Rybka 3, but it is nothing for Rybka to compete against because it is Rybka ... albeit with way too many bugs, no endgame knowledge, no EGTB support and only single processor support.  BUT it understands bishop under promotion, whoooppeeeee !!!!  Whoever is claiming this engine is stronger than Rybka 3 I suggest the following, try a match using zero contempt for Rybka 3 and a time control of 5 0 or more.  I tried that last night on my Octa using a SP Rybka 3 contempt zero, but gave up after the other engine crashed 3 times in 18 games.  By the time I had given up Rybka 3 had 4 more wins if we do not include that in 2 of the games the crashes were losing games for the other engine.  I know 18 games is nothing, I will be the first to admit it, but I am pretty sure that my original hunch is true.  This engine is simply the equivalent of a buggy, incomplete Rybka 3 single processor beta ... nothing more nothing less.  I already have Rybka 3 MP which is not a beta, and although it has the MP EGTB bug, I know how to work around that.  I am more interested in seeing what Rybka 4 has to offer.  The sooner that Vas releases it the happier I will be.
Parent - - By Wayne Lowrance (***) Date 2009-10-22 22:08
Mansari
You do not know it is a  R3 clone.  My last comment on it
Wayne
Parent - By Felix Kling (Gold) Date 2009-10-22 22:21
Is it some "oιδα oυκ ειδως" (I know not to know) diskussion or what's your point?
Parent - - By M ANSARI (*****) Date 2009-10-23 04:56 Edited 2009-10-23 05:07
I am sorry, but you are completely wrong.  I DO know it is a R3 clone.  While it might not be a 100% line by line copy of R3, it was not due to lack of trying.  Whatever result they got, the strength of the program would not have been possible had it not been for the parts that were stolen from Rybka 3 ... in my book that is describes a clone.
Parent - By Quapsel (****) Date 2009-10-23 05:25 Edited 2009-10-23 05:27

> I DO know it is a R3 clone


And I cant't imagine that.
I do believe, that it is a R4beta clone.
For what reasons do you declare, it must be R3?
Quap

PS:
Your argument
"the strength of the program would not have been possible had it not been for the parts that were stolen" is very weak!!
Parent - By Silvian (***) Date 2009-10-23 13:21

> The engine mentioned is made from stolen code of Rybka 3 and patched up together to get it to run.


?????

Something missed me ? Or are some Rybka's extremists here ?

Silvian
Parent - - By FWCC (***) Date 2009-10-23 14:24 Edited 2009-10-23 19:46
I must say that after further reflection and looking at the results of the matches I have created and reading the facts on [Deleted],here and the like I DO believe that the other engine is made from Rybka's code.The mystery and excitement of a potetial "Rybka Killer" has declined,Vas did indeed know all about this.So we await Rybka 4 and the sooner she is out the better.It would be nice to have Her as a Christmas gift.
Parent - - By Nelson Hernandez (Gold) Date 2009-10-23 16:13
Rybka out of danger?
Parent - - By Capa (***) Date 2009-10-23 21:35
Isn't it common knowledge that Rybka was essentially a Fruit clone?  I'm baffled by all of the self-righteousness. If Rybka can be begotten via cloning, why can't its offspring as well?
Parent - - By Uly (Gold) Date 2009-10-23 21:37

> Isn't it common knowledge that Rybka was essentially a Fruit clone?


It isn't.
Parent - - By Capa (***) Date 2009-10-23 21:38
It isn't common knowledge or it isn't true at all?
Parent - - By Uly (Gold) Date 2009-10-23 21:41
It isn't true, thus it can't be knowledge at all.
Parent - - By Capa (***) Date 2009-10-23 21:55 Edited 2009-10-23 22:22
How do you know it isn't true?  When Rybka 1st came out there was a lot of discussion that it was "somewhat" based off Fruit, just like this new program is "based " off Rybka.  Here is an interview in 2005 where Alexander Schmidt asks Vas:

We had our first contact when I had questions about a similarity to Fruit in the search, others found similarities in the evaluation. Some people were a little bit suspicious that Rybka could be a clone of the open source engine. In the meantime it is clear that Rybka is no clone but you used ideas of Fruit (I guess all other serious engine programmers had a look at Fruit too). How strong would Rybka actually be if the Fruit code would have never been published?

Here is the link to the whole interview which some may find interesting:

http://www.superchessengine.com/vasik_rajlich.htm

As can be seen from the interview, Vas admits (but not to what degree) that Rybka was loosely based off Fruit.
Parent - - By Nelson Hernandez (Gold) Date 2009-10-23 22:57
I think you have a point: programmers looked at Fruit code and used it each in their own way.  But I think it should be obvious that the search and evaluation features (just two name two things) have probably gone through literally hundreds of iterations and improvements since Rybka 1.0 Beta.  Fruit was open source; this clone was flat-out stolen.  Rybka 3 contains a ton of original ideas that Vas came up with; the clone contains virtually no new ideas apart from a few trivial headlined changes (bishop underpromotion, which comes up maybe once in five million actual games).

We stand on the shoulders of giants, yet we can see farther than they.  Except these guys cannot see farther than Rybka 3 even after breaking and entering (how is this any different than some punks busting into your home and rifling through your computer, then posting anonymous taunts on the Internet?).
Parent - - By Capa (***) Date 2009-10-23 23:29
I've missed a lot of posts on this discussion because I was on Fall break and haven't had time to read most.  Has it been conclusively proven that this program is a Rybka clone?  If so, and then it is definitely unethical and they should be prosecuted and banned from any official tournaments.

However, what I was trying to get at was that if this program is "loosely" based off Rybka 3 and then how is that different from Rybka 1 being "loosely" based off Fruit?

This also raises the issue of how much reverse engineering (similar search, eval, etc) is allowable before it is rightfully and lawfully considered a clone.
Parent - - By Nelson Hernandez (Gold) Date 2009-10-24 00:31
If Vas says it is a clone (see his posts on the subject), it's a clone.  Your premise is that is it "loosely based on Rybka" is hardly the case.  I agree with your final point, which I myself have made on a different thread.  I don't think there is a clear-cut answer.
Parent - By Uri Blass (*****) Date 2009-10-24 03:38
There are other reasons to believe that the engine is a clone but no proof for it that I know, and
Vas says is certainly no proof(if you consider Vas says as a proof you may believe that rybka is number 1 even in the case that some original engine become better than rybka because Vas can claim that the original engine is a clone).

Uri
Parent - By Uri Blass (*****) Date 2009-10-24 03:29
I doubt if the clone contains virtually no new ideas.

Based on my knowledge evaluation of the clone is different from rybka3 and the analysis of the clone is different than the analysis of rybka3 not only in some simple endgames.

I also know no evidence except trusting other people that the new engine is a clone of rybka but I choose to believe the people that I believe.

Uri
Parent - By George Tsavdaris (****) Date 2009-10-24 09:27

>Rybka 3 contains a ton of original ideas that Vas came up with; the clone contains virtually no new ideas apart from a few trivial >headlined changes (bishop underpromotion, which comes up maybe once in five million actual games).


It seems to have some endgame knowledge also that Rybka misses:
http://rybkaforum.net/cgi-bin/rybkaforum/topic_show.pl?pid=191549#pid191549

It seems also to have an evaluation very different from Rybka's. Some times much more powerful by thinking immediately moves that R3 needs a lot of time to think, and some times much more dubious by playing stupid moves that R3 would reject immediately.

>Except these guys cannot see farther than Rybka 3 even after breaking and entering (how is this any different than some punks busting >into your home and rifling through your computer, then posting anonymous taunts on the Internet?).


Perhaps they can see if you give them more time. I'm already surprised how they did what they did in such a short time from the release of Rybka 3.
Parent - By George Tsavdaris (****) Date 2009-10-24 09:30

>As can be seen from the interview, Vas admits (but not to what degree) that Rybka was loosely based off Fruit.


Maybe but this is no problem at all, if he just took the ideas of it and put them on Rybka.
Parent - By Roland Rösler (****) Date 2009-10-24 01:03
It´s common knowledge, current Ferrari is clone of Ford's Model T!
Parent - By Miroslav Kvíčala (**) Date 2009-10-24 06:42
I have already tested it. Well, let me conclude my impressions. First I don´t think that it is stronger than Rybka 3 even I tested versions a to b. If you are interested if it is R3 clone or not. It is hard to say because some positions proposed moves and evaluation is almost the same, but there are also positions with very different evaluation (not only in endgame). More important is that it changes proposed moves very often (typical for all non-Rybka engines) and rapidly changes evaluation on the same depth or between two depths. Moreover using of hash tables is completly different (it doesn´t have R3 bug with losing transpositions or pawn pushing to h6 - as white against black´s structure f7g6h7). Variety of proposed (after reverse position set up) moves is very wide (typical e.g. for Shredder). Generally said my impression is that it is Fruit based engine. I have absolutly no knowledge about chess engines programming (I am happy that I am mastering my Windows Vista  :-) ) but I have many years experience with chess engines handling and I frankly said I am not sure that it is R3 clone...
Parent - By Nelson Hernandez (Gold) Date 2009-10-22 23:24

> Here on rybka forum we tend to be too pro Rybka.


If this is a RYKBA forum, would it not make sense for us to TEND to be pro Rybka?  Do people write in product-centric fora to pillory the product?  OF COURSE we are biased.  That's why we are here and not in the Hiarcs forum or the Chessmaster forum, hmm?
- By billyraybar (***) Date 2009-10-22 18:05
Wayne,

I agree.  The real excitement will be when Rybka 4 is released.  I can't wait and of course will buy it immediately.
- - By Uly (Gold) Date 2009-10-23 00:46
Rybka has always been Rybka's main competitor.
Parent - - By Uri Blass (*****) Date 2009-10-23 06:07
Yes and I guess that rybka will never be able to beat this competitor.

Rybka4 will also not score more than 50% against rybka4.

Uri
Parent - - By Quapsel (****) Date 2009-10-23 12:29
If you will let play Rybka4 against Rybka4, you will get with a very good likelyhood one Rybka4, that has scored better than 50%
;-D
Quap
Parent - By M ANSARI (*****) Date 2009-10-23 15:52
Yes, that is very true.  Sometimes the result can be quite surprising even with identical engines.  However the more matches that are played the closer it will reach to 50%.
Parent - By h1a8 (***) Date 2009-10-23 21:47
Yes but the other engine will score under 50% and thus Rybka4 has not scored better than 50%.
Parent - - By Uly (Gold) Date 2009-10-23 19:47
Rybka as white has always beaten Rybka as black.
Parent - - By Uri Blass (*****) Date 2009-10-23 20:53
I am not sure about it

I think that it is dependent on the opening book and with the right opening book (that is based on opening lines that black is winning) rybka black is going to win.
Parent - - By Uly (Gold) Date 2009-10-23 21:37
It always depends on the opening book, if you give Rybka a book that plays poor variations and Zappa a book tuned against the first book Zappa is going to win, that's not the point, I assume neutral books.
Parent - - By Uri Blass (*****) Date 2009-10-24 03:05
I assumed only the same book for both sides.

Note that zappa can win against rybka even in these conditions because the same book may lead to positions that zappa knows to play better but in the case of rybka white against rybka black the reason  is different.

Uri
Parent - By Uly (Gold) Date 2009-10-24 05:56
But that is true for all engines on all contexts, it doesn't change the fact that under normal circumstances Rybka scores better with white.
Parent - By Roland Rösler (****) Date 2009-10-24 01:15
But Rybka 5 will crush Rybka 4! And she will play endgames better than our human legends (Capa, Smyslov, Karpov, Kortchnoi, Salov, Carlsen)! :-)

PS: Today Kortchnoi lost an (rook) endgame, which R3 would have hold (?). Okay, Rybka isn´t 78 years old! :-)
Up Topic Rybka Support & Discussion / Rybka Discussion / Lets get on with it.

Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill