Not logged inRybka Chess Community Forum
Up Topic The Rybka Lounge / Computer Chess / SCCT Engine Test – Without Books
- - By Sedat Canbaz (*****) Date 2012-06-08 11:49
Hello Dear Friends,

I have run a new testing-played by Fruit 2.0 vs Engines (released around 2004 year)

Some Notes (if we test the engines without books):
http://www.sedatcanbaz.com/chess/scct-no-books/

1)Fruit 2.0 test (no books):
•Fruit 2.0 and Zappa 1.0 (when Zappa is with Black pieces) have similarities in openings
•Fruit 2.0 vs Fruit 2.0, both same versions have identical opening playing style
•The rest engine openings are almost different than Fruit 2.0 opening playing style
Games:http://www.sedatcanbaz.com/chess/games/Fruit20test.rar

2)SCCT SSE/Non-SSE test (no books):
•There are some similarities in openings, played by the Top chess engines:
- Houdini / Strelka / Komodo / Rybka / Ivanhoe / Critter / Stockfish
http://www.sedatcanbaz.com/chess/scct-sse-no-sse/

3)SCCT 3 Type Tournament (no books):
•Some engines (Rybka;Glaurung;Thinker;Bright;Toga;Loop…) have similarities in openings too
http://www.sedatcanbaz.com/chess/sccts-3-type-tournaments/

More Details:
•It seems many of the above engines are based on similar opening-related code
•These testings are done mainly for different purposes (not for proving which engine is a clone)
•Once more we noticed that many chess engines are each other relative
•Mostly of the played openings are far away from the current strong opening theory



Best Wishes,
Sedat Canbaz
Parent - - By rocket (****) Date 2012-06-09 21:32

>"It seems many of the above engines are based on similar opening-related code"<


There's no such thing as opening-related code. They are simply clones of each other and then tweaked(sometimes radically different) according to preference.
Parent - - By Sedat Canbaz (*****) Date 2012-06-09 22:34 Edited 2012-06-09 22:36

>There's no such thing as opening-related code. They are simply clones of each other and then tweaked(sometimes radically different) according to preference.


Are you a engine programmer ?

I can't say:mostly of the engines are clones,but i strongly believe that nowadays its hard to find 100 % original work
And in my opinion,many of the engines are each other relative, at least these testings (without books) confirms ...

Btw,you can find more information about opening-related code (see the message by Edmund Moshammer):
http://www.talkchess.com/forum/viewtopic.php?t=41764&highlight=sedat

Best,
Sedat
Parent - - By rocket (****) Date 2012-06-10 03:44 Edited 2012-06-10 03:54
I don't know the exact definition of clone but most people tell me that there is stuff from Fruit in almost all engines. They may have at one point been defined as clones.
Parent - By Banned for Life (Gold) Date 2012-06-10 06:09
Not sure about all engines, but there is certainly stuff from Fruit in Crafty! :lol:
Parent - - By Sedat Canbaz (*****) Date 2012-06-10 09:03 Edited 2012-06-10 09:34

>I don't know the exact definition of clone but most people tell me that there is stuff from Fruit in almost all engines. They may have at one point been defined as clones.


In my understanding (for chess),i can call a clone engine:without any significant improvement,just modifying or changing some values or renaming the exe file

But in case of +50 Elo improvement over the original engine (e.g X engine based on free open source) its clearly not a clone engine

Actually i am not engine programmer,but my feelings say its very very hard to add +50 Elo improvement over the free open source engine

One thing more,really i am very disappointed regarding new commercial engine releases,in case of no any significant Elo improvement than previous version

About Fruit-Rybka issue,
Hmm...It will be a BIG mistake,if we say 'Rybka is a clone'
Just i'd like to mention that Rybka 4.1 is approx.250 Elo stronger than Fruit 090705
And its clear proof that Rybka has own brilliant ideas,that's why this monster was unbeatable almost 6 years

Some notes about open letter to the ICGA regarding Rybka-Fruit issue,
In shortly,the opponents of Rybka preferred the easiest and cheapest way, well-done to them !!
Ok..i can understand Fabien,but why the rest programmers are against Rybka ?
As far as i know Rybka-Fruit issue is not their problem ?!
Or maybe,they complained... due to Rybka includes stuff/ideas from theirs engines ?

For more details:
http://www.talkchess.com/forum/viewtopic.php?t=43946&start=80

Clone Definition:

1   [noun] a person who is almost identical to another
    Synonyms: ringer, dead ringer

2   [noun] a group of genetically identical cells or organisms derived from a single cell or individual by some kind of asexual reproduction
    Synonyms: clon

3   [noun] an unauthorized copy or imitation

    Synonyms: knockoff
4   [verb] make multiple identical copies of; "people can clone a sheep nowadays"


Best,
Sedat
Parent - - By rocket (****) Date 2012-06-10 11:53

>"Actually i am not engine programmer,but my feelings say its very very hard to add +50 Elo improvement over the free open source engine"<


That's a myth. Fruit was improved more than 100 points with Toga.

>"And its clear proof that Rybka has own brilliant ideas,that's why this monster was unbeatable almost 6 years"<


Sure... but aloth of it is Fruit, you can still see traces from that engine in very specific positions despite the big elo leap.
Parent - - By Banned for Life (Gold) Date 2012-06-10 11:59
Nonsense.
Parent - - By Homayoun_Sohrabi_M.D. (***) Date 2012-06-10 12:28
He (rocket) said the same thing 4 months ago and Houdart shot him down.   I'd like rocket for once to provide us with some of those very specific positions where he can supposedly see traces of fruit.
Parent - - By Banned for Life (Gold) Date 2012-06-10 16:10
you can still see traces from that engine in very specific positions despite the big elo leap.

This is a silly statement. With the huge number of chess positions, you could certainly go through a huge number and find a few positions where any two engines both liked an unusual move. Or you could do something more meaningful, like looking at dendograms, to see that the two engines' move choices aren't that similar.
Parent - By Homayoun_Sohrabi_M.D. (***) Date 2012-06-10 16:12
correct.
Parent - - By rocket (****) Date 2012-06-10 20:34 Edited 2012-06-10 20:38
I gave Houdart a concrete example and motivation behind it and never heard anything more from apart  from "statistical possibility" nonsense when both engines commit the same non-opening book misstakes and same suboptimal line.

If it looks like a duck, walks like a duck  and quacks like a duck chances are it's a duck.
Parent - - By Homayoun_Sohrabi_M.D. (***) Date 2012-06-10 21:10
I am referring to your conversation with Houdart from end of February.   It was a very interesting one, let's see it again.

you say:  The basis or call it philosophy from fruits evaluation is still in use by most but not all top engines.

Houdart says:  Can you provide any evidence to support this claim?
What exactly is, according to you, the "philosophy from Fruit's evaluation"?

the next day Houdart says:  Should we understand from your silence that you haven't got a clue?

Today you say:  > I gave Houdart a concrete example and motivation behind it and never heard anything more from:confused:
Parent - - By rocket (****) Date 2012-06-10 21:21
I did not see the message actually but now it's too late:smile: I don't always visit here
Parent - - By Stonehenge (***) Date 2012-06-10 21:33

> I did not see the message actually but now it's too late:smile: I don't always visit here


Oh no, it's not too late. Please be so kind to give us the answer to the questions.

> I gave Houdart a concrete example and motivation behind it and never heard anything more from apart  from "statistical possibility" nonsense when both engines commit the same non-opening book misstakes and same suboptimal line. "


Amazing fiction you write.
Thank you, Homayoun_Sohrabi_M.D., for showing what really happened.
Parent - - By rocket (****) Date 2012-06-10 21:55
if two programs sacrifice a pawn long term(a pawn they don't get back) and the other non-suspected clones don't like it we have the duck scenario.

I know fully well it could be coincidences but if you add it all together it sure looks like a duck, quacks like a duck  and walks like a duck:razz:

Fruit is not even in the same league streight wise anymore and I still find these  peculiar common grounds
Parent - By Stonehenge (***) Date 2012-06-10 22:25
There is a clear statistical possibility that both engines commit the same non-opening book mistakes and use the same suboptimal line.
Parent - - By rocket (****) Date 2012-06-10 21:43

>"Today you say:  > I gave Houdart a concrete example and motivation behind it and never heard anything more from:confused:"<


the message you are referring to was not meant to be sent at that stage. I was not finished writing and I shut down the computer for the day- later so see I accidentally posted it anyway. And now you refer to it!:

Determinism indeed: everything happens by neccesity(congratulations to be part of the chain of events!) :wink:

When I say philosophy I mean some basic controll of squares settings(which are still key) and some other aspects. I know this because when I see rybka deviate with a clearly distinguished preference(such as an unusual pawn sacrifice) I see fruit with the same preference and choice. And most other engines did not really like the sac.

Those type of things make it evident for me that they are based on the same foundation of controll of squares.
Parent - By Homayoun_Sohrabi_M.D. (***) Date 2012-06-11 00:56
Let's try to see if we can meet halfway on this.    I think you are saying that Rybka's playing style is similar to fruit.    Let's say that's true (I don't know that, I have never used Fruit before).    Vas did say that he read through the Fruit's code backwards and forwards, so we would not expect the two engines to be totally opposite.     However, I will not accuse someone of cloning or cheating, simply based on your evidence of a similar pawn sacrifice, etc, here and there.
Parent - By Guenther RWBC (**) Date 2012-06-11 07:52
You walk like a troll, quack like a troll and look like a troll and you have neither
a clue about chess programming nor chess...
(ignored)
Parent - By Sedat Canbaz (*****) Date 2012-06-10 19:17 Edited 2012-06-10 19:37

>That's a myth. Fruit was improved more than 100 points with Toga.


Do you believe in that, if i will start testing Toga, it will be rated more than 3120 Elo ???
http://www.sedatcanbaz.com/chess/scct-rating/

Note:I expect (in case of testing Toga II 1.4 beta5c 6c under SCCT conditions) Toga will be rated around 3000 Elo
I mean,probably the strength of both engines (Toga and Fruit) will have almost same Elo points

See Toga Elo performance, approx.250 Elo weaker than Rybka:
http://www.husvankempen.de/nunn/40_4_Ratinglist/40_4_AllVersion/rangliste.html
http://computerchess.org.uk/ccrl/404/

In other words,your statement is wrong !

Best,
Sedat
Parent - - By Homayoun_Sohrabi_M.D. (***) Date 2012-06-10 12:38
Hi Sedat,

did you see what the president of the Turkish chess federation do?   http://chessbase.com/newsdetail.asp?newsid=8227

I guess he has been assigned to do llyumzhinov's dirty work and settle all the old scores, but that looks pretty embarrassing if you ask me. 

I think we need new leadership, I nominate Sedat Canbaz for president of Turkish chess federation.
Parent - By Sedat Canbaz (*****) Date 2012-06-10 19:04
Dear Homayoun,

Yes...I see again open letters

Thanks a lot,but unfortunately i have no wish be a president,due to i have no free time for all :wink:

Greetings,
Sedat
- - By rocket (****) Date 2012-06-10 20:43 Edited 2012-06-10 20:47

>Do you believe in that, if i will start testing Toga, it will be rated more than 3120 Elo ???<


No Toga was an admitted clone of <Fruit> and achieved what you said is not possible.

Fruit 2.1  2696 

Toga II 1.4.1SE  2832 

http://computerchess.org.uk/ccrl/4040/cgi/compare_engines.cgi?family=Fruit&print=Rating+list&print=Results+table&print=LOS+table&print=Ponder+hit+table&print=Eval+difference+table&print=Comopp+gamenum+table&print=Overlap+table&print=Score+with+common+opponents

there's an improvement of more than 100 points on an open source engine.
Parent - - By Sedat Canbaz (*****) Date 2012-06-11 00:55 Edited 2012-06-11 08:32

>there's an improvement of more than 100 points on an open source engine.


Congrats to Toga team !

Just to make it more clear,
-I am not against Fruit,i like this engine a lot and there is no doubt Fabien Letouzey is very talented programmer
-Plus,i strongly believe 'Fruit' is a mother (or grandmother) of many chess chess engines
-I never said:nobody can create +50 Elo improvement stronger (which is based on free open source engine)
-I repeat again:its very hard to add +50 Elo improvement...and such work should be counted as original work
-And i say again:the latest Fruit versions are not weaker than Toga,but Rybka is approx.250 Elo stronger than Toga/Fruit

Do you think that only Toga and Rybka include stuff/ideas from Fruit ?
And as far as i remember (in one of the interviews) Vasik Rajlich stated clearly that Rybka has ideas from Fruit
In my opinion,almost all current top engines (which are stronger than Fruit2.1) include many stuff/ideas from Fruit
For example,one of them are (which include Fruit stuff/ideas too):Komodo,Stockfish,Strelka,Houdini,Critter,Ivanhoe...

Without to not mention this i cant:
-Fruit 2.0 includes also stuff/ideas from Zappa 1.0

But what is really hurting me,some clever people are trying to prove that Rybka,Houdini,Ivanhoe...are clones
And only their engines are 100 % original work - come on ...we are not so naive and that's really sad and ridiculous

I have a wish from those clever people:
-Its will be great and i will congratulate all of them,who can create stronger engines than Houdini,Rybka,Ivanhoe,Critter...

Best Regards,
Sedat
Parent - By Homayoun_Sohrabi_M.D. (***) Date 2012-06-11 01:25

> But what is really hurting me,some clever people are trying to prove that Rybka,Houdini,Ivanhoe...are clones
> And only their engines are 100 % original work - come on ...we are not so naive and that's really sad and ridiculous
>
> I have a wish from those clever people:
> -Its will be great and i will congratulate all of them,who can create stronger engines than Houdini,Rybka,Ivanhoe,Critter...


very well said:smile:
Up Topic The Rybka Lounge / Computer Chess / SCCT Engine Test – Without Books

Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill