Not logged inRybka Chess Community Forum
Up Topic Rybka Support & Discussion / Rybka Discussion / new .ctg book for Rybka (by Harry Schnapp)
- - By Vasik Rajlich (Silver) [hu] Date 2007-07-11 11:09
We've put a new .ctg book by Harry Schnapp for download from our web site.

Direct download link:

http://www.rybkachess.com/download/HS_for_Rybka.zip

Book downloads:

http://www.rybkachess.com/index.php?auswahl=Downloads

Vas
Parent - - By ok373 (**) [de] Date 2007-07-11 13:30
Is it better than Rybka II?
Parent - - By ok373 (**) [de] Date 2007-07-11 14:19
Which is the best book in the Rybka website for Fritz 10 interface?
Parent - By Rybka_King88 (*) [us] Date 2007-07-11 18:28
My Favorite is Sheebar... it is huge but GREAT book...
Parent - - By garyf919 (**) [us] Date 2007-07-11 20:51
What is the purpose of this new book? Does it replace the Rybka II final book?
Parent - - By Vasik Rajlich (Silver) [hu] Date 2007-07-12 07:43
Jeroen's RybkaII.ctg is the current official Rybka book. This will be the case until it's replaced by Jeroen's RybkaIII.*.

HS_for_Rybka.ctg is a nice strong additional book for Chessbase interfaces which Harry was kind enough to provide for us.

Vas
Parent - - By bedouin (**) [gb] Date 2007-07-12 17:16
I hope the HS_for_Rybka.ctg is more comprehensive than the mainbook.abk by the same author for Arena. Has anyone tried it?
Parent - - By Graham Banks (****) [nz] Date 2007-07-13 01:08
The HS-Book.abk is the name of the most up to date mainbook.abk.
It's a generic book as opposed to an engine specific book and it is of very high quality.
It's one of the few Arena books that I'm prepared to use for my CCRL testing.

Regards, Graham.
Parent - By bedouin (**) [gb] Date 2007-07-14 08:55
Thanks I will get both books the hs ctg and the hs abk version.
Parent - By Gregory Owett (*) [fr] Date 2007-07-13 06:43
When I use RybkaII in the tournaments between Rybkas, I often obtain identical games. With HS-book that never arrives. The games are most varied.

Gregory
Parent - By Harry Schnapp (**) [de] Date 2007-07-12 18:49
  Hello friends,
  In the HS_for_Rybka.ctg book please mark in green the
  initial move 1)e4 as tournament move!
   Thanks,
     Harry Schnapp
Parent - - By Arrière Pensée (Gold) Date 2007-07-12 18:58 Edited 2007-07-12 19:03
Hi Vas,

That was very considerate of Harry Schnapp to take time out to do this- If he happens to peruses this web site-
" Thanks Harry for your great work.!"

Best of All Wishes For a Pleasant Day-

Robert

Well, you were posting while I was writing! :-)
Parent - - By Mark Mason (***) [gb] Date 2007-07-13 05:56
These are the setting for using the book, just in case anyone didn't spot it on Talkchess Forum.

http://64.68.157.89/forum/viewtopic.php?t=15011

Mark
Parent - - By ok373 (**) [il] Date 2007-07-13 09:15
Which is stronger Sheebar or HS_for_Rybka?
Parent - By Harry Schnapp (**) [de] Date 2007-07-13 10:40
 
  Hello,

   http://www.rybkachess.com/index.php?auswahl=Downloads

   Link with short description of the HS_for_Rybka.ctg book.

     Best regards,
      Harry Schnapp
Parent - - By Harry Schnapp (**) [de] Date 2007-07-13 11:05
   Hello,
  Settings for the HS_for_Rybka.ctg:
1) Rightclick 1)e4 in the book window and make that move to a Main move.(in some Chessbase GUIs 1)e4 is already a main move).
2) Writeprotect the book files now (not necessary,but my recommendation).
3) In the book options click first on "Optimize" and afterwards set Minimum games "1" and Up to move "30"

     That's all.

       Harry   

 
Parent - - By ok373 (**) [il] Date 2007-07-13 11:31
They are both defenited "very strong" books,so how can I know which one is stronger?
Parent - By fhub (**) Date 2007-07-13 11:35
Because it´s HIS one ... :)
Parent - By Felix Kling (Gold) [de] Date 2007-07-13 12:44
You will have to test it! Just let Rybka play against Rybka, one using Sheebar.ctg and one using the new book by Harry Schnapp.
If I would know which one is stronger, i would say it on the download page.
Parent - - By George Tsavdaris (****) Date 2007-07-14 16:51

>They are both defenited "very strong" books,so how can I know which one is stronger?


I will try to see this for very fast blitz time 1'+1'':

I'm running a 1+1 tournament with Rybka 1.1 having: Harry's book(95 MB), my book(3.42 GB) and Sheebar book(512 MB).
So it is Rybka 1.1 playing each other with 3 different books.

I will play 300-400 games for each engine i think.....
So far:

Rybka 1.1three    *********  36.0-28.0  30.5-33.5   66.5/128  52.0%
Rybka 1.1two      28.0-36.0  *********  38.0-26.0   66.0/128  51.6%
Rybka 1.1 32-bit  33.5-30.5  26.0-38.0  *********   59.5/128  46.5%

Rybka 1.1three uses your book, 1.1two my book and 1.1 32-bit Sheebar.

Harry's book is destroying mine while it loses to Sheebar, while my book is destoying Sheebar. Strange world! :)

Although it is losing, my book is full of traps and Harry's fall in two of them: :)

In Harry's book he has 3...e6 in the following game but it is on a grey colour and Chessbase GUI does not play it.
Yet the 3...dxe4 is playable but black should play in this line 8...Kc6! instead of 8...Kc7 Rybka played....
I will not reveal the other trap that the "HS_for_Rybka" book fall, sorry....

[Event "RRRbooks"]
[Site "Thessaloniki"]
[Date "2007.07.14"]
[Round "10.2"]
[White "Rybka 1.1two"]
[Black "Rybka 1.1three"]
[Result "1-0"]
[ECO "B21"]
[PlyCount "83"]
[EventDate "2007.07.14"]
[Source "Tsavdaris"]
[TimeControl "60+1"]

1. e4 {[%emt 0:00:00]} c5 {[%emt 0:00:00]} 2. f4 {[%emt 0:00:00]} d5 {
[%emt 0:00:00]} 3. Nf3 {[%emt 0:00:00]} dxe4 {
Black out of book. [%eval -24,11] [%emt 0:00:03]} 4. Ng5 {[%emt 0:00:00]} Nf6 {
[%eval -20,10] [%emt 0:00:04]} 5. Bc4 {[%emt 0:00:00]} Bg4 {
[%eval -46,11] [%emt 0:00:03]} 6. Qxg4 {[%emt 0:00:00]} Nxg4 {
[%eval -73,10] [%emt 0:00:01]} 7. Bxf7+ {[%emt 0:00:00]} Kd7 {
[%eval -98,5] [%emt 0:00:00]} 8. Be6+ {[%emt 0:00:00]} Kc7 {
[%eval -63,11] [%emt 0:00:04]} 9. Bxg4 {[%emt 0:00:00]} Qe8 {
[%eval -64,10] [%emt 0:00:02]} 10. Ne6+ {[%emt 0:00:00]} Kb6 {
[%eval -51,10] [%emt 0:00:02]} 11. Na3 {[%emt 0:00:00]} a6 {
[%eval -48,10] [%emt 0:00:02]} 12. d3 {[%emt 0:00:00]} Ka7 {
[%eval -61,9] [%emt 0:00:02]} 13. Be3 {[%emt 0:00:00]} Qg6 {
[%eval -66,11] [%emt 0:00:05]} 14. Bh3 {[%emt 0:00:00]} exd3 {
[%eval -41,10] [%emt 0:00:01]} 15. O-O-O {[%emt 0:00:00]} b6 {
[%eval -11,11] [%emt 0:00:06]} 16. Rxd3 {[%emt 0:00:00]} Qe4 {
[%eval 0,11] [%emt 0:00:02]} 17. Re1 {
White's last book move. A winning position. [%emt 0:00:00]} Qc6 {
[%eval 29,11] [%emt 0:00:04]} 18. Nc4 {(Re1-d1) [%eval 112,10] [%emt 0:00:03]}
Nd7 {[%eval 118,10] [%emt 0:00:04]} 19. Red1 {[%eval 176,11] [%emt 0:00:02]}
Nf6 {[%eval 172,11] [%emt 0:00:02]} 20. Nxb6 {[%eval 178,12] [%emt 0:00:05]}
Qxb6 {[%eval 186,11] [%emt 0:00:02]} 21. Bxc5 {[%eval 207,11] [%emt 0:00:02]}
Qxc5 {[%eval 225,13] [%emt 0:00:02]} 22. Nxc5 {[%eval 234,12] [%emt 0:00:02]}
e5 {[%eval 241,12] [%emt 0:00:02]} 23. Nd7 {[%eval 245,12] [%emt 0:00:03]} exf4
{[%eval 245,11] [%emt 0:00:01]} 24. Nxf6 {[%eval 253,11] [%emt 0:00:01]} gxf6 {
[%eval 256,11] [%emt 0:00:02]} 25. Rb3 {(Rd3-d7+) [%eval 268,11] [%emt 0:00:02]
} Rb8 {[%eval 264,12] [%emt 0:00:01]} 26. Rd7+ {[%eval 267,11] [%emt 0:00:01]}
Ka8 {[%eval 257,12] [%emt 0:00:01]} 27. Bf5 {[%eval 261,12] [%emt 0:00:04]}
Rxb3 {[%eval 261,12] [%emt 0:00:02]} 28. cxb3 {[%eval 268,12] [%emt 0:00:03]}
h6 {(Ka8-b8) [%eval 265,10] [%emt 0:00:01]} 29. Be4+ {
(Rd7-f7) [%eval 258,10] [%emt 0:00:02]} Kb8 {[%eval 264,5] [%emt 0:00:00]} 30.
b4 {(Rd7-b7+) [%eval 266,12] [%emt 0:00:03]} Kc8 {
(h6-h5) [%eval 269,12] [%emt 0:00:02]} 31. Ra7 {[%eval 266,12] [%emt 0:00:03]}
Bd6 {[%eval 276,12] [%emt 0:00:03]} 32. Ra8+ {[%eval 277,11] [%emt 0:00:02]}
Bb8 {[%eval 275,11] [%emt 0:00:00]} 33. Rxa6 {[%eval 281,12] [%emt 0:00:02]}
Be5 {(Rh8-f8) [%eval 280,11] [%emt 0:00:02]} 34. Ra8+ {
[%eval 288,11] [%emt 0:00:02]} Bb8 {[%eval 288,11] [%emt 0:00:00]} 35. Kc2 {
(b4-b5) [%eval 288,12] [%emt 0:00:04]} Re8 {
(Kc8-d7) [%eval 295,10] [%emt 0:00:01]} 36. Bf3 {[%eval 325,12] [%emt 0:00:03]}
Re1 {(Re8-d8) [%eval 323,11] [%emt 0:00:02]} 37. a4 {
(b4-b5) [%eval 360,12] [%emt 0:00:02]} Rf1 {
(Re1-e7) [%eval 350,11] [%emt 0:00:04]} 38. a5 {
(b4-b5) [%eval 404,14] [%emt 0:00:02]} Rxf3 {
(Rf1-f2+) [%eval 404,14] [%emt 0:00:02]} 39. gxf3 {
[%eval 1023,16] [%emt 0:00:04]} Kb7 {[%eval 1116,17] [%emt 0:00:04]} 40. Rxb8+
{[%eval 1749,17] [%emt 0:00:09]} Kxb8 {[%eval 1039,15] [%emt 0:00:01]} 41. b5 {
(Kc2-d3) [%eval 2160,17] [%emt 0:00:01]} Kc8 {
(Kb8-c7) [%eval 971,14] [%emt 0:00:01]} 42. Kd3 {[%eval 2885,16] [%emt 0:00:03]
} 1-0

My impression so far is that Harry's book is very deep although "only" 95 MB in comparison with Sheebar which is 512 MB and mine which is 3.42 GB.
I've compared some lines and my book is much deeper, but it seems Harry found a way to put lines that Rybka likes very much. I have seen this in 3 games until now. Although Sheebar and my book created a position which is better for one color, Rybka played badly with that color and lost.
So Harry's books seems indeed an optimized for Rybka book....

As for the depth which is very deep in Harry's book and in 2 cases even "out-booked" mine and Sheebar, i wondered how is it possible with such a small book in comparison with my 3.42 GB and Sheebar's 512 MB. I think that the answer is that my book for example(i don't know about Sheebar) has millions of risky lines that contain all the different variations up to the endgame while Harry's book is more solid with only "logical" moves. Ideal for Rybka i would say.... :)
Parent - - By bedouin (**) [gb] Date 2007-07-14 16:58
Your book is 3.42 GB? Isn't that extraordinary? How much RAM have you got on your system to use such an opening book. Imagine having 3-4-5 TBs loaded into memory as well as your book and you still need RAM for hash?
Parent - - By George Tsavdaris (****) Date 2007-07-14 17:19
It was around 4.1 GB but because when i was adding lines, Chessbase GUI sometimes run a bit slow, i decided to cut many unnecessary lines.
It works fine now with 3.4 GB....

I have 4 GB ram, but when i load the book it does say only 80.000 K in the task manager for the Chessbase GUI. I guess it means 80.000 kB. So i guess even bigger size of the book with lower RAM, would not be a problem as it seems.....
Parent - - By bedouin (**) [gb] Date 2007-07-14 17:39
4 GB of RAM? If you give hash 1 GB and tablebases 256 Mb then your opening book is consuming too much RAM. It may not be worth having such a big book.
Parent - - By George Tsavdaris (****) Date 2007-07-14 17:54

>4 GB of RAM? If you give hash 1 GB and tablebases 256 Mb then your opening book is consuming too much RAM.


That doesn't make sense.....

As i've said, the book as the task manager shows, does not consume much RAM at all....
Also i haven't noticed any slowdown at all when using the book....
Parent - - By bedouin (**) [gb] Date 2007-07-15 08:31
A 3.4 GB book probably means you have lines from blitz or games from amateur players and the result is a weaker book (than sheebar or hsbook). If some guy won the tours with the HS book then you can bet that your book contains a lot of irrelevant stuff. Does rybka emerge from your 3.4 GB book with favourable evals?
Parent - By George Tsavdaris (****) Date 2007-07-15 14:16

>A 3.4 GB book probably means you have lines from blitz or games from amateur players and the result is a weaker book (than sheebar or hsbook).


I don't know if this is true.
The book was about 1.2 GB when i generated it before some years(5-6). Then i added manually lines which are good for sure.
Still has many many things missing, but i can't find time and i don't really want to find time to work more seriously in the book....

>If some guy won the tours with the HS book then you can bet that your book contains a lot of irrelevant stuff.


What does that mean? I haven't participated in that tournament....

>Does rybka emerge from your 3.4 GB book with favourable evals?


Sometimes yes, sometimes no.
Yet, the ratio between yes/no until now in the tournament is smaller than with Harry's book.
Although the difference is very small to say Harry's book is better for Rybka as it seemed in the beginning....

Update for the tournament:
Score seems that is equal for the 3 books. It seems that these 3 books does not make any difference in this 1+1 time control for Rybka 1.1.

Rybka 1.1(Sheebar)   *********  61.5-48.5      50.0-60.0     111.5/220  50.7%
Rybka 1.1(Harry)       48.5-61.5      *********   61.5-48.5     110.0/220  50.0%
Rybka 1.1(My book)   60.0-50.0     48.5-61.5      *********  108.5/220  49.3%

Notice that Sheebar book wins Harry's book with the same score Harry's book wins mine!
Differences are insignificant for now..... I will play 400 games for each book to see if anything will change.....

Note that if we exclude my book then we would have that Rybka 1.1 with Sheebar would win 61.5-48.5 Rybka 1.1 with Harry's book.
That would give an evidence that Sheebar is better for Rybka 1.1 than Harry's book is.

Note also that if we exclude Sheebat book then we would have that Rybka 1.1 with Harry's book would win 61.5-48.5 Rybka 1.1 with my book.
That would give an evidence that Harry's book is better for Rybka 1.1 than my book is.

BUT combining all these results we have a seemingly contradiction that no book is better for Rybka 1.1 at time control of 1+1. Until now....
Parent - - By Nelson Hernandez (Gold) [us] Date 2007-07-15 13:46
3.42 GB is emphatically not extraordinary.
Parent - - By Felix Kling (Gold) [de] Date 2007-07-15 13:52
My book is 4,4 GB big, I always had problems in win98 SE with it because the OS doesn'T support such big files :)
It's very usful to have such a big book when exploring new opening lines...
Parent - By MoKy (**) [sk] Date 2007-07-16 11:12
Today we need new system for storing books.
Each opening would have one self.
Parent - - By bedouin (**) [gb] Date 2007-07-16 18:11

>My book is 4,4 GB big, I always had problems in win98 SE with it because the OS doesn'T support such big files :-)


4.4 GB on Windows 98? I thought Windows 98 was retired many years ago. Where do you get updates for Internet Explorer etc? If you get a virus you are finished.
Parent - By Felix Kling (Gold) [de] Date 2007-07-16 19:06
I'm using Opera and switched to Ubuntu 1 year ago.
Parent - By Vasik Rajlich (Silver) [hu] Date 2007-07-16 11:12
But enough about tablebases .. :)

Vas
Parent - - By Aung (**) [kr] Date 2007-07-15 01:01
Good test running George Tsavdaris...
If u have time please run another test 3+0 Rybka2.** with Sheebarblitz...
Parent - By George Tsavdaris (****) Date 2007-07-15 14:18
When i will buy a quad or even 8-CPU i will buy Rybka 2.xx and then perhaps i will do such a test, sorry.....
Parent - - By Harry Schnapp (**) [de] Date 2007-07-15 17:19
 
   Hello my friends,
  Time control 1'+1" is a lottery.
Parent - - By George Tsavdaris (****) Date 2007-07-15 19:48
I agree in the sense that the quality of games is lower than it is on longer time controls.
That's why i'm repeating all the time that: So far no book seems to be better for Rybka 1.1 at time control of 1+1.
Parent - - By Harry Schnapp (**) [de] Date 2007-07-15 21:38
  Hi George,
  My english is not enough to explain to you clearly.
  I try it:
  Very important is the evaluation by out of book.If your book comes into the game with +0.30 and
  nearly each move is played into 1-2" is everything hazard!. The +0.30 rises or falls rapidly fast in so a "game".
  How can have the "result" something to do with the quality of the line?? If you win were yours book better
  and if you lose were the other book better??
  With at least some seconds more per move,so that Rybka has time TO USE the good line,then one could judge
  somewhat better the difference in quality of the books. But one needs games of at least 3'+2" or even 5' on a fast computer.
    Best regards,
      Harry
Parent - By Aung (**) [kr] Date 2007-07-15 22:49
agree with Harry...
1+1 is not book result..
Harry's book is uptodate..and Sheebar book is oldie now..
Engine need to find some position for need little more time..
BTW..good job Harry...
Parent - By Lukas Cimiotti (Bronze) [de] Date 2007-07-14 20:00
Thank you for this book.
Today there were 2 tournaments at chessbase held by William Fuller. In both of them i used your book. In the first tour i made 2. place, in the second tour i won. :)
Parent - By Arrière Pensée (Gold) Date 2007-07-13 14:53
Hi Mark,

Thank you!

Best Wishes,

Robert
Parent - - By Harry Schnapp (**) [de] Date 2007-07-13 16:00
 
   Hello Robert,
  Thank you very much for your kind words !!!

    Best regards,
     Harry.
Parent - By z0qxz (**) Date 2007-07-13 21:27
Harry,...  thanks for h.s. rybka,ctg book... i have use it..very good book and the settings that u suggest to use it..  great job!!  :-)  z0qxz
Parent - - By Elindros97 (*) [us] Date 2007-07-13 15:46
I know almost all the books are of "chessbase" format of *.ctg

but I have only the free version of Chessbase Light which can only load *.cbh and PGN databases...I know converting the *.ctg into PGN is unlikely if not impossible, but what about the *.cbh?

If I have Chessbase Light, shouldn't I be able to access the Chessbase format of these books in *.ctg?
Parent - - By bedouin (**) [gb] Date 2007-07-14 16:59
Elindros the CBLight can also access the CTG book.
Parent - - By Elindros97 (*) [us] Date 2007-07-16 19:10
I'm fairly sure I tried to do this and it couldn't recognize it. I will atempt to do this again before inquiring further. Thanks for your response.
Parent - - By Elindros97 (*) [us] Date 2007-07-16 19:21
This is now confirmed. I downloaded the Rybka Opening book from the website in *.ctg format and after extracting the files into a folder on the desktop, I tried to open the .ctg with CBLight and it told me it was an invalid database format.

Am I doing something wrong? There are also 2 other much smaller files in the ZIP folder which I downloaded (*.ctb and *.cto).
Parent - By bedouin (**) [gb] Date 2007-07-16 19:54
You are using it the wrong way. CTG=opening book and CBH=database. You open a CBH with cblight but a CTG is for opening book reference. In the CBH database that comes with cblight, open that (CB Light Databse.cbh) and then click on the game e.g. Navara-Motylev. On the opening book tab point to a CTG book so that you can see the book moves.
Parent - - By Nelson Hernandez (Gold) [us] Date 2007-07-17 01:18
Without going into a lot of detail, I've tested the properties of this new Schnapp book using my own methods, which are unlike those of others and far more rigorous and statistically-based.  I have to say that this book appears to be the strongest .ctg book I have ever seen at "Normal" settings.  My tournament book did measurably better against it, but all the relevant superiority margins weren't impressive at all. 

I was impressed by its choice of openings in particular.  Can you believe this book actually dared to play 1.h3?  Spurned B90 lines as white?  Amazingly well done for such a compact book.
Parent - By Harry Schnapp (**) [de] Date 2007-07-17 09:59
   Hello Nelson.
  I thank you much for this competent description.
  Probably also Vasik will be pleased about it.

     Best regards,
       Harry
 
Up Topic Rybka Support & Discussion / Rybka Discussion / new .ctg book for Rybka (by Harry Schnapp)

Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill