HS-Masterbook 4.0.ctg

[Le Fou numerique]

Hi,

Harry Schnapp publish a new book: HS-Masterbook 4.0.ctg.

Harry wrote me:
This release is the hugest eng-eng book in computerchess.
The HS-Masterbook 4.0.ctg is created :
-For opening training
-For opening analysis
-To play against other classical big books
-To play against itself with very important settings to ensure correct functionality.

Regards,
Patrick

[Graham Banks]

Thanks to Harry for this book and also to you Patrick for hosting it.
(2gb unzipped! )

[tomgdrums]

Thanks to Harry for this book and also to you Patrick for hosting it.
(2gb unzipped! )

How is Harry's book? Is it good for opening training and such?

[Graham Banks]

Thanks to Harry for this book and also to you Patrick for hosting it.
(2gb unzipped! )

How is Harry's book? Is it good for opening training and such?

I can only comment on the usage of Harry's books for neutral engine v engine testing. In this regard, his books are still the best (in my opinion).

Cheers,
Graham.

[Sedat Canbaz]

Thanks to Harry for this book and also to you Patrick for hosting it.
(2gb unzipped! )

How is Harry's book? Is it good for opening training and such?

I can only comment on the usage of Harry's books for neutral engine v engine testing. In this regard, his books are still the best (in my opinion).

Cheers,
Graham.

Hello Graham,

First of all,many thanks to Harry Schnapp and there is no doubt that his books are excellent work

But anyway,i strongly believe in that:nobody (no any book author) is best !
I don't want to start counting the strong book creators,because i am afraid to not mention/miss some names

But nowadays there are many book authors who create really very strong/superior books,they (the super strong book creators) are changing periodically...

Actually instead of comments/polls,i prefer testings,standings,games...
In other words,the best way of measuring the opening book authors work is the system by elo calculation

BTW,sometime i wonder about my ratings:why the elo difference between Rybka 4 and Stockfish 1.8 is more than 70 elo ?!

For example,in most other ratings,the elo difference of the both same engines is approx.30-50 elo ?!

Is that can be the reason/influence of using Perfect 2010 book ???


SCCT (15m+10s) Gladiators:
http://sedatchess.110mb.com/index.php?p=1_62

Code: Select all

    Program                            Score     %    Av.Op.  Elo    +   -    Draws

  1 Rybka 4 x64 T4                 : 102.5/132  77.7   3080   3296   50  48   37.1 %
  2 Stockfish 1.8 JA x64 T4        :  87.0/132  65.9   3088   3203   45  44   45.5 %

SCCT (10 minutes/game) Rating List:
http://sedatchess.110mb.com/index.php?p=1_59

Code: Select all

Rank   Name                   Elo    +    -  games score oppo. draws

3 Rybka 4 x64 T4             3286   17   17  1460   78%  3068   33%
7 Stockfish 1.8 JA x64 T4    3214   21   20   850   70%  3060   38% 

Note also that Houdini 1.5 has 14 games forfeit on time,even with these time loses,its rating is 70 elo stronger than Rybka 4

Code: Select all

Rank   Name                   Elo    +    -  games score oppo. draws

1 Houdini 1.5 x64 T4         3356   27   26   780   85%  3051   19%
3 Rybka 4 x64 T4             3286   17   17  1460   78%  3068   33% 

One thing more,since almost two years i don't play on Playchess server (i have stopped testing my Perfect opening books on the server,due to nowadays i have no much free time,plus there is a such possibility of some people who will start to think that i use SCCT private books)

A little note (without not mentioning this i can't and i feel a big proud):my Perfect books were with highest rank/best performance on Playchess server
I mean about Perfect 13x/14/15/16 books (on Playchess server),this test is done before SCCT private book registrations (in those days i was testing mainly free books)


Best Wishes,
Sedat

[beachknight]

Hi,

Harry Schnapp publish a new book: HS-Masterbook 4.0.ctg.

Harry wrote me:
This release is the hugest eng-eng book in computerchess.
The HS-Masterbook 4.0.ctg is created :
-For opening training
-For opening analysis
-To play against other classical big books
-To play against itself with very important settings to ensure correct functionality.

Regards,
Patrick

Thanks Patrick for the announcement. Congrats to Harry for
his new biggest book. I just want to point out a minor thing.
The topic is the book size. I assume we dont want to mention
powerbooks by chessbase. I have created some books over
4 GB in the past. The first one was a really huge one, 12 GB
in size and created in 2007. The latest one was Fritz12ultra.ctg
whose size is 6+ GB. Its creation took around one year. I was
able to complete it just before the Fritz 12 release. I have been
using it mainly in my fritz gui test tournaments.

Best,

[ernest]

-To play against itself with very important settings to ensure correct functionality.

Bonjour Patrick,
Where are those very important settings explained?

[mwyoung]

Thanks to Harry for this book and also to you Patrick for hosting it.
(2gb unzipped! )

How is Harry's book? Is it good for opening training and such?

I can only comment on the usage of Harry's books for neutral engine v engine testing. In this regard, his books are still the best (in my opinion).

Cheers,
Graham.

Hello Graham,

First of all,many thanks to Harry Schnapp and there is no doubt that his books are excellent work

But anyway,i strongly believe in that:nobody (no any book author) is best !
I don't want to start counting the strong book creators,because i am afraid to not mention/miss some names

But nowadays there are many book authors who create really very strong/superior books,they (the super strong book creators) are changing periodically...

Actually instead of comments/polls,i prefer testings,standings,games...
In other words,the best way of measuring the opening book authors work is the system by elo calculation

BTW,sometime i wonder about my ratings:why the elo difference between Rybka 4 and Stockfish 1.8 is more than 70 elo ?!

For example,in most other ratings,the elo difference of the both same engines is approx.30-50 elo ?!

Is that can be the reason/influence of using Perfect 2010 book ???


SCCT (15m+10s) Gladiators:
http://sedatchess.110mb.com/index.php?p=1_62

Code: Select all

    Program                            Score     %    Av.Op.  Elo    +   -    Draws

  1 Rybka 4 x64 T4                 : 102.5/132  77.7   3080   3296   50  48   37.1 %
  2 Stockfish 1.8 JA x64 T4        :  87.0/132  65.9   3088   3203   45  44   45.5 %

SCCT (10 minutes/game) Rating List:
http://sedatchess.110mb.com/index.php?p=1_59

Code: Select all

Rank   Name                   Elo    +    -  games score oppo. draws

3 Rybka 4 x64 T4             3286   17   17  1460   78%  3068   33%
7 Stockfish 1.8 JA x64 T4    3214   21   20   850   70%  3060   38% 

Note also that Houdini 1.5 has 14 games forfeit on time,even with these time loses,its rating is 70 elo stronger than Rybka 4

Code: Select all

Rank   Name                   Elo    +    -  games score oppo. draws

1 Houdini 1.5 x64 T4         3356   27   26   780   85%  3051   19%
3 Rybka 4 x64 T4             3286   17   17  1460   78%  3068   33% 

One thing more,since almost two years i don't play on Playchess server (i have stopped testing my Perfect opening books on the server,due to nowadays i have no much free time,plus there is a such possibility of some people who will start to think that i use SCCT private books)

A little note (without not mentioning this i can't and i feel a big proud):my Perfect books were with highest rank/best performance on Playchess server
I mean about Perfect 13x/14/15/16 books (on Playchess server),this test is done before SCCT private book registrations (in those days i was testing mainly free books)


Best Wishes,
Sedat

I am not sure what to think of all the work some put into making engine books. I don't know if there is any value in someone taking all that time tuning a book. I don't know as I have never tested it.

For fun, and I am not picking on anyones book, I will test a random book I just made against perfect 2010. I will run two Houndini 1.5 engines one with perfect 2010 book, and one with the book I just made. I don't know what the outcome will be, but I am curious.

I will let them play 1+1 games for a few hours and see what results I get. Everything is the same on the engines except the opening books.

I make a book using a database of 5.2 million games. I filtered the games into a smaller database. Using the terms year 1970 or later, and one of the players must have a rating of 2500 elo. A very simple book. I then imported the games using ECO-R.L. of 20.

[Le Fou numerique]

Bonjour Patrick,
Where are those very important settings explained?

Salut Ernest,

In the Readme HS-Masterbook 4.0.rtf file inclued in the download archive.

Cordialement,
Patrick

[Rubinus]

Hi Sedat.
Your books very preferred B90 = advantage for Rybka. If use width repertoir, or some testsets, 30-50 ELO difference Rybka:Stockfish is probably.
Regards Pavel Háse

My little ratings:
1. Deep Rybka 4 x64_4CPU 3219 150
2. Houdini 1.03a x64 4_CPU 3216 60
- Deep Rybka 4.1 x64 Beta 1_4_CPU 3213 30
3. IvanHoe 9.50b x64_4_CPU 3203 60
4. Stockfish 1.9.1 JA 64bit_4_CPU 3190 150
5. Fire 1.3.1 x64KLO_4_CPU 3187 60
6. DeepSaros 2.3c3w32_4_CPU 3161 30
7. Critter 0.90 x64_4_CPU 3134 60

But use always my width book: PH-TestBook2010. SF:Rybka -29 ELO ...