Cookin' The Books

[Norm Pollock]

Geschrieben von:/Posted by: Norm Pollock at 07 May 2004 15:15:42:

Here are a couple of opening books I cooked. The max depth is 40 plies. One is for ruffian (adjust cfg) and the other is for BookThinker. Files are zipped.
They are based on my personal database 0f 45,000 games. The database is 75% players 2530+ elo, and 25% computer-computer games (SSDF and WBEC). The database is heavily filtered to remove "bad" games.
One book is for ruffian, and the other book is for use with BookThinker.
np-ruffian
np-Thinker
cheers
-Norm

[Dann Corbit]

Geschrieben von:/Posted by: Dann Corbit at 08 May 2004 02:42:57:
Als Antwort auf:/In reply to: Cookin' The Books geschrieben von:/posted by: Norm Pollock at 07 May 2004 15:15:42:

Here are a couple of opening books I cooked. The max depth is 40 plies. One is for ruffian (adjust cfg) and the other is for BookThinker. Files are zipped.
They are based on my personal database 0f 45,000 games. The database is 75% players 2530+ elo, and 25% computer-computer games (SSDF and WBEC). The database is heavily filtered to remove "bad" games.
One book is for ruffian, and the other book is for use with BookThinker.
np-ruffian
np-Thinker

I would be interested to see results for Fruit using your thinker book compared to the one that I created:
ftp://cap.connx.com/pub/chess-engines/n ... hinker.dat



my ftp site {remove http:// unless you like error messages}

[Norm Pollock]

Geschrieben von:/Posted by: Norm Pollock at 08 May 2004 03:58:31:
Als Antwort auf:/In reply to: Re: Cookin' The Books geschrieben von:/posted by: Dann Corbit at 08 May 2004 02:42:57:

Here are a couple of opening books I cooked. The max depth is 40 plies. One is for ruffian (adjust cfg) and the other is for BookThinker. Files are zipped.
They are based on my personal database 0f 45,000 games. The database is 75% players 2530+ elo, and 25% computer-computer games (SSDF and WBEC). The database is heavily filtered to remove "bad" games.
One book is for ruffian, and the other book is for use with BookThinker.
np-ruffian
np-Thinker

I would be interested to see results for Fruit using your thinker book compared to the one that I created:
ftp://cap.connx.com/pub/chess-engines/n ... hinker.dat

OK, will do. I'm now running a 50 game match at 3 minutes + 1 second under Jori's wbtm. I'm using Thinker.dat from your site (562,013). I'm using an amd 2000xp+ with 128M per engine. Yesterday I ran a similar match of 20 games against fruit with the Thinker.dat from Thinker. Nothing conclusive: 11.0/9.0 in favor of my book. Chance plays a big part in the outcome.

[Norm Pollock]

Geschrieben von:/Posted by: Norm Pollock at 08 May 2004 13:16:50:
Als Antwort auf:/In reply to: Re: Cookin' The Books geschrieben von:/posted by: Norm Pollock at 08 May 2004 03:58:31:

I would be interested to see results for Fruit using your thinker book compared to the one that I created:
ftp://cap.connx.com/pub/chess-engines/n ... hinker.dat

OK, will do. I'm now running a 50 game match at 3 minutes + 1 second under Jori's wbtm. I'm using Thinker.dat from your site (562,013). I'm using an amd 2000xp+ with 128M per engine. Yesterday I ran a similar match of 20 games against fruit with the Thinker.dat from Thinker. Nothing conclusive: 11.0/9.0 in favor of my book. Chance plays a big part in the outcome.

The result is 26.5/23.5 in favor of my book. Nothing conclusive about that. Chance is the main factor. The pgns, table and verification file can be downloaded below. (Note: "fruit" is with my book, "fruit3" is with Dan's book):
fruit v fruit3