A Book For a Vintage Program

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Rebel, chrisw

User avatar
mhull
Posts: 13447
Joined: Wed Mar 08, 2006 9:02 pm
Location: Dallas, Texas
Full name: Matthew Hull

A Book For a Vintage Program

Post by mhull »

I'm looking for an opening book for a vintage computer program from the late 1970s. I'm guessing this program has not played any public games since 1978. It will likely not search deeper than 9 plies at human tournament time controls.

The program will not be a giant killer but it might be a novelty for humans of a certain strength.

I'm thinking perhaps 200 - 300 variations to start with human opposition in mind. There could even be different versions of the book to change-out the openings if desired. A final book could be much larger but I'm just testing at the moment.

Any help or advice would be appreciated.
Matthew Hull
User avatar
Roland Chastain
Posts: 640
Joined: Sat Jun 08, 2013 10:07 am
Location: France
Full name: Roland Chastain

Re: A Book For a Vintage Program

Post by Roland Chastain »

Hello! I don't know exactly what kind of book you are looking for, but for me I was glad when I discovered that I could use pgn-extract to build books in what I called "TSCP format" (maybe the good format for a vintage program), from pgn files. So I downloaded collections of grandmasters games and, for example, build a book for white from games where Fischer played white, and so on. You can find an example of command line to do that here:
http://talkchess.com/forum3/viewtopic.p ... 33#p826533

Still in a "vintage" spirit, I converted my books to the "Kathe Spracklen format":
https://github.com/rchastain/alouette/b ... /white.txt

My two cents.

Regards.

Roland
Qui trop embrasse mal étreint.
User avatar
mclane
Posts: 18748
Joined: Thu Mar 09, 2006 6:40 pm
Location: US of Europe, germany
Full name: Thorsten Czub

Re: A Book For a Vintage Program

Post by mclane »

IMO the book should be kind of mirror of the playing style of the engine.

If it plays like Petrosian I would not bring aggressive openings into the book.
If the engine plays Like tal it would be nice to bring open positions into the book.
If it plays very nimzowitch....


The engine must come out of book in a position it can handle.
What seems like a fairy tale today may be reality tomorrow.
Here we have a fairy tale of the day after tomorrow....
User avatar
mhull
Posts: 13447
Joined: Wed Mar 08, 2006 9:02 pm
Location: Dallas, Texas
Full name: Matthew Hull

Re: A Book For a Vintage Program

Post by mhull »

mclane wrote: Tue Oct 20, 2020 2:15 pm IMO the book should be kind of mirror of the playing style of the engine.

If it plays like Petrosian I would not bring aggressive openings into the book.
If the engine plays Like tal it would be nice to bring open positions into the book.
If it plays very nimzowitch....


The engine must come out of book in a position it can handle.
Thanks for the pointers.

This program does very little pruning and move ordering seems to mainly depend upon the score of the previous search iteration (it does iterative deepening). So its search is wide and not deep. 9 plies would be a maximum (on modern hardware) and more likely is 5-6 plies in a middle-game.

A book tailored for a 1980s non-speculative program might be good for testing at-least. This program is going to play more like a club-house "Petrosian" (not like an aspiring Tal or Rashid Nezhmetdinov).

So far, I have setup a program to process PGN format. What I don't have is a way to set the depth of the book or to tailor the opening for white and black.

So if there are books (in the public domain) that already set to a middle-game positions for white or black, that would be ideal.
Matthew Hull
User avatar
mhull
Posts: 13447
Joined: Wed Mar 08, 2006 9:02 pm
Location: Dallas, Texas
Full name: Matthew Hull

Re: A Book For a Vintage Program

Post by mhull »

Roland Chastain wrote: Tue Oct 20, 2020 11:06 am Hello! I don't know exactly what kind of book you are looking for, but for me I was glad when I discovered that I could use pgn-extract to build books in what I called "TSCP format" (maybe the good format for a vintage program), from pgn files. So I downloaded collections of grandmasters games and, for example, build a book for white from games where Fischer played white, and so on. You can find an example of command line to do that here:
http://talkchess.com/forum3/viewtopic.p ... 33#p826533

Still in a "vintage" spirit, I converted my books to the "Kathe Spracklen format":
https://github.com/rchastain/alouette/b ... /white.txt

My two cents.

Regards.

Roland
I think the Spracklen books could work fine for testing.

Thanks, Roland!
Matthew Hull
User avatar
Rebel
Posts: 6991
Joined: Thu Aug 18, 2011 12:04 pm

Re: A Book For a Vintage Program

Post by Rebel »

mhull wrote: Mon Oct 19, 2020 9:00 pm I'm looking for an opening book for a vintage computer program from the late 1970s. I'm guessing this program has not played any public games since 1978. It will likely not search deeper than 9 plies at human tournament time controls.

The program will not be a giant killer but it might be a novelty for humans of a certain strength.

I'm thinking perhaps 200 - 300 variations to start with human opposition in mind. There could even be different versions of the book to change-out the openings if desired. A final book could be much larger but I'm just testing at the moment.

Any help or advice would be appreciated.
http://rebel13.nl/dump/hg440.txt

It's from the Mephisto HG40 module, a move with a "-" sign means don't play.
90% of coding is debugging, the other 10% is writing bugs.
User avatar
mhull
Posts: 13447
Joined: Wed Mar 08, 2006 9:02 pm
Location: Dallas, Texas
Full name: Matthew Hull

Re: A Book For a Vintage Program

Post by mhull »

Rebel wrote: Tue Oct 20, 2020 9:35 pm
mhull wrote: Mon Oct 19, 2020 9:00 pm I'm looking for an opening book for a vintage computer program from the late 1970s. I'm guessing this program has not played any public games since 1978. It will likely not search deeper than 9 plies at human tournament time controls.

The program will not be a giant killer but it might be a novelty for humans of a certain strength.

I'm thinking perhaps 200 - 300 variations to start with human opposition in mind. There could even be different versions of the book to change-out the openings if desired. A final book could be much larger but I'm just testing at the moment.

Any help or advice would be appreciated.
http://rebel13.nl/dump/hg440.txt

It's from the Mephisto HG40 module, a move with a "-" sign means don't play.
Thanks Ed! Excellent.

Do you know if the "-" is a prefix or suffix?
Matthew Hull
chrisw
Posts: 4315
Joined: Tue Apr 03, 2012 4:28 pm

Re: A Book For a Vintage Program

Post by chrisw »

mhull wrote: Tue Oct 20, 2020 10:23 pm
Rebel wrote: Tue Oct 20, 2020 9:35 pm
mhull wrote: Mon Oct 19, 2020 9:00 pm I'm looking for an opening book for a vintage computer program from the late 1970s. I'm guessing this program has not played any public games since 1978. It will likely not search deeper than 9 plies at human tournament time controls.

The program will not be a giant killer but it might be a novelty for humans of a certain strength.

I'm thinking perhaps 200 - 300 variations to start with human opposition in mind. There could even be different versions of the book to change-out the openings if desired. A final book could be much larger but I'm just testing at the moment.

Any help or advice would be appreciated.
http://rebel13.nl/dump/hg440.txt

It's from the Mephisto HG40 module, a move with a "-" sign means don't play.
Thanks Ed! Excellent.

Do you know if the "-" is a prefix or suffix?
Looking at the F2F4 lines in the text file. Almost certainly a suffix. Eg don’t initiate any of the F2F4 lines
User avatar
mhull
Posts: 13447
Joined: Wed Mar 08, 2006 9:02 pm
Location: Dallas, Texas
Full name: Matthew Hull

Re: A Book For a Vintage Program

Post by mhull »

chrisw wrote: Tue Oct 20, 2020 10:27 pm
mhull wrote: Tue Oct 20, 2020 10:23 pm
Rebel wrote: Tue Oct 20, 2020 9:35 pm
mhull wrote: Mon Oct 19, 2020 9:00 pm I'm looking for an opening book for a vintage computer program from the late 1970s. I'm guessing this program has not played any public games since 1978. It will likely not search deeper than 9 plies at human tournament time controls.

The program will not be a giant killer but it might be a novelty for humans of a certain strength.

I'm thinking perhaps 200 - 300 variations to start with human opposition in mind. There could even be different versions of the book to change-out the openings if desired. A final book could be much larger but I'm just testing at the moment.

Any help or advice would be appreciated.
http://rebel13.nl/dump/hg440.txt

It's from the Mephisto HG40 module, a move with a "-" sign means don't play.
Thanks Ed! Excellent.

Do you know if the "-" is a prefix or suffix?
Looking at the F2F4 lines in the text file. Almost certainly a suffix. Eg don’t initiate any of the F2F4 lines
Ah! So no Reti lines either. Thanks for the excellent clue!
Matthew Hull
User avatar
Rebel
Posts: 6991
Joined: Thu Aug 18, 2011 12:04 pm

Re: A Book For a Vintage Program

Post by Rebel »

Yes, suffix.
90% of coding is debugging, the other 10% is writing bugs.