Not logged inRybka Chess Community Forum
Up Topic The Rybka Lounge / Computer Chess / Question For Fritz Book Makers---
- - By Italian81 (****) [us] Date 2009-02-17 05:06
Does anyone know how to check for multiple games in a database so that duplicates can be weeded out?

I know how to do this in shredder for .bkt books but not sure how to do so in Fritz for .ctg books

Also when importing games from your database into your new book do you go up to move 100 and check the box to include variations? 

Can someone please explain the difference and importance of importing games in their entire length, rather than selecting the option that is only related to the specific elo of the game. 

Which is better?

I have been currently experienting by adding weights to the moves in my book from 1-10 but it seems to have no effect.

Also a nice thing about .bin books is you have total control to assign a % on a move that should be played for each move in a book.....IE I can say I want 1 e4, 1d4, and 1 Nf3 to all have a 33% chance of being played in a match so it will play those 3 opening moves an equal amount of times in a tournament but I can't figure out how to do this in Fritz

Any help is much appreciated, thank you.  :)
Parent - - By tano-urayoan (****) [pr] Date 2009-02-17 06:08
For your first question you will need a database program: for example chessbase #x or chessbase light premium. I don't think it could be done under Fritz.

Importing 100 moves seems a lot, you must be deep in the middlegame if not in the ending by that point.
Parent - By Italian81 (****) [us] Date 2009-02-17 15:50
Thank you, Tano ;)
Parent - - By Mark Mason (***) [gb] Date 2009-02-17 09:14

Checking for duplicates - as tano-urayoan has said, no you cannot do this using Fritz you need Chessbase 9 or 10 or Chessbase Light  Premium.

Importing Games -  if you tick "absolute length"  then all games will be imported to this length regardless of how common or unusual the opening in question is. This will result in a deep book for all lines. If you tick "ECO-Relative Length" then this limits the variations relative to the ECO classification. This results in long variations for theoretical main lines (ECO classification position found late in the game) and short variations for side lines (i.e., early deviations) - so overall a less deep book.  The reasons I can think to limit in this way are either to save hard disk space or if you don't want to stay in book long for the less common openings.  I always use absolute length to get as deep a book as possible .  By the way there is one very important point here that a lot of people are not aware of and which can cause confusion. When you import games,  the length that you specify is in half moves (plies) not full moves. So specifying length 100 will mean that the book will play to move 50 then stop. Therefore if you want a book to play to the Fritz optimised default of 32 moves deep - you have to use a value of at least 64 when importing the games.  Don't let anyone tell you this is not the case  - it is !   If in doubt just make a very small test book and import to length 10 and you will see it will only play 5 moves deep. However, and here's where it can cause confusion, when you set the book up to actually play using Fritz (on or off line), the number that you set in Book Options for the "up to move" value, is full moves not half moves.

Variations  - if you tick to include variations then this will include any annotated variations (i.e. analysis that might be included in the games), but these are not included in the statistical analysis that affects the moves in the book. Therefore including variations will make the book bigger but will not affect game play - it is only for analysis purposes.

Weights  -  change the 'weighting' of an existing move makes it more/less likely to be chosen. The likelihood of a move being chosen is the product of the weighting (in the FACT column) and the statistical properties derived from the game databse (in the PROB column) which results in a value in the [%] column.. The value is from -125 (very unlikely to be picked) to +125 (very likely to be picked). You can change the weighting manually  to any value in this range. Actually this is what the automatic book learning does based on whether games are won or lost  (depending on how you set the middle slider in Book Options). It sounds as though you need to be setting higher values for your weightings to influence move choice.

Apologies if you already knew all of this !

Hope this helps,

Parent - By Dr.Wael Deeb (***) [jo] Date 2009-02-17 10:22
Thanks Mark,although I am aware of all that you posted,it was a pleasure to go through the thread :-)
Parent - - By Italian81 (****) [us] Date 2009-02-17 14:51
Actually Mark most of what you told me was new news to me thanks.  I learned a lot from your post and I appreciate you taking the time to explain it to me :)
Parent - - By Mark Mason (***) [gb] Date 2009-02-17 19:33
Your welcome. I have been fascinated by this subject for some time now so it's good to be able to share some of what I've learned with fellow enthusiasts.

Parent - By Italian81 (****) [us] Date 2009-02-17 19:37
and the advice is very much appreciated, thank you ;)
Up Topic The Rybka Lounge / Computer Chess / Question For Fritz Book Makers---

Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill