Reply
 
LinkBack Thread Tools Display Modes
  #1   Report Post  
Old January 1st 04, 09:34 PM
RW
 
Posts: n/a
Default question about TBS

TBS(tablebase) in Fritz 6 is 132 MB big. TBS in Fritz 8 is 261 MB big.
Does it mean that TBS in Fritz 8 content 2x more endgame positions in?
Or that the statistics in Fritz 8 are much better then in Fritz 6 but the
number of positions is the same.
Or do the tablebases include all "up to 5 five pieces" positons or not?

I try to open the TBS from Fritz 8 using Scid and some of endings (f.e. 5
piece) are inactive. Am I doing anything wrong?




  #2   Report Post  
Old January 2nd 04, 07:08 AM
Anders Thulin
 
Posts: n/a
Default question about TBS

RW wrote:
TBS(tablebase) in Fritz 6 is 132 MB big. TBS in Fritz 8 is 261 MB big.
Does it mean that TBS in Fritz 8 content 2x more endgame positions in?


Not necessarily -- the tablebase files are compressed, so it's difficult
to draw any direct conclusions from file size only, unless you also happen to
know the degree of compression achieved.

It's better to check the number of files -- each file represents a typical
endgame (such as KRQ vs KQ). More files means more endgames. And you can
also ensure that you have endgames that are likely to pop up in your
games.

Or that the statistics in Fritz 8 are much better then in Fritz 6 but the
number of positions is the same.


Each position is reprensented by one single number -- the score of that position.
Once you have that, you can't improve on it -- well, in some cases,
involving castling it is possible.

Or do the tablebases include all "up to 5 five pieces" positons or not?


The full set of 5-piece endgame is around 7 Gb. The names of the files
tell you which you actually have. There's an Fritz 8 addon with a number of
DVD disks which contain all 5-pieces endgames and some 6-piece endgames,
though you can download the files from some FTP sites as well.

I try to open the TBS from Fritz 8 using Scid and some of endings (f.e. 5
piece) are inactive. Am I doing anything wrong?


Are the inactive endgings, endings for which you have the appropriate files?
If you don't have them, they're not going to be active.

Check and endgame that you *know* you have. If that doesn't work, you probably
have got the endgame table paths wrong. Check Tools | Options | Tablebases.

--
Anders Thulin http://www.algonet.se/~ath

  #3   Report Post  
Old January 2nd 04, 12:47 PM
RW
 
Posts: n/a
Default question about TBS

The number of TBS files in each of the programs is the same.
However, for example krpkrnbw file is in the first program 16MB big and in
the
second program 79MB big!

Of course, if the all "up to 5 pieces" base should have about 7Gb, now it is
clear
for me that some of the endgames are missing in the mentioned tablebases.

That is the reason that some files in Scid are inactive.

Thank you for ypur very valuable answer.



  #4   Report Post  
Old January 2nd 04, 04:04 PM
Anders Thulin
 
Posts: n/a
Default question about TBS

RW wrote:
The number of TBS files in each of the programs is the same.
However, for example krpkrnbw file is in the first program 16MB big and in
the
second program 79MB big!


I've been using these files (among many others) for both Fritz 6 and 8:

krpkrnbw.emd is 81247218 bytes
krpkrnbb.emd is 29506318 bytes

without problems.

Both files together cover the krp vs kr endgame -- the 'nbw' file is for
white-to-move, and the 'nbb' file for black. 'emd' indicates that the files are
compressed.

I don't understand the 16Mb size -- are you sure you got the right file?
The names can be confusing. If you *did* get it right, there's probably
something wrong with that one, and usually the only way you discover it is
that that particular endgame won't use the database files. The only
alternative I know about are the uncompressed files, which don't have an
'emd' extension. But they should not be used, except in special circumstances.

You'll find the authoritative collection of endgame database files at

ftp://ftp.cis.uab.edu/pub/hyatt/TB/

In each subdirectory, you'll also find list of md5 checksums, which you
can use to verify that the files are indeed correct.

--
Anders Thulin http://www.algonet.se/~ath

  #5   Report Post  
Old January 2nd 04, 11:32 PM
HD
 
Posts: n/a
Default question about TBS

RW skrev:
TBS(tablebase) in Fritz 6 is 132 MB big. TBS in Fritz 8 is 261 MB big.
Does it mean that TBS in Fritz 8 content 2x more endgame positions in?
Or that the statistics in Fritz 8 are much better then in Fritz 6 but the
number of positions is the same.


Fritz 6 came on a single cd, but Fritz 8 is a double cd with all the
chat files placed at cd 2, so I believe that CB just wanted to use some
of the extra space for some extra TB's.

Or do the tablebases include all "up to 5 five pieces" positons or not?

I try to open the TBS from Fritz 8 using Scid and some of endings (f.e. 5
piece) are inactive. Am I doing anything wrong?







  #6   Report Post  
Old January 9th 04, 12:39 AM
Dieter Buerssner
 
Posts: n/a
Default question about TBS

Anders Thulin wrote:

Both files together cover the krp vs kr endgame -- the 'nbw' file
is for white-to-move, and the 'nbb' file for black.


This is not totally accurate. In krp vs kr - krpkr.nbw.emd (for example),
nbw does not really mean white to move, but rather the krp side to move.
So, when black has krp and white has kr, and black has the right to move,
one would find the position in krpkr.nbw[.emd]. Of course, from user point
of view (and even from programmer's point of view) this is not that
important. For specific analysis, a user who understands it well, and does
not have all the TBs but wants to download some critical TBs for the
position, it might help to save a bit of download (when the engine he uses
will not have problems with uncomplete TB-sets).

Regards,
Dieter
  #7   Report Post  
Old January 9th 04, 05:37 PM
Mark Hay
 
Posts: n/a
Default question about TBS

HD wrote in message . dk...
RW skrev:
TBS(tablebase) in Fritz 6 is 132 MB big. TBS in Fritz 8 is 261 MB big.
Does it mean that TBS in Fritz 8 content 2x more endgame positions in?
Or that the statistics in Fritz 8 are much better then in Fritz 6 but the
number of positions is the same.

phrocrew
Fritz 6 came on a single cd, but Fritz 8 is a double cd with all the
chat files placed at cd 2, so I believe that CB just wanted to use some
of the extra space for some extra TB's.

Or do the tablebases include all "up to 5 five pieces" positons or not?

I try to open the TBS from Fritz 8 using Scid and some of endings (f.e. 5
piece) are inactive. Am I doing anything wrong?




it is
Reply
Thread Tools
Display Modes

Posting Rules

Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are On


Similar Threads
Thread Thread Starter Forum Replies Last Post
Question on analysis Ralph Jones rec.games.chess.analysis (Chess Analysis) 4 March 17th 04 06:46 PM
Puzzle solving: an urgent question Samik rec.games.chess.analysis (Chess Analysis) 4 February 20th 04 09:25 PM
The "50-move" rule: a question, a remark, and the exact interpretation. Sidney Cadot rec.games.chess.computer (Computer Chess) 6 December 8th 03 07:39 PM
Question About Move Order 1.d4 Nf6 2.Nf3 Nf6 3.e3 edwood rec.games.chess.analysis (Chess Analysis) 7 October 3rd 03 02:08 AM
Chessbase Question Mike Ogush rec.games.chess.computer (Computer Chess) 0 August 11th 03 10:23 PM


All times are GMT +1. The time now is 06:06 AM.

Powered by vBulletin® Copyright ©2000 - 2019, Jelsoft Enterprises Ltd.
Copyright 2004-2019 ChessBanter.
The comments are property of their posters.
 

About Us

"It's about Chess"

 

Copyright © 2017