Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Compression tools

Author: Heinz van Kempen

Date: 11:44:09 09/28/05

Go up one level in this thread


On September 28, 2005 at 14:30:31, Günther Simon wrote:

>On September 28, 2005 at 14:03:03, Heinz van Kempen wrote:
>
>>On September 28, 2005 at 13:41:41, Uri Blass wrote:
>>
>>>On September 28, 2005 at 13:23:31, Heinz van Kempen wrote:
>>>
>>>>On September 28, 2005 at 13:15:37, Günther Simon wrote:
>>>>
>>>>>On September 28, 2005 at 09:30:36, Heinz van Kempen wrote:
>>>>>
>>>>>><<And where exactly are the games WITH comments in that page?
>>>>>>I mean there are 3 thousant links there, so i'm confused.....:-)
>>>>>>
>>>>>>Does the: |||CEGT 40/40 (2Ghz) all games so far (> 12 Mb)|||  include the
>>>>>>depth,search...,etc......... or is in another file?>>
>>>>>>
>>>>>>
>>>>>>Hi George,
>>>>>>
>>>>>>the games with Fritz 9 are with comments, second file from above.
>>>>>>
>>>>>>The 12 Mb file contains all +37000 CEGT 40/40 games up to now and is unstripped
>>>>>>from comments. Otherwise the file would be 100 Mb.
>>>>>>
>>>>>>http://www.husvankempen.de/nunn/downloads.htm
>>>>>>
>>>>>...
>>>>>
>>>>>Hi Heinz,
>>>>>
>>>>>Is it possible that a lot of 'commented' games _don't_ include
>>>>>the depth?
>>>>>
>>>>>Best regards,
>>>>>Guenther
>>>>
>>>>Hi Guenther,
>>>>
>>>>as already posted that 12 Mb file with more than 37000 games does not include
>>>>comments. Guess how big such a file would be with comments.
>>>
>>>Let see one comment
>>>[%eval 7,16] [%emt 0:01:15]
>>>
>>>it can be compressed to 7,16,75
>>>depth is usually 10-17 and I guess that it can be compressed to average of
>>>something near 3 bits.
>>>
>>>evaluation probably can be compressed to something near 8 bits as average(there
>>>are big numbers but they are minority) and time can be compressed to 7 bits.
>>>
>>>total number of bits is near 18 bits per comment.
>>>
>>>I claim that
>>>24 bits that are 3 bytes are needed for both comment and move.
>>>
>>>explanation:
>>>If we talk about a single move then it can be compressed to 6 bits because we
>>>need only move generator that generate the moves in specific order and usually
>>>there are less than 64 moves so it is 6 bits per ply(if the number of moves is
>>>not more than 32 then it is even only 5 bits.
>>>
>>>37000 games probably include something near 4,000,000 plies
>>>
>>>4,000,000 plies*3 byte=something near 12 Mbytes.
>>>
>>>My conclusion is that 12 Mbytes are enough to compress all the games with
>>>comments and 3 mbytes are enough to compress all the games without comments.
>>>
>>>If you need more than it then the compression tools that you use are not good
>>>enough(for example they probably do not translate moves in the pgn to numbers).
>>>
>>>Uri
>>
>>Hi Uri,
>>
>>Winzip is not good in compressing CB and Arena pgn files including comments.
>>WinRAR is only a bit better here. An alternative would be .cbv files with a high
>>compression rate,  but not all people use the ChessBase 8 or 9.
>>
>>Best Regards
>>Heinz
>
>There is a much better one especially for PGNs, so if you have lots
>of big PGN files for CEGT now, you could try the free compressor by
>George Lyapko. (3 times better compression than e.g. WinRar for PGN)
>http://www.geocities.com/lyapko/winboard.htm
>
>Guenther



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.