Author: Dann Corbit
Date: 15:16:27 10/19/01
Go up one level in this thread
On October 19, 2001 at 18:03:35, Andreas De Troy wrote: >I once mentioned a +/- similar idea here, to reduce the storage requirements for >PGN-files (read: to "ZIP them more efficiently): if you agree on a certain move >generator (MG), you could store your move as the number of the move the MG >generates. > >For instance, in the start position, if this move generator produces moves a3, >a4, b3, b4, ... in that sequence, move 1. a4 would be stored as the number 2 >(it's the second move generated). > >You don't have to count the maximum number of moves possible in any position to >determine how many bits you need; you can examine the position (generate all >possible moves) and use just as many bits as you need for that particular >position (for instance, if there are 4 legal moves, you use 2 bits; and since >you know that only 4 moves are possible, you just look at the next 2 bits in >your datafile to decode the "next move"). > >Disadvantage: to decode, you need to know how many moves are possible and that >takes time. > >I opened this post with "I once mentioned a +/- similar idea here, ...". I will >tell you what happened then: some guy replied that this "move NUMBER"-idea was >exactly what was used in the (now old) Chessbase-format. > >There it ended. But indeed -maybe a similar idea can be used in tablebases? That is the technique used in Pigbase -- an Amiga chess database that comes with source code. I don't know why you need to know the distance to mate (which is given as the reason not to store the move). If you make the best possible move, then it does not matter what the distance is. If it is too far away, then you can only draw anyway. Whatever the case, you need to make the best possible move and if that has been encoded, it seems like it should be enough. But I imagine there are horrible holes in this idea also, that I am simply unaware of.
This page took 0.01 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.