Author: Geoff
Date: 01:53:02 06/08/03
Hi
I haven't so far come across a proper explanation of how EGTB's work in
principle. Can anyone explain it in a sentence or two please ?
I will have a stab at how I think it might work.
Typically, for 5 or less pieces each position is stored in the table with just
the board position and a mate in x moves figure. The program can then just read
these figures for each possible move and navigate its way to a shortest mate.
But there are some more detailed questions such as
1) What format is each position stored in ?
Does it just store position for say white and infer info for black ?
2) Does it store the whole position or use some symmetry tricks ?
3) How were the tables created in the first place, quite a tricky algoritm I
would imagine !!
Without sitting down and trying to do the maths I had naively assumed that the
tables would all fit in 10 Megs or so. It is surprising, well to me at least
that the tables are so big
regards Geoff
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.