Author: Dieter Buerssner
Date: 13:16:13 11/12/01
Go up one level in this thread
On November 12, 2001 at 13:10:44, Robert Hyatt wrote: [much snipped] >It [space] _could_ be more than 50 for win/lose/draw depending on the indexing >scheme of course. 50 MB is about right. In http://www.icdchess.com/forums/1/message.shtml?197044 I estimated about 35 MB for a rather unsophisticated indexing scheme. >[...] Yet they [the 4-men tables] obviously represent >tables of size 64^4th roughly. Which should be around 16 megs. The nalimov >indexing/compression certainly helps a _lot_ but not as much as the compression >done outside of the indexing. Certainly, compression is a big factor for the "primitive" tables: all draw, or all won, etc. For the other tables, I would assume that the indexing helps more (or at least comparable much) than compression in the pawnless tables. Compare for example 462 * 64^2 to 64^4. This is almost a factor of 9 already without any further tricks. 462 is the number of legal unique KK position, where one K is restricted to a small triangle a1-d1-d4 (it can allways be put there by rotation and/or mirroring), and when the K is on the diagonal, it is restricted to one half big triangle (a1-a8-h8 or a1-h1-h8). Regards, Dieter
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.