Author: Steve Maughan
Date: 09:14:15 01/09/00
Go up one level in this thread
Bruce A root processor program tries to assess the knowledge that will be pertinent to a position BEFORE it starts to search ie at the root. For example if the black king is on g8 then it may be a good idea for white to have a rook on the g file. The program can then create a table of rook bonus points where the g file gets a high bonus eg a rook on g2 may be worth +5.20 pawan while a rook on a2 may be worth 4.8 pawns. All of this knowledge is integrated into this lookup table, which is only created once. >and what are the advantages/disadvantages of such a thing? Since the knowledge is in a simple lookup table, programs that use this approach are usually fast. The downside is that they cannot cope with dramatic changes to the position and as the search gets deeper the difference between the root nodes and the leaf nodes increase. This results in 'inaccurate' evaluation. For example in the above case, if the black king moves from g8-f8-e7-d6, white still thinks that a rook on g2 is worth 5.2 pawns. Hope that helps! Steve Maughan
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.