Author: J. Wesley Cleveland
Date: 10:24:45 09/18/03
Go up one level in this thread
On September 18, 2003 at 11:15:13, Edward Seid wrote: >On September 18, 2003 at 11:05:15, Joel wrote: > >>Actually, the more I think about writing an evaluation function for that game, >>the more easy I think that problem is. > >Maybe this is a stupid question, but I'm new to this so I'll ask anyway... > >What I'm proposing is a minimax traversal of the move tree, all the way to the >terminal node (position) where it's either a win for White, a draw, or a win for >Black. Then that result gets minimaxed back towards the initial position. As I >see it, there is no doubt what the assessment is... either is it a win, loss or >draw. > >Having said that, is the use of the phrase 'evaluation function' really >appropriate here, since there is no weighting of positional features or material >balance? If you use minimax, then you don't need an evaluation function, but the search time will be too long. If you use alpha-beta (or varients) with iterative deepening, then you need an evalution function for move ordering. If you use use alpha-beta (or varients) without iterative deepening (sezrch deeper until you have a true result), then you don't need evaluation, but bad move ordering may kill the search.
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.