Author: Robert Hyatt
Date: 20:01:57 11/19/02
Go up one level in this thread
On November 19, 2002 at 20:32:27, Joel wrote: >Hey all, > >I find the inordinate amount of discussion in regards to the underlying data >structures quite amusing. > >Isn't the fact that there are strong programs based on a variety of difference >approaches good enough to 'prove' that the particular approaches have merit? That is the point. However, one person claims bitboards are inherently flawed. And it simply isn't true... > >Hell, when I first started, I looked into each method somewhat, decided to go >with an array based method, then after that learning experience, decided to >change to bitboards (I think it is elegant) for a more serious attempt at making >a reasonable engine. > >My reasoning for this decision was that if other strong coders could make a >strong chess AI using either approach, then well, I may as well chose the one >which I like the best. (and hope that I am a good coder!) > >What confuses me is that a lot of people here get very involved in these data >structure zealotry discussions! I read numbers such as a 'factor of 2' being >thrown around here and there, but don't fully understand the importance? Isn't >the branching factor in chess suffiently high enough for these factors to make >only a small difference to the programs chess playing performance? The point is that even if a bitboard move generator is 2x slower than an array- based approach, that generally is less than 10% of the total time, so it is not that important. Bitboards offer interesting opportunities in evaluation, as several have mentioned, but it takes some experience and thought to develop the right mindset to come up with elegant solutions to what appear to be hard problems... > >Keep in mind my perspective is slanted - my first serious attempt at a chess >engine is not complete (need a few more weeks), and my only other experience >with chess programming is a quick engine knocked up in a month, plus lots of on >paper design for my current one. > >Regards, >Joel
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.