Author: Uri Blass
Date: 22:14:00 06/07/02
Go up one level in this thread
On June 07, 2002 at 19:21:52, Russell Reagan wrote: >On June 07, 2002 at 15:39:53, Gian-Carlo Pascutto wrote: > >>On June 07, 2002 at 15:37:59, Russell Reagan wrote: >> >>>the only thing making >>>it hard is it's incredibly high branching factor when trying to analyze >>>variations. >> >>'Ahem' > >If chess were played on a 19x19 board with 38 pieces on each side, it would be >extremely more complex as well. I do not think that in that case computers are going to play so poor that humans can beat them after one year of learning. I guess that the main problem is that humans know better to explain their evaluation function in chess relative to go. In chess counting material can in most positions help to get a good estimate for the evaluation when it is not the case in go. With 38 pieces for side in chess there may be 100 legal moves per side. It can make the effective branching factor bigger but even with effectiive branching factor of 8 computer may search 7 plies in the middle game and even in regular chess a depth of 7 plies is clearly better than rating of 1500 and when the board is bigger 7 plies for humans is going to be even harder. The game of go is on a complexity level of >tic-tac-toe. If tic-tac-toe were played on a 19x19 board it would be complex >too. I do not think that you are right. The difference is the target of the game. I guess that a simple evaluation function can practically help to solve tic tac-toe in every dimension(I read that it is solved in another post). Uri
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.