Author: Vincent Diepeveen
Date: 03:08:52 10/29/02
Go up one level in this thread
On October 28, 2002 at 12:28:06, Dan Andersson wrote: Nullmove works great in GO. It is called nullmove and not pass move reduction. It is not nice in the academic world to rename something that has a name to a new name and claim copyright on it. That is kind of 'not nice' behaviour. Search module is independant from evaluating a branch. Of course it gets influenced by the score of the evaluation, but it is a complete independant module. Why do you confuse it? The first 100+ moves or so obviously you don't have the problem that passing might be a good alternative. I never heard from those misere games, so they are not important i bet. >The main point I try to make is that the name null-move is somewhat of a >misnomer in Go. Pass reduction pruning might be better. The recursive version >might work well. But the main problem is the scoring function. You have to have >a good one for the Pass-Pass move combination. And the opening is the most >troublesome part in Go. As for missing pertinent variations in B*, the same >problem exist in nullmove. It is only less likely to occur due to the fact that >doing a move is superior in allmost all situations. And that the search depth >that can be attained must be lower in B*. A modern B* implementation would be >most interesting to behold IMO. It would certainly kick ass in slow endgames. >The most popular misere games on the top of my head would be Nim and other pile >games, low-ball poker and pass whist. >MvH Dan Andersson
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.