Author: Ulrich Tuerke
Date: 04:54:59 01/28/00
Go up one level in this thread
On January 28, 2000 at 07:39:17, Mark Taylor wrote: >In null move, (as I understand it- please correct me if I'm wrong), a player >misses a move (i.e. lets the opponent play 2 moves in succession), and if no >significant change in the eval occurs then further searching from that node is >abandoned (on the basis that if you can up a move without suffering then your >position must be solid)? > >My question is this - being forced to make a move in Chess can be disastrous in >certain positions (esp. in the endgame), and in this type of position misleading >results would be obtained. You have very well recognized a critical point in the assumptions leading to the null move algorithm. Nullmove has problems in zugzwang situations. In practice, most of the programs do not use null move search in endings; others use nullmove but verify the result by a shallow conventional search. In my program, I apply nullmove only when the total material is above a certain threshold and when the generated move list does not seem too small. However you can never be sure; such kind of zugzwang situations can also occur in middle game though they are rare. This has sometimes disastrous consequences. By the way , position nr. 9 from gs-2930 is a nice example: i think that program which do still use nullmove after exchange of the rooks, do not have a chance to find the solution Rd6. Uli > >It seems to me that a better approach would be to make a single move (any move) >rather than no move at all - the resulting tree would be the same size and >therefore the overheads should not be that much greater. > >Can anyone help me with an explanation? > >Thanks, > >Mark Taylor.
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.