Author: Don Dailey
Date: 14:53:03 01/03/98
Go up one level in this thread
On January 02, 1998 at 17:35:29, Thorsten Czub wrote: >>It's my understanding that Genius has a full width base. > >I understand. But I don't understand WHY this is your believe ! Because Richard told me this years ago. It was one of the few things he "gave up" about his program. >I have always (for 10 years or more) believed in an asymmetric search. >Strange that 2 persons can come to the opposite ideas... > > >> I do not >>believe there is any asymetry going on here at all. > >Ok- but why not ?! > > >> It may appear >>that way because if you lose several ply of search on a problem >>because of asymetry it will not be obvious why. But it's my >>opinion that any problem is guaranteed to be solved eventually >>by Genius. > >But why finds genius the problem earlier when going back one play. If it >has a full width-search, isn't this a contradiction ? shouldn't it be >vice versa ? I said it has a full width BASE, not a full width search. Here is why. In my diagram I am assuming a critical move is being prunned at the 11th ply. An F means full width (all moves) and a '-' means severe pruning. This "critical" move prevents the tactic from being seen. I'm marking the critical move with a '**' In this example Genius is doing 4 + 12 selectivity: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 F F F F - F - F - F ** F - F - F If Genius needs to see the 11th move it will require 7 more iterations or 11 + sel! But take back a move and it will see it instantantly: 2 + 12 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 F F - F - F - F - F - ** - F - F Do you see how the critical move is now on the 12th ply? And the 12th ply is no longer prunned. The full width base (which I believe is there and Richard confirmed) has nothing to do with this behavior. If you are right, that pruning happens right from the very beginning, the behavior will be the same except that Genius would NEVER see this particular tactical shot. I believe Genius will eventually see any tactic, like most of our programs, and if I'm right this means it has a full width base meaning that on each iteration it expands on the amount of full width searching it does. -- Don
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.