Author: Dann Corbit
Date: 15:09:07 06/18/99
Go up one level in this thread
On June 18, 1999 at 17:51:36, William H Rogers wrote: >On June 18, 1999 at 14:21:13, Dann Corbit wrote: >>On June 18, 1999 at 09:40:00, William H Rogers wrote: >> >>>On June 17, 1999 at 17:30:00, Dann Corbit wrote: >>> > >>>>Ply: Number: >>>>---- ------- >>>>0 1 >>>>1 20 >>>>2 400 >>>>3 7,602 >>>>4 101,236 >>>> >>>>I'm doing ply 5 right now. >>> >>>Just what exactly is a "distinct/unique" position and what are the proposed >>>moves for both sides leading to your numbers.? >>Distinct/unique removes all positions that are exact duplicates generated >>through a different path. The normal ply counters just add up all the legal >>positions that can be generated. But I have done a study to find out how many >>*different* positions there are. It is a *lot* less than the set of legal >>positions with duplicates, and seems to shrink faster and faster with ply count. > >I am having problems with the above calculations, i.e. >Ply Number > 1 20 - assume e4 > 2 20 x 20 = 400 - assume e7 > 3 29 x 400 = 11,600 > 4 29 x 11,600 = 336,400 >Those represent all legal moves, however 4 depends on the move made at 3, it >could be a little smaller. I don't think that there are any duplicates in this >list. These are the nodes my program generates and a few others that I can think >about. Although some programs use forward pruning and some would not consider >some of the moves as they are of less value than others. >Where am I wrong or misunderstanding? What I am trying to show is *not* how many legal moves there are as we go from ply to ply, but how many distinct legal positions. They are altogether different. See the other branch of the thread for a brief explanation. My calculations are still definitely long, but the numbers can be used as a ceiling.
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.