Author: Dann Corbit
Date: 16:33:48 07/30/99
Go up one level in this thread
On July 30, 1999 at 18:01:15, Terry Ripple wrote: > I noticed that several of the top programs will initially do a brute force >followed by a selective search. For instance, when it shows a 9/30 in it`s >search, does this mean that it`s performing 9 ply brute force and then >performing up to 30 ply selective to complete its search for the position? I`am >using the Hiarcs 7.32 program! > > So, getting back to my question: If my program performs a 9 ply brute force >search, then this means it`s searching every possible move on the chess board to >complete a 9 ply brute force. This seems that it would need to search billions >of moves at times to complete such a task that comes from a brute force search. >This seems almost impossible with a home PC to do this kind of search with a >time control of 40/2hrs., but i know that it is doing this very task as this is >what it shows! > > How many moves or positions does it look at to perform a 9 ply brute force? Here is a repost of an article by Michel Langeveld ML] Subject: FULL ply 9 is 2.439.530.234.167 positions!!!!!!!!!!!!!!! ML] From: Michel Langeveld ML] E-mail: rudolf@stad.dsl.nl ML] Message Number: 57423 ML] Date: June 22, 1999 at 12:11:54 ML] ML] Hello Chess World!! ML] ML] the complete table becomes: ML] ML] Ply(0) Nodes= 1 ML] Ply(1) Nodes= 20 = 20,00x ply(0) ML] Ply(2) Nodes= 400 = 20,00x ply(1) ML] Ply(3) Nodes= 8902 = 22,26x ply(2) ML] Ply(4) Nodes= 197281 = 22,16x ply(3) ML] Ply(5) Nodes= 4865609 = 24,66x ply(4) ML] Ply(6) Nodes= 119060324 = 24,47x ply(5) ML] Ply(7) Nodes= 3195901860 = 26,84x ply(6) ML] Ply(8) Nodes= 84998978956 = 26,60x ply(7) ML] Ply(9) Nodes= 2439530234167 = 28,70x ply(8) ML] ML] Ply 9 is interesting because white-promotions and white long castling is added ML] for the first time. ML] ML] Volunteers for doing ply 10 or verify ply 9 :-P ? Maybe we can calculate this ML] distributedly... ML] ML] Kind regards, ML] ML] Michel Langeveld Michel Langeveld, Andreas Stabel, and Steven J. Edwards post interesting information on node counts from time to time. It is not necessary to do a brute force search of 9 plies to find the best move in that same arena. If you use Alpha/Beta, only a fraction of that many nodes are actually examined, yet the search provably comes to the same conclusion.
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.