Author: leonid
Date: 16:23:46 06/11/99
Go up one level in this thread
On June 11, 1999 at 10:46:07, Daniel Karlsson wrote: >On June 10, 1999 at 19:46:52, leonid wrote: > >>How many nodes per ply game is forced to see in order to "understand" the >>situation? If in the ply we have 35 nodes, how many program must see? >>Five, six, seven... > >In some nodes you get a cutoff on the first move, in others you have to search >all legal moves to get a score. > >Assuming no transpositions, constant branching factor, perfect move ordering and >so on, the average number is around sqrt(moves) using alpha-beta pruning. So six >to seven moves is about right for many middlegame positions. > >Clever pruning techniques can further reduce this number by skipping parts of >the tree. The desciption obove goes very well with my game development. Had the impression that person could see my game. Number of the nodes that logic see in the middle of game is also right. After my statistics, logic is forced to see around 22% of the nodes from each ply. I don't count the highest ply and two lowest plys that have some special features. If in the ply we have 30 nodes, logic will see 6.6 nodes. The next very curious thing for me to find is: At what spead (positions per second) usual, or strong game can see at this stage? Thanks for response! Leonid.
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.