Author: Robert Hyatt
Date: 21:22:03 05/28/03
Go up one level in this thread
On May 28, 2003 at 17:24:56, William H Rogers wrote: >On May 28, 2003 at 17:07:23, Andreas Herrmann wrote: > >>On May 28, 2003 at 16:25:54, William H Rogers wrote: >> >>>I'll give you only one or two examples. >>>Initially there are 20 possible moves for white and assuming that the Kings pawn >>>makes the first move then there are 20 possible moves for black. >>>That equals 20*20=400 moves for a two ply search. >>>Continuing on with the same opening moves now there are 29 possible moves for >>>white once again, that equals 400 * 29 = 11,600 moves for three plys. >>>Then numbers grow from that point up. Set up a board and try it yourself. >>>His statement is that at two plys there are a total of 420 moves. He is getting >>>confused by adding the original 20 to the new 400 possible moves that already >>>include the original 20, and he carries this flawed logic throughoput his >>>equations. It is a simple matter of doing the math... >>>I hope this helps to see the errors of his ways. In all other respects his page >>>looks real good, and I am sure that some people will benifit from it. >>>Bill >> >>There is nothing wrong. Please read his description. If you count only the >>frontier nodes, than you got the perft value. If you add all moves from ply 1 to >>n then you got the total moves count. >>And his perft values are verified by a lot of engines, which supports the perft >>command. You can verify this for excample with Crafty, Yace, Gaviota, Pepito and >>a lot of other engines. >> >>Andreas > Then I don't understand because 20 * 20 = 400 not 420 as he brings the totals >to the right of his boxes. The original 20 are included in the 400. There must >be more than what I could figure out because that is the only things that makes >sense to me. >Bill In a two ply search, at ply 1 you have 20 moves. At ply 2, for each positin you reach (there are 20 of them) you have 20 possible responses. That is a total node count of 420. 20 of them caused by playing one of the 20 legal root moves, 400 of them caused by playing each of the legal replies to those 20 moves...
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.