Author: leonid
Date: 11:09:58 09/26/99
Go up one level in this thread
On September 26, 1999 at 13:49:21, Antonio Dieguez wrote: >On September 26, 1999 at 07:52:45, leonid wrote: > >>On September 25, 1999 at 16:50:30, Antonio Dieguez wrote: >> >>>a difference between an average ply from 8 to 10, is very much than a difference >>>from 15 to 17? >>>so if a big difference in evaluation is a factor like that it will become to be >>>more and more important in the future? or the difference stays about the same? >>> >>>thx. >>> >>>This is my first post, please be nice with me... Greetings to everybody! :) >> >>Probably your question is if "branching factor" is all the time the same, >>never mind between what plys it was caluculated? If this is the question, my >>response will be yes, in general. At least, this is what I found by trying the >>branching factor for the best games that I bought. All the time games were asked >>to search the position by "brute force". One additional finding - all >>those games had practically identical "branching factor". >> >>Branching factor. If plys five take 1 second and plys six 8 second, branching >>factor will be 8. > >Thanks leonid! my friend... >altough my question were how much easy to win is with 10 plys against 8, than 17 >plys against 15, and so, the eval is more important at deeper search... > >Anyway a factor 8 is very bad... wich is the best branching factor that a >program have? 3.something? > >thx. > >>Leonid. > >Antonio. Everything depend on average number of nodes (moves) that each play have for given position. After what I could see, all best games usually see 16% of the nodes in ply before finding theirs results. Sometime game see even less that three nodes before saying the result and sometime more. Verification was done for search by "brute force". 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.