Author: Randall Shane
Date: 06:59:48 10/17/05
Go up one level in this thread
On October 17, 2005 at 06:12:48, chandler yergin wrote: >On October 16, 2005 at 10:02:12, Tord Romstad wrote: > >>On October 16, 2005 at 08:18:48, Maurizio De Leo wrote: >> >>>I tought the branching factor of modern programs was more of the order 3.5. Is >>>the 2.4 value typical of Fruit, or is it due to the particulare position, or I >>>was just wrong and most modern programs have a branching factor smaller than 3 ? >> >>I think you are just wrong. I am fairly sure most modern programs have an >>average branching factor closer to 2 than to 3 in the middle game. >> >>Tord > > The Branch Tree is usually about 35 Branches from any position Full trees from normal positions tend to have a branching factor averaging about 35, that's true, but that's not the issue here. All but the simplest chess problems implement some form of alpha-beta searching to cut down on the branching factor. The entire tree doesn't have to be searched to 'prove' a proposed move at a level is best. Naive implementations of alpha-beta can be expected to have a branching factor of about 6 in most positions. Using good move ordering, hashing, pruning, and many other techniques can reduce this further. The flow of the conversation above your reply concerned how far the branching factor had been reduced.
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.