Author: Ratko V Tomic
Date: 13:14:06 10/15/99
Go up one level in this thread
>>Roughly >>estimating, there are perhaps 3-4 non-junk moves in a typical middlegame >>position and the effective branching factor may be 6-8 in such position. > >I've never seen such a high branching factor in any sort of typical position. >3-4 (even sometimes 2) is a much more typical branching factor for today's >programs. > The best you can do with the full width search (I was not counting extensions here, which are highly selective) is the square root of the number of legal moves, which in a middle game would be typically 35-45 or so, giving thus B as stated earlier. With some form of forward prunning, you can do better, of course. Do you have some branching factor figures for the stronger among the selective search programs? In any case, as long as the effective branching factor is greater than the average number of non-junk moves per node, the percentage of non-junk nodes examined goes exponentially to zero as depth increases. While I have put above 3-4 as non-junk moves, that may be a bit generous as well. Experiment, along the lines suggested earlier, would probably answer this question reasonably accurately.
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.