Author: Rich Van Gaasbeck
Date: 13:40:26 03/08/02
Go up one level in this thread
> >No >CL1=5s, 2 plies >CL2=15s, 3 plies >CL4=20 minutes, 4 plies >CL5=24 hours, 5 plies > > >it means effective branching factors of 3,80 and 72. > >Uri You're right. Those don't make sense. What sort of branching factors would be reasonable? For minimax, what do people use 35? I seem to recall for Alpha-Beta with ideal ordering you get 3 or 4? There is no interative deepening, so it can't use previous values to sort moves. It has some idea of strong moves, since we know it does selective search on some of the levels. It's possible (given the year it was built and the rom size) that it could be using a killer move for ordering. What kind of effective branching factor would that give? Anyone else want to take a crack at matching search depths with times. Guess I didn't do a very good job :-). Rich
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.