Author: Vincent Diepeveen
Date: 12:25:54 08/12/99
Go up one level in this thread
On August 12, 1999 at 14:16:56, Robert Hyatt wrote: >On August 12, 1999 at 13:11:45, Vincent Diepeveen wrote: > >>On August 11, 1999 at 22:31:43, Bas Hamstra wrote: >> >>>Vincent, >>> >>>Your story goes for the starting position, yes. But how about the WAC suite for >>>example? According to my experiments the difference becomes much less then... >>> >>>What other positions did you test on? And what were the results? >> >>for the always returning 0 >>version exactly the same of course. > > >it isn't quite the same. At move one white has 20 moves. The branching >factor is low. At move 20 it is possible that white has 50 moves. That >makes the tree _much_ larger, even if it is searched perfectly in order. > So that means that it still returns 0, but at move 0 it's searching more nodes as it GETS to those 50 move positions, where at move 50 it exchanges a lot of material on average which gets it to positions with say 10 or at most 20 possibilities, then hashtable does the rest... Greetings, Vincent > > >> >>But the tactical version was of course not tested on other positions, >>as my whole experiment was only to see what was the minimum of nodes i >>could need, and as we know in openingsposition one can sort very >>well as white can prevent tactics. >> >>I don't remember how deep it searched in other positions, as very >>soon i added piece square tables too. >> >>You might want to see at the internet and look for misssilicon and >>talk to Andrew. It gets to real huge depths in positions very quickly. >> >>It's tactical very strong :) [cutted down]
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.