Author: Omid David Tabibi
Date: 09:48:05 08/02/03
Go up one level in this thread
On August 02, 2003 at 12:15:18, Sune Fischer wrote: >On August 02, 2003 at 11:12:29, Tony Werten wrote: > >>>To summarize, a score can be considered "exact" only if it has been achieved >>>independent of alpha and beta bounds. This is the case only if we are at a leaf >>>node with full evaluation, or other node with alpha < value < beta. >> >>Or when a checkmate score is returned or a drawscore or .... > >This is not true. > >Consider a case where beta<=0, the first move you search bring back a draw score >(say, forced repetition). >Now you fail high without checking the rest, but you could have had a mating >move or another winning line you didn't search. So it will just be a lower >bound. Again, the decision was based on the bounds, so it was by no means an exact score. When you encounter a threefold repetition, you may store it as exact score in the current node, but when you return 0 to the father, you treat it there just as a normal score, so if it is not between alpha and beta, you cannot store it as exact. (Looking at my code I see that I don't store repetition draws in the hash table at all, I have commented it out, but I'm not sure why :) > >Mate scores are another matter I think. > >>Sorry if I tell you something new, this is old stuff. >> >>When not doing any extensions ( not realistic but anyway) the values returned >>from the last 2 plies from normal search are exact. (As was published 15 years >>ago) > >I don't believe that, and I think my example above shows it. > >-S. > >>Tony >>
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.