Author: Scott Gasch
Date: 08:55:30 09/09/04
Go up one level in this thread
Andrew, Let me see if I understand what you are saying: You reach some node N, search its subtree, store a value/bound in the hash table. Later you reach the same node N under a null move. Because of the reduced depth you hit the hashed value of the node. This causes you to miss some mate/tactic. The fundamental problem here, I think, is that the subtree grafting is "wrong". You have, in essence, grafted the subtree searched below node N the first time onto the second point in the search tree where you encounted the same node. So my question for you is: why is it wrong? What do you see the second time if you let it search the subtree instead of take the hashed score? Rh8#. _Why_ do you see this the second time and not the first? Some extension? Some other interaction with the hash table? If you can determine this then maybe you can add this information to your hash table and avoid this type of situation. Interesting problem, I'm sorry I don't have any other advice. Good luck! Scott
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.