Author: Severi Salminen
Date: 12:10:47 01/05/01
Go up one level in this thread
On January 05, 2001 at 12:23:06, Gian-Carlo Pascutto wrote: >On January 05, 2001 at 12:12:29, Heiner Marxen wrote: > >>I.e. if you retry the exact same search, you may well get another result, >>because early in the search tree you now find deep results produced >>later in the first search. You find other values in the TT >>(deeper than requested), use them, and... return different values. > >I do not think this is true when the hash table is only used for >move ordering. > >My complaint was that I got the fail high/low effect even when the >only difference was the move ordering and the root search bounds. > >Since move ordering is influenced by other factors than the hash >table, disabling it may very well lead to the same problem. Hi! Today I programmed SEE and got also weird results (now it works). Are you using PVS? My problem was to decide what to use as the best score if all the moves were pruned because off futility pruning, also in qsearch. In qsearch I decided to return alpha and nothing less. I don't know if it is right, but it worked. And in normal search I use material_balance + FUTILITY_MARGIN to alter the bestscore so far. I don't know whether I should include the material gain of a move or exclude the MARGIN, but now things work. Maybe you should check out those and experimant. Severi
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.