Author: Vincent Diepeveen
Date: 03:24:08 12/08/00
Go up one level in this thread
On December 07, 2000 at 16:09:32, Ernst A. Heinz wrote: >>If you use _any_ kind of pruning, you can perturb the root score. Thanks >>to the transposition table that grafts parts of the tree into odd places. >>There is no way to anticipate which branches are useless _here_ but the >>hash results might improve the results over _there_. And as a result, you >>can get different root move scores no matter how "sound" the pruning. > >But the effects you describe, Bob, are rather due to the >hash-table usage than to the pruning. They might also >occur if you just change your move ordering. > >Anyway, in the case of futility pruning at frontier nodes >with a remaining depth of 1 ply they hardly matter (if at >all) because it simply lifts inevitable "stand-pat" beta >cutoffs at horizon nodes up one level in the search. > >=Ernst= This depends upon the margins you use and the positional score your evaluation can deliver. If i prune on a few pawns last ply but a certain queen move near opponent king can deliver because of the piece configuration already around the king like +10 pawns positionally, then i obviously have a problem. If capturing a pawn after which my pawn gets a connected passer at say for example the 7th row or 6th row, then we have another obvious case of a move which possibly gets a huge positional score, then the futility again is pruning wrong and possibly influencing the root score. There are a lot of less obvious cases which also go wrong. Vincent
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.