Author: Andrew Dados
Date: 13:43:39 07/15/04
Go up one level in this thread
On July 15, 2004 at 15:16:55, Gopikrishna wrote:
>Fail highs occur more often than Fail Lows right? I mean in a good searching
>normally Beta cut offs % is around? Thanks.
(for the moment lets disregard 'exact nodes, with non-zero alpha-beta window)
Consider this: for node to be a fail high (FH) node all its children must fail
low (and all of its children must be searched).
For node to be a fail low (FL) node only one of its children must fail high (so
we stop at first failing high move).
Those nodes alternate in search when going 'down depth' :
node: FL-FH-FL-FH-...
depth: 0 1 2 3 ...
What is the ratio of FH to FL nodes depends on how good is your move ordering:
in ideal move ordering it would be close to 50% while in real move ordering
20-30% of FH is good.
This above has some implication in replacing schema for transposition table:
always replace FL nodes with FH nodes for same remaining depth ('draft') because
FL nodes are less costly to compute (cost of FH at depth=d is close to cost of
FL at d+1).
-Andrew-
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.