Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: YARMOQ! Yet Another Root Move Ordering Question

Author: Dann Corbit

Date: 13:02:34 05/03/01

Go up one level in this thread


On May 03, 2001 at 15:38:44, Dann Corbit wrote:

>What is the target percentage for getting the first move right?
>90%? <<this is pretty much a minumum goal, isn't it?>>
>95%?
>99%?
>Besides node counts for the move in question, what other schemes have been
>tried. (time spent analyzing the move in question is obviously an equivalent to
>node counts).

From the mega-thread below, I percieve that there is one other way to do move
ordering that is commonly tried: an estimate of the score.

So... How does one arrive at this estimate -- have an alpha/beta/gamma where
gamma is returned by the function?

Is there any method besides:
1.  Node counts/time spent
2.  Estimate of goodness {with a magical incantation to achieve it}

>What about the very first pass?  How important is ordering _before_ you have any
>sort of node count, and what do you use to order before anything has been
>searched?

OK, on the 1st pass thingy -- GCP's response shows me how silly I was even to
ask.



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.