Author: Robert Hyatt
Date: 07:48:18 11/22/02
Go up one level in this thread
On November 22, 2002 at 06:35:56, Gian-Carlo Pascutto wrote: >On November 21, 2002 at 20:01:36, Robert Hyatt wrote: > >>You said it was completely useless to you. I don't see how since he didn't add >>or subtract anything to change his raw search speed. Meaning fixed depth is a >>reasonable metric to measure and report. > >But what does it say? Nothing, since the pruning rules changed. > >Moreover, depending on the rest of the program, such a change _can_ >change raw search speed in ways that are not always immediately obvious. > >-- >GCP It says that R=3 + verification produces smaller trees than R=2 normal search, which means that R=3 + verification is faster since it is the same program doing the tests. And R=3 + verification solved more test positions. Seems clear to me that for his program, it is a win in terms of both speed and accuracy. Whether it works for others is another issue. It isn't uncommon to see something work here and not work there...
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.