Author: Charles Roberson
Date: 08:08:43 12/24/05
Stuart,
I looked it over. Looks nice and well thought out.
However, it is a tuner and it takes effort to base
an evaluation function on it. I implemented a simplified
version some years ago for NoonianChess and it made my
program worse. I tried tweaking NoonianChess' eval parameters
based on the parameters it has and it made it worse.
I must admit this is just my results. Done differently
maybe it would have made it better.
Also, a paper was published in the ICCA journal on the tuning
efforts using the code. Refer to ICCA Journal Vol. 20 No. 4 December
1997, T.S. Anantharaman: Evaluation Tuning for Computer Chess: Linear
Discrimination Methods.
It is a very good read. The conclusions are interesting and some
are obviously good areas to pursue.
Charles
This page took 0.01 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.