Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Shredder 7 === Review by Stephen Ham [ChessCafe.com]

Author: Stephen Ham

Date: 08:14:50 02/27/03

Go up one level in this thread


On February 27, 2003 at 00:38:31, K. Burcham wrote:

>
>
>"I also found that its ability to evaluate positions was not equally good. In
>fact, I think Fritz 6a and Fritz 7 have more accurate evaluation functions".
>
>Hello Stephen,
>      could you post some examples where you think Shredder has eval not as
>accurate as Fritz.
>
>best to you,
>kburcham

Hi K!

Actually, I did exactly that in my review. There I sited several examples.

http://www.chesscafe.com/REVIEWS/books.HTM

In fact, I ran many more tests, using my own Coorespondence Chess games as test
subjects, so that I could be reasonably certain that I understood what the
"real" evaluation was. But, since my review ran quite long anyway (some thought
the final draft was already too long), I only sited some examples and then
summarized that my other tests confirmed the conclusions from these examples.

I think Robin Smith illustrated the situation clearly (thanks Robin!) when he
wrote that Shredder 7's evaluation was too "optimistic." If an advantage exists,
Shredder 7 often magnifies that advantage out of proportion to reality.

Also, to some degree, Shredder 7 changes its assessment with each iteration and
move more than Fritz 7, which I used for a comparison benchmark.

For many people who merely want the engine to find strong moves, then Shredder 7
is an excellent choice in open positions. However, I was disappointed in it's
play in closed positions with locked pawn chains (see my review). But I use a
chess engine merely to test my own opening novelties. Often I think I've found
"improvements" and so want a strong and objective partner to confirm/reject my
ideas. And if my position is indeed better, how much better is it? Do I have an
edge or a clear advantage? That's why I'd like my partner to have a trustworthy
evaluation function to compare to my own assessments.

Hopefully I've clarified this better for you, K. Thanks for your interest.

All the very best,

Stephen





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.