Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Rebel wins Shep Championship 99

Author: Thorsten Czub

Date: 05:23:37 09/20/99

Go up one level in this thread


On September 20, 1999 at 08:01:34, Shep wrote:

>Since I have had several emails from Chris about this tournament and he never
>mentioned anything in that direction, I guess I have permission. :)


:-)) we will see what the programmer says afterwards :-))

>As for the "CST 2.50 broken" issue, I will make some experiments later this week
>and report the results here. I have no reason not to believe Thorsten, but I
>want to see for myself before I add a comment in that respect on my site.

easy way to find out about:
use 2.03 and play the 2.50 games on the same machines under the same conditions
from the same positions and wait what happens. if this does NOT change the
result positively than i am wrong, if it changes the result positively (as
i guess) you will understand that the version is broken.
I an also explain you why (if you are interested).
Version 2.5 has a new feature done by a user-interface programmer.
that is the styling-feature.
it works not accurate. some weightings seem not to work as the style should
set them. the user-interface programmer was - of course - unable to find
out about, because he does not play games and test-series with it.
thats chris' and my job. we found this out later.
since this version (2.5) was only for programmers to download to implement
their engines into CSTal, chris and i never thought this would be a problem
because it is said on the web site that version 2.5 is only for programmers
and not for testing purposes. we - of course - never thought that people
cannot read but download without reading what is written.
Next time we put in the license-agreement: the user/customer has the capability
to read and understand written-text on the package and user-manual.

:-))


>BTW, I think it's interesting that everyone involved in the "2.03 vs. 2.50"
>issue (including Thorsten and yourself) had enough time to write dozens of
>postings but never took a few minutes to run some test positions to prove that
>either 2.50 does play different from 2.03 or that it doesn't... :)

Test positions are not very much disturbed by wrong styling. The key moves
is often found via extensions and these extensions are often not
influenced by styling. the best way to find out about the difference is:
to play with the version 2.03 from the same positions 2.5 did and watch out
the difference.
Test-positions are not the same thing games are. in a game there is often no
key move to find and the styling gets more important.



>
>---
>Shep



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.