Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Hmm... your info is not clear enough.

Author: Mike S.

Date: 16:13:15 03/02/02

Go up one level in this thread


On March 02, 2002 at 16:01:24, Eran wrote:

>Jorge,
>
>Your info does not show me clearly how well Fritz7 with the new settings does.
>Your info about games must be in PGN format please.
>Also, you need to create two tournaments, Fritz7 vs. Shredder6, with two
>different Fritz7 settings each, new ones and default. So that I can compare the
>Fritz7 difference of strength.

How do you handle the openings in your comparison matches?

To compare the performance with the default setting, or any other setting, it
would be necessary to play from the *same opening lines* than in the comparison
match. Without that, I think the openings chosen will have more influence than
the settings.

If I would do that (which I don't), I'd set the priority to a higher number of
games rather than to long time controls. I think, 120 games from 40/40:00 give a
more reliable result that 40 games from 40/2h. Furthermore it would probably be
useful to use a set of opening positions similar to the Nunn openings (as an
openings database) for that experiment.

Regards,
M.Scheidl

(my message 216068 disappeared, I'm curious what will happen with this one :o)



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.