Author: Peter Fendrich
Date: 10:55:44 06/07/99
Go up one level in this thread
On June 06, 1999 at 16:29:14, Jeroen van Dorp wrote:
>Some questions about determining reproducable and somewhat reliable comparison
>between engines and their strenght.
>
>I think I read here on CCC that in engine - enigine games pondering should be
>off.
>Is that so, and if yes, why?
Only if you are playing engine-engine on the same machine. Pondering takes cpu
power and memory from the opponent and it's difficult to know what the real time
settings are. If the engines have separate machines I see no reason to turn
pondering off to let the programs play in the mode they are tunesd for.
>
>If you're disabling opening book(s) -apart from the randomness of choosing a
>good or bad opening- will it tell me more of the relative strenght of the
>engines?
I know that there are different opinions about this but I don't like the idea of
turning the book off. My reasons are:
- The book is part of the chess program and the program is tuned to use
it. Of course it is "unfair" when someone has a bigger bok but it's also
a matter of how the book is used, how the learning is working and
how well the book guides the program to positions it manage to handle.
It's as well "unfair" to turn a good such ability off.
- If both programs have books, the play tends to be more variable than if
they start with no books at all. A variable play is really meassuring
the engine capacity better in the end, despite the books.
IMO it's more interesting to avoid setting any unnecessary constraints and let
the programs play with their maximum strength, rather than trying to isolate the
"pure" engines in order to find out some more abstract comparison.
What will you use it for? :-)
//Peter
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.