Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: What Gambit New Paradigm could be...if it exist

Author: Ratko V Tomic

Date: 16:34:45 10/23/00

Go up one level in this thread


> Changing the evaluation is not changing a number or two in the source code.
> It may be changing hundreds of lines in the source code.

But once you come up with a new function in a prototype you can switch between
the old and the new in a single program using single a number to represent the
relative weight of the two (e.g. 0-1, where 0=use old, 1=use new, else anything
in between). So why would you need two programs and make plans of future
integration. Integrating two leaf node evaluators, both developed on the same
Tiger base code, might take from under an hour to few hours or a day at most,
depending how disciplined your coding is. That is, if the leaf evaluator is the
only difference.

It doesn't fit what little we already know about. There must be another
ingredient that both, requires much more work than just cut, paste & call to
integrate with regular Tiger, and that makes the increased king safety score
work in comp-comp play.




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.