Author: Dirk Frickenschmidt
Date: 11:21:11 11/21/97
Go up one level in this thread
Hi Mike, you wrote: >Thanks for the feedback guys. But I still have questions. IS a test like >this truly ojective? Are they equal positions which eliminate poor or >excellent opening books? Any of this kind of tests should be played with opening book *off*. Else it makes no sense at all. The aim of thematic playing tests (which I used as student and CSS hobby journalist a long time ago) is to get some insight into a programs playing strenth as well as playing style. This kind of test is as "objective" as any :-) insofar as you have to know what you are looking for. A lot depends on searching for appropriate positions: do they have some exemplary value and are not just accidential samples. If you succeed in getting the appropriate positions I think in a 20 positions test (40 games) you might be able to learn more and faster about new porgrams than with any other testing method - knowing enough about the testing positions (pawn structure, various possible key moves and plans, skill in developing advantages and leading those to won attacks or endgames etc). It would be like one big fingerprint-test for programs. I have just begun developing such a test of my own and would like to invite the whole group in discussing the positions I will post here soon. And I would be glad if those who have two computers with equal or very similar hardware could have these play some (autoplayer) games with these test positions. Who would be ready to do so if time allows it? More soon. Kind regards from Dirk
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.