Author: Christophe Theron
Date: 09:26:02 10/31/03
Go up one level in this thread
On October 31, 2003 at 06:30:42, Mike S. wrote:
>On October 31, 2003 at 04:40:16, Luis Smith wrote:
>
>>(...)
>
>>Remind me once again of how its "unfair" for TheKing to be used under the UCI
>>protocol which weakens chess engines under the Chessbase GUI, along with a 3
>>year old book?
>>
>>Besides what if they tested the original AND the SKR settings? Hell they tested
>>Shredder 7 (out of the box) and 7.04 (not out of the box) while 7.04 being
>>clearly better.
>
>Good arguments! (Although I don't think that King is weakened by the Wb2Uci
>setup; i.e. the SKR settings were done using this adaption method too.)
>
>Of course "out of the box" is a good general guideline, but also, the testing
>situation shouldn't be too artificial when the real usage in the computerchess
>community is much different to the first release version and it's defaults.
>Shredder 7.04 was such a big improvement, it would not have made sense to limit
>testing to an old engine version. SSDF has shown wisdom in both cases :-)
>
>I don't believe that somebody really thinks that the old general.ctg is much of
>an advantage or a threat for the other top progs. For example, AFAIK Nimzo 8
>used a newer book by A.Kure than general.ctg is (which came with Fritz 6). Fritz
>6 itself with that book is ranking 47 Elo behind Tiger 15 (K6-2/450).
>
>A book is also necessary in SSDF tests, to ensure that an engine doesn't have a
>*time disadvantage*. This may even be more important than any sophisticated book
>details which may decide one game out of onehundred.
>
>Regards,
>M.Scheidl
Testing with settings recommended by the author and patches that are available
for free to the customers is perfectly fine.
Testing a MIX of products is definitely not fair.
Christophe
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.