Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Usage of general.ctg book+CB learner by ChessMaster in SSDF testing

Author: Ed Schröder

Date: 08:21:32 10/31/03

Go up one level in this thread


On October 31, 2003 at 06:02:33, Luis Smith wrote:

>On October 31, 2003 at 05:45:38, Ed Schröder wrote:
>
>>On October 31, 2003 at 04:27:54, Luis Smith wrote:
>>
>>>>I have to agree with your criticism. I was delighted to see the SSDF initiative,
>>>>but this is the wrong way of doing things and it does more harm than good. It's
>>>>better not to test this wonderful chess program than doing it this way.
>>>
>>>I disagree.
>>>
>>>What do you mean by "It's better not to test this wonderful chess program than
>>>doing it this way."?  Do you mean because of the book?  I saw no one crying
>>>"unfair" when Crafty used general.ctg.  Besides the book must be 2-3 years old
>>>by now, and I am sure everyone has already optimized against it.  Perhaps Johan
>>>can grab a few PGN's and let SSDF make it into a book?  Would that satisfy all
>>>that critics?

>>The SSDF policy is out-of-the-box testing, everything else is wrong. Own book,
>>own learner stuff. This only represents the true strength of a program.

>So does out "out-of-the-box" testing include Shredder 7.04?

This is okay. Keyword: public availability.

Also the software should be the total work of the programmer or programmer team
(company). It can't be the case 5-10-20% of its strength (opening book +
learning) comes from another source (programmer, company). That's not right.



>>>>Suppose CM9000 will top the list in the next edition then I foresee the longest
>>>>thread in the history of CCC. Suppose it does real bad it will cause a similar
>>>>stir.
>>
>>>I disagree here also, when they announced the testing of it there wasn't as much
>>>of a stir as I thought there would be.  There might be a few "Yays!" or "Johan
>>>is the greatest programmer ever" but I don't think it will cause much of a
>>>rutkus.

>>Well, tell me if you want to put your money on it and how much :)
>
>haha sorry Ed you're asking the wrong person, I'm broke =O)

Good excuse :)



>>>>Unfortunately good intentions not always guarantee wise decisions, perhaps we
>>>>have pressured them too much.
>>
>>
>>>Perhaps so, at one time I thought SSDF was an organization who only had
>>>Chessbase products top the list.  Now seeing that they are testing TheKing 3.23
>>>especially with SKR settings they have shown they have different intentions.
>>
>>Agreed, that's the good part, it shutdowns the conspiracy theorists.
>
>Yep, after a while I was believing it myself.
>
>
>>
>>My best,
>>
>>Ed

>How come the SSDF hasn't tested other versions of Rebel (besides Century 4)?  Is
>it because of the past problems that you've had with them?  Is the new Rebel
>even capable of auto-testing?

I did not ask.

The new Rebel 12 will play.

My best,

Ed



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.