Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: SSDF always fails to test the latest and greatest. Can Rybka be Tes

Author: Dann Corbit

Date: 10:47:32 01/09/06

Go up one level in this thread


On January 08, 2006 at 15:13:35, Stuart Cracraft wrote:

>On January 08, 2006 at 01:35:26, Uri Blass wrote:
>
>>On January 07, 2006 at 22:48:31, Joseph Anthony Merolle wrote:
>>
>>>First fruit now Rybka can we please make a last minute entry.  Thank you
>>>
>>>Joseph
>>
>>
>>No
>>
>>SSDF wait for rybka1.2 that is going to have an opening book and other
>>improvement.
>>
>>There is no reason for ssdf to test a version without opening book.
>>
>>Uri
>
>Seems to me testing without an opening book, all programs, would be more
>accurate comparison of chess strength.

If you like to see exactly the same moves played over and over.

>One could eliminate the opening issues by starting from a standard
>set of positions like the Nunn suite.

And eliminate the thousands of hours of work that the book builders have spent.

>Ken Thompson used to do this when testing Belle.

I bet if he had a book with 5000 hours of effort, he would not have done so.



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.