Computer Chess Club Archives


Search

Terms

Messages

Subject: Kramnik will have a field day with Deep Fritz

Author: Joshua Lee

Date: 16:59:10 11/16/01


The match will wake a lot of people up look at Junior's game against Kramnik has
anyone attempted to see exactly how long it takes Junior to make moves that
would've kept it from loosing?

I really think that even on a 320Ghz machine any chess program will still be
able to be beaten but the correct positions have to be found from games that
test positional play. I have a decent test taking all main epd test sets and
solving a great deal of these problems originally about 10-15,000 including
duplicates  now removed and i was left with around 1700 which i previosly
posted. I am now testing instead of 15minutes per position 1Hr so given a few
months i will have cut this hopefully in half.
I had looked 15minutes as this would be Tournament level on a 4Ghz but after
looking at Nolot positions i noticed that most are solved within 90minutes some
take days (3,6,9) #10 and 5 which are harder are solved within 18 hours on a
1Ghz machine. so i figured the problems not solved within 1hr are Worthy of
being included in a revision of a new Nolot Like Test Set.  Provided after this
71+ days i can isolate the problems with incorrect solution or position.

Already there are a few that are not solved in under an hour by Gambit Tiger and
one of which that takes 8+ hours. Still more programs will have to be tested.



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.