Author: Joshua Lee
Date: 20:52:53 09/11/01
Go up one level in this thread
On September 11, 2001 at 19:15:00, Dann Corbit wrote: >On September 11, 2001 at 18:58:15, Joshua Lee wrote: > >>I noticed that most WCSAC positions are not solved with Fritz or any other >>program due to incorrect solution has anyone kept a lit for revision to the test >>set or others? > >There was some work on ECM. > >The only throughly debugged, large test set (more than 20 problems or so) is >WAC. The other large sets are full of bugs. ECM is probably the worst one. > >There are also repeat positions (e.g. some position in WAC might also reside in >WCSAC). I will back you up on that one , You can take all the epd files from your site (not including mega the older version of this file had all the positions from the other files)put them together you'll have over 10,000 positions with (ROCKPIBM being nearly 2000 extra just without specifics i.e what test set what position came from) Then you have Many ECE and MES positions being duplicates . I deleted the doubles and then the positions where two programs solved quickly alot of the work is done however i still have around 866 not solved in 15 minutes by Fritz 6 , another 150 in one file that Gambit Tiger Solved that i want to test with fritz which i am doing with my other computer and under 1300 a good 600+ are just MES/ECE and Shereshevsky positions. Of the several thousand that were previously solved within 3 minutes there are only a few that i need to test with Gambit Tiger (some of which Fritz solved within 15minutes but GT2 did not.) So i think it is very safe to say there are hundreds with either incorrect solutions and or are too difficult for Top programs. I will let you know how it turns out the problem is just it is so time consuming .
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.