Author: Dann Corbit
Date: 07:55:43 05/10/05
Go up one level in this thread
On May 10, 2005 at 10:50:33, martin fierz wrote: >On May 10, 2005 at 05:22:57, Steven Edwards wrote: > >>On May 10, 2005 at 05:07:12, martin fierz wrote >>>On May 09, 2005 at 17:43:17, Dann Corbit wrote: >>>>On May 09, 2005 at 17:24:36, martin fierz wrote: >>>>>On May 09, 2005 at 14:54:54, Steven Edwards wrote: >> >>>>>>WCSAC.0107: >>>>> >>>>>forget it - WCSAC is no good. use ECMGCP instead... >>>> >>>>It is still interesting to debug the test suite. >>>> >>>>After his efforts, we will have a better understanding of this particular set. >>>> >>>>I expect every test suite will always have bugs in it until we drive every >>>>position all the way forward to formal proof of the outcome. >>>> >>>>I think that WAC is probably the most carefully debugged, but I am also sure >>>>that there are bugs still in it. >>> >>>both WACnew and ECMGCP are good. WCSAC as a book is just full of errors. if you >>>want to spend your time debugging it, fine. if you want to spend your time >>>debugging your program, use WACnew (for weak programs only) or ECMGCP. >> >>A debugged WCSAC will be useful for many, including myself. Also, I feel a >>little bit of personal responsibility here: I would have extensively debugged >>WCSAC and several others had I had the resources years ago when I first made the >>EPD test suites available on the net. > >as i said, there are already debugged test suites available. one more won't >hurt, but it won't be really useful IMO. It depends on what you are looking for. If you want to be able to solve checkmates, then BWTC should be useful. If you want to be able to solve sacrifices, then WCSAC should be useful. If you want to be able to solve simple tactical problems, then WACNEW. If you want to be able to solve middle game problems, then ECMGCP. Most test suites are targeted to a particular purpose -- albeit often with overlap. Even if a million hours of CPU time should produce a new test suite that does not make any engines any stronger, it still produces something useful: Beautiful new solutions (or even overturnings) for famous chess positions. The outcome is art, in my eyes.
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.