Author: Bruce Moreland
Date: 14:42:59 06/22/98
Go up one level in this thread
On June 22, 1998 at 17:04:56, Don Dailey wrote: >I still always run test suites first too, and for an algorithm like >this is especially useful. I agree they can often tell you with >reasonable certainty that an idea DOES NOT work. Also test sets >are great at finding bugs. Yes. If my program does great on a test suite, maybe it means something. If I run a tactical suite, and a new version scores really really badly on it, I probably didn't make a tactical improvement. Likewise, if I make a performance improvement, and I end up reaching less depth in the same amount of time, the version is probably bogus. bruce
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.