Author: Gian-Carlo Pascutto
Date: 23:36:30 01/07/01
Go up one level in this thread
On January 07, 2001 at 18:30:04, James Swafford wrote: >> >>The catch here is that it keeps failing in the same direction most of the >>time, so one bound will often be enough. > >Why does it keep failing in the same direction most of the time? > >My understanding is that you're basically playing a game of "hi-lo." >In other words, take a guess. If it's too high, take a lower guess. >Too low, make a higher guess. Eventually you'll zero in on the >correct score. If it kept failing in the same direction, you'd >continue to make lower and lower (or higher and higher) guesses >until you hit the correct score. Doesn't sound right, but maybe >I'm misunderstanding you... It's exactly how the basic algorithm works though. I makes one guess, then keeps failing low or high till it reaches the correct score. This may sound ineffective, but in practise it works very good. >>If you use convergence-accelerating heuristics, dual bounds will probably >>be better 'cos you will be failing in both directions then. > >Maybe that's what I'm thinking about. (?) Yes. Many people try to accelerate the convergence by increasing the interval of the guesses. This makes it possible to overstep the real value, in which case you fill fail in the opposite direction. -- GCP
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.