Author: Robert Hyatt
Date: 15:19:08 10/14/99
Go up one level in this thread
On October 13, 1999 at 17:24:57, Georg v. Zimmermann wrote: >Strange, because when your eval drops to -1.5 it is probably too late already >cause u missed something _before_. This means: try to use more time in positions >where your program makes wrong moves ;-). > >--Tec. Right... although on occasion your position is fine. You have just chosen a move at the root that has a tactical refutation. Recognizing this and using more time to find one of the acceptable moves is good. One of the things that Hsu had done that caught my eye on several occasions was that they could 'smell' these positions better than Cray Blitz. IE they sensed when something was wrong, because somehow the size of the tree changes in a way they could measure, and they were able to recognize that some sort of 'horizon effect' was involved and they could go into a time panic early enough to really solve the problem. IE once you play a move, and start to ponder, and fail low on _all_ possible moves, you are 'done'. You have to hope that you are able to avoid those kinds of 'singular' lines where you have only one good move, because if that good move isn't good enough, you are in deep kimshee. They recognized this quite well in the cases where I was watching them. But I never quite understood their somewhat cryptic explanation when someone would ask... as I was always busy in a game of my own and could only listen in with half of one ear... :)
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.