Author: Mark Young
Date: 18:38:19 05/18/99
Go up one level in this thread
On May 18, 1999 at 20:16:35, Dann Corbit wrote: >On May 18, 1999 at 19:55:02, Mark Young wrote: >[snip] >>Incorrect logic and reasoning...Humans find moves this was all the time...its >>called positional play. >Of coures. And when they do find it, they know it is a good move. They would >not say, "Now I am down 1.5 pawns to my opponent." Are you a chess player, this statement is total ********. They don't know its a good move or winning move all the time...but they THINK its the best move (BM) in the position. The eval is meaningless, you can have a eval that shows anything. What is importent is if it plays the correct move and correct line. If it finds the correct move and line because it see all the other moves and lines are worse, or because it sees that 1 move and line wins, is no difference. Both methods are valid, and both are use by computers and humans. Think about what BM stands for, Best Move... Both methods can find the correct move and line of play, and have equal value in the real world and in a over the board game. > >In the same way, a chess program that thinks it is behind has not found the >right move yet. > >If you don't put down a bm for H7 to look for, it does an excellent job. If you >do put one down it does some bad things that make for crappy data. That was my >point, which you have failed to grasp. > >When you put down a bm and H7 immediately stops, it is not because of some >positional brilliancy. It is because it happened, at that iteration, to look >better than the other moves. Yet if the eval is still negative, it does not >know why the move was chosen. It just stopped. I will write slow so you understand. I don't care what or what not hiarcs7 finds because a BM input, or if it does something strange with the BM. I did not search the position with the BM. If Hiarcs7 were playing a real game, and got to those positions it would have played the BM and correct line.
This page took 0.01 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.