Author: Fernando Villegas
Date: 16:36:15 05/14/99
Hi all: Suppoose you are playing a game and you arrive to a decisive position. Suppose you have only two options, move A and move B. And each of them gives ten posible counter moves to your opponent. OK? Now suppose that of the ten moves the opponent can do after you move "A", the very best of them gives him an score of, say, -0,21. The rest are even less good for him. And suppose that of the ten moves the opponen has for your move "B", 9 are awfully bad for him, a lot worst than -0,21, but he has one move that give him, say, +1,1. And now the last suposition: your adversary has five seconds left in his clock to complete his move. Now tell me which is the "best" move to do, A or B? You see, in this extreme example I think the problem of what is "best" appear clearly. Clearly is not so clear. I bet than any human player would play B on the ground that with such few time at his disposal, there are very good chances the adversary will commit a mistake. In other words, practical considerations change radically what can be considered as best. If you see a chess game like a fight and NOT like an abstract problem, where a theoretical move is the very best, then you have ground -I believe, correct me if I am wrong- to think that programs like CSTAL could be performing a somewhat different approach and to understand why they are good againts human and not so good againts computers. Maybe the same apply for the anti-human feature of Rebel. I wonder if some programmer here has implemented something of the sort to define in other, practical way what is best in a real game. Fernando
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.