Author: Robert Hyatt
Date: 20:17:02 06/14/98
Go up one level in this thread
On June 14, 1998 at 17:30:02, Fernando Villegas wrote: >Hi again mark: >Just another and I hope last shot...for today :-). >When you say structural approach is not possible, I think maybe you dont >understand what I mean by such thing. I dont say there are some laws to >discover. What I say is that from a superior level of analysis, that is, >from positioonal structure instead of moves, we have or could have a >better point of departure for calculate or guessing. In fact, dear Mark, >save in specific cases where the game ends -let us say, a forced mate- >even the nmost conclusive and concrete tactical calculation is a kind of >a guess AS MUCH AS it is, nevertheless, a non exhaustive calculation. >Why a computer goes wrong after, let us say, capturing the typical >poisoned B pawn? Simple: because inside the horizon of his tree to win a >pawn is something concrete, solid, and then he supposes or guess that >beyond that, in the mist of future, that will be good enough. Tactical >and positional considerations has both a lot of guessing as much the >operation of calculation cannot be finished. If it can be, then we have >no problem at all and all this discussion is irrelevant. All this debate >is based precisley on the ground of the non resolvable aspect of chess >most of the time. The issue, then, is not between concrete calculations >and dumb guessing, but between one or another kind of guessing. of >course, you can fail with any of them. I am not talking of an unfalible >methoid of playing chess, I am talking of a most eficinet mechanism to >do gueses, to do sensible decisions when guesses and decisions are the >issue because not definitive output is at sight. >Regards >Fernando you might be right, but I'd bet that a computer taking the "b pawn" results in a win for the computer *far* more times than it results in a loss. But take this up to a piece, and it gets much better, because winning a piece almost always wins the game. Occasionally no, but 99% of the time, yes. And today, we are beginning to see situations occur where old "positional ideas" simply fall apart. I'll bet in 20 years, that chess is going to be a bit different than it is today... because computers are going to begin to prove that a backward d pawn in the Sicilian is not usually bad, or an isolated d-pawn in qga openings is not always bad, or always good. So experience is going to show that some old ideas were wrong, some need to be modified, some are unreliable, and some are even better than originally thought. Because tactics will continue to improve over time, and tactics is what chess is all about. You can *not* win games, if all you do is lose material... :)
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.