Author: Fernando Villegas
Date: 06:23:40 10/23/98
Go up one level in this thread
On October 22, 1998 at 20:37:12, Dann Corbit wrote: >On October 22, 1998 at 16:44:34, Fernando Villegas wrote: >[snip] >>Hi Dan: >>In the case of an entity that is designed to fight another, probably that sheer >>capacity to change should be considered as part of his strenght. So we should >>not pretend to get a valid result only if that result is static, permanent, >>something that another searcher can repeat or falsify as it is said in the >>canon. We should say "this program actually is 2400 Ello, BUT have an >>unmensurable capacity to grow so probably will be 2400+ at short notice". >We won't know how well it gains until we measure it. And we won't know how >linear the gain is until we measure that. By the time all of our measurements >are done, there may be something much better that comes along anyway. > >>Equaly, perhaps we should give more room to the concept of prediction. Clearly >>you are a scientist and so I am not going to teach you nothing, but sometimes it >>is forgotten that prediction not necesarily are discrete. >I doubt that you will not teach me anything. I constantly learn from this >newsgroup. It is, I am sure, as Will Rogers said: "Everybody is ignorant -- >only in different areas." > >>In fact, there are >>non discrete predictions. With just to refine the tool, the measure change. The >>world also is chasnging all the time and so the mechanistic glance of the past, >>as if the world looked at by science is some kind of watch, is erroneous. >>Another thing: is not a prediction absolutely grounded in, let us say, a >>trillion events, something that lose all relevance? What I mean is that to >>predict is an act guided by practical, current problems that must be faced in a >>world of uncertainty wirth uncomplte facts and so maybe it is in his essence to >>be partial, uncomplete, tentative, something you do with just ten games. What I >>want to say is that if you have the trillion games with a precise result, then >>you do not need prediction at all; you just have a current knowledge. You"Know" >>the game will be won by this program with 99,99999% probabilities and that is >>not a prediction, but an statement. Of course all this is more a semantic >>discussion that a scientific one :-) >Indeed. The quantity of our measurements increases the certainty of the result. > But the cost of measurement might exceed the value of the result if we are not >careful! At some point, it is surely time to stop measuring some object or >process and move to the next. ...and the same I suppose we must do with this issue. It has been a pleasure to debate this point with you, Dan, the kind of thing that makes worthy to be part of this group. In fact, the real value of CCC is for me less to talk about chess computer than to meet smart people with which to talk about anything, eveen including chess computers... Sheers 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.