Author: Dann Corbit
Date: 13:28:10 07/04/01
Go up one level in this thread
On July 04, 2001 at 06:43:04, José de Jesús García Ruvalcaba wrote: >On July 03, 2001 at 19:12:01, Dann Corbit wrote: > >>On July 03, 2001 at 12:59:15, Dan Andersson wrote: >> >>>It has been successfully argued that common sense does not exist, until someone >>>needs to make a shaky argument. The computers performance level seems beyond >>>doubt i.e. 'hard data'. The only thing remaining is formulating hypotheses and >>>trying them against available information. There are many factors that are >>>possible but, more than a few are in the programs favour: mental fatigue, >>>stubborn defence, tactical combacks and very good opening preparation. >> >>I don't think there is any question that many computers have had GM level >>performances or that they are the tactical equal (often superior) of a GM. >> >>The question is of certainty and proof. >> >>I would say the following: >>It has been 'demonstrated' that computers are of GM caliber. >>It has not been mathematically 'proven' yet in a rigorous manner. >> > >That can not be proven mathematically. Math does not have any extramathematical >consequence. It can be demonstrated to any degree of confidence that is desired. This can be performed under carefully controlled conditions to insure fairness and accuracy. It can be repeated to demonstrate repeatability. In the mathematical sense, it can clearly not be proven, but if we want to be pedantic, Kurt Gödel showed that nothing can be proven without using axioms that are unproven to start with. And Heisenberg showed we can't rigorously know everything about the physical world either. But we can do better than we have. The only thing holding back an accurate determination is money. If the experiment were interesting enough, that would happen also. I do think we have enough evidence to say that the current ability may be GM level to within the confidence of a coin toss or so.
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.