Author: Uri Blass
Date: 22:05:08 05/04/01
Go up one level in this thread
On May 04, 2001 at 23:45:26, Pham Minh Tri wrote: >On May 04, 2001 at 05:04:23, Uri Blass wrote: > >>On May 04, 2001 at 01:46:06, Pham Minh Tri wrote: >> >>>Hi friends, >>> >>>Some recent posts raise some interesting questions to me: What is chess the >>>problem and has it been solved or not? Many people believe that we can not solve >>>the chess problem because of exponential cost and can only approximate the >>>solutions, which is good enough. However, we have not been given clear and >>>unique definitions of the chess problem and of what is a solution. If you give >>>me one, the answer would probably be different. I think, with almost all >>>definitions, we can be happy to say, the chess problem was solved (or been >>>solved in this year). For example: >>> >>>Some old books (actually, I did not read about their definitions but could infer >>>from them) define a chess problem in that with any given positions of chess, we >>>would know the results and how to defeat (or draw) the opponent. I am sure those >>>problems were solved with some evidence: >>> >>>1) Opening positions: To me, any good chess programs of people in this club >>>could beat me easily all games, even if I play the white or black side. >> >>It proves only that chess proghrams are better than you. >>It does not prove that chess is solved. >> >>Suppose all the humans in the world have rating of at most 1400. >>somebody with rating of 2000 comes from the moon to visit them and win against >>all of them. >> >>Did (s)he solve chess? >> >>It is clear that (s)he did not solve chess but the humans may think that he >>solved chess if they use your logic. >> >> >>If you claim that chess is solved you need to prove that there is something that >>cannot be beated. >>If the best computer can be beated by better hardware that can be used in 2010 >>then it is a proof that chess was not solved. >> > >I think you are not clear here. If a problem is solved by a computer, it meas >the solution was found, even it may be not the best one. The better computer in >2010 will find the __better__ solution of a solved problem (look like they >compute faster any fomulars). The question is what is the definition of a solution to chess. I do not define it as beating all the humans but as finding a strategy to get at least a draw against everyone(or a win with white against everyone). There are some simple games that are solved by my definition. It is not interesting to see comp-comp games in these games because all the games will give the same result unless the hardware or the software is not good enough. Fortunately it is not the case with chess and I expect that computer tournaments will continue to be interesting at least in the next few years. Uri
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.