Author: Uri Blass
Date: 01:16:55 05/11/01
Go up one level in this thread
On May 09, 2001 at 19:24:36, Dann Corbit wrote: >It does not matter whether we use the 50-move rule or not. 5900 moves is quite >sufficient to render the problem 100% absolutely infeasible. > >Chess is computationally intractible. With current algorithms, it will never be >solved. No one will ever make a 32 piece tablebase and if they could, they >would have no where to put it. It is possible that with current algorithm chess will be practically solved and when I say practically I mean that it is possible that in 2050 every comp-comp game between the top programs will be drawn. It is possible that you can always find the best move by searching deep when deep is something that is not practical today but is going to be practical in 2050 when computers will be 10,000,000,000 times faster than today. You can use null move and avoid null move errors if you search deep enough by some simple tricks and you can avoid hash collisions by storing all the board in the hash table. In this case the problem of solving chess is relatively easy to the problem of proving that you solved it. Today it is clear that programs did not solve chess because finding always the best move mean that games always end with the same result and it is not the case. It is clear to me that in the near future programs are not going to solve chess but I cannot know about the results in 2050. 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.