Author: Graham Laight
Date: 04:23:16 01/08/03
Go up one level in this thread
On January 08, 2003 at 06:42:38, Drexel,Michael wrote: >On January 07, 2003 at 23:00:40, Russell Reagan wrote: > >>On January 07, 2003 at 07:56:28, Drexel,Michael wrote: >> >>>Nonsense. computer chess will improve slower and slower. >> >>>again you are wrong. computer chess will never demonstrate this. >> >>>no computer will EVER be able to prove that. not in 10000 years. a prove would >>>require a 32-piece tablebase. hope you know what that means. >> >>That is quite a bold (and pessimistic) statement. Unfortunately, it's also >>wrong. It is quite foolish to predict what will be possible in 10,000 years IMO. > >Sorry, I was wrong. Chess will be solved in year 2168 :-) > >http://user.cs.tu-berlin.de/~kunegis/hack/chess/ Wrong calcualtions. They say in that link that in 2002 we can investigate 3 million positions per minute. If one is only checking for checkmate or draw in the eval, then I'm sure that in 2002 we could have done several billion positions per second on optimal hardware. My prediction: after 2020, no top chess computer will ever lose a game. See my previous posts in this thread for the reasoning (which is basically that one can extrapolate that all games will be drawn above 3500 elo, and I posted a link to a well-researched graph from which this level is extrapolated. In picking 2020, I'm assuming that computers advance 40 elo per year - a figure I plucked out of the air - but which is of the right order of magnitude). -g >do you really know what 10^43 means? I dont think so.
This page took 0.01 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.