Author: Ratko V Tomic
Date: 06:42:44 08/16/99
Go up one level in this thread
> In 10 years, it's clear that no human will be able to contend > with commercial programs running on off-the-shelf hardware. That's not clear at all. In checkers, where it is possible to search 20+ plies the search had reached the plateau of diminishing returns and the top human player (Tinsley, age 65) beat the top program (Chinook) in a regular match. Thus it is quite plausible (and there are already data supporting the plateau hypothesis) that in a few more plies, searching deeper will not produce any better move in vast majority of positions. Whether the plateau will be above the top human level at that time is questionable. Even if it turns out so initially, I wouldn't doubt that the top players would soon find a strategy to beat the machine, since having reached the plateau of its algorithms the machine cannot easily improve further. To get beyond its plateau, the program would have to search 100+ plies deep (i.e. to the real end nodes of the chess tree), and that is not anywhere within a forseable technological advance. Something significant beyond the brute force techniques will have to be created to move beyond the plateau.
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.