Author: Dann Corbit
Date: 13:59:26 07/31/01
Go up one level in this thread
On July 31, 2001 at 16:43:05, John Dahlem wrote: >>Strangely, if you play enough games, you will eventually play a perfect game >>that would beat Kasparov. > >Is there a way to calculate the chances of such a thing? For every move it must >play the one considered "right" over all other available moves. Chances of that >are what? 1 in a billion, 1 in a trillion? If it were that simple, it would be easy to calculate. Unfortunately, there can be more than one right move. Consider the first move of a chess game. A player might move 1. b4 if they wanted to play the Orangutan. Is that the wrong move? We can do a rough guestimate. There are 30-40 moves possible on average. Usually (if you look at what GM's and good chess programs do on average) there are often 3-4 good choices from an average position. So the odds of picking a good move might be roughly .1. The odds of making n good moves in a row are therefore (very roughly): (1/10)^n To play 30 good moves in a row would therefore be about 1e-30 in probability. Like I said, I would not hold my breath waiting for the thing to beat Kasparov.
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.