Author: blass uri
Date: 02:41:42 05/22/99
Go up one level in this thread
On May 22, 1999 at 03:02:06, greg moller wrote: >On May 21, 1999 at 14:21:28, KarinsDad wrote: > >>On May 21, 1999 at 12:55:13, Dann Corbit wrote: >> >>[snip] >>>In any case, it still puts a lid on the maximum number possible. If, for >>>instance, I can encode all chess board positions in 100 bits, then there are >>>only about 10^30 possible board positions. The legal positions will be a >>>subset of that. Hence, finding a minimal coding for board positions also has a >>>fascinating mathematical result: It puts a cap on the maximum number of possible >>>board positions. >> >>My estimate is 10^48 since I think that it can be done in 160 bits, but not much >>lower. >> >>KarinsDad :) > >I wonder, would it be possible to generate a (random) sample of random >positions, and based on the percentage of illegal positions in that sample to >infer what the actual probability of legal/illegal would be for all positions. >Any thoughts? > >regards, >gm I think that it is possible but you have to do a program that generate random positions and count them. if I suppose x1 white pawns,x2 white knights,x3 white bishops... then I have conditions like x1+x2<=10,x1+x2+x3<=12 for every possible case of x1,x2,x3,... the program can compute the number of possible random positions and add all these numbers. we have not too many numbers to add(less than 100,000,000) and the only problem is that these numbers are big and the program should know to add integers like 10^30 and I do not know a language that let me do it without overflow errors(of course it is possible to do an array for numbers and teach the computer to add and multiply big numbers but I hope that we do not need it). 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.