Visualize a large cubic room, say 100 feet on a side. 1,000 ping pong balls will be removed from an urn and placed in the room via random drawings from a uniform distribution for the x, y, and z position. The balls will be held stationary in their assigned position by means of magic. Once all the balls are in place compute the distance between each ball and its nearest neighbor. Take the mean of all those 1,000 computations. *That* is what I want to be able to derive analytically. That number may be called mean free distance, it may not. But the proper name to affix to the number is not what I am in search of. I can write a Monte Carlo simulation but an analytic solution seems possible.
If the balls are small enough, their size is immaterial, the only rule is that the only one ball can occupy one space. Assume infinitesimally small balls if you wish. If the room is big enough boundary conditions at the walls can safely be ignored. This was triggered by a statement I read that, on average, if galaxies were represented by a tennis ball, the nearest neighbor to a ball would be 50 feet away. I have tried to distill this down to get to what I see as the essence of the problem. I have no interest in pursuing the *actual* method used or the estimates *actually* used in formulating that statement. I am for now only interested in the mathematics of the sterilized situation. ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at http://jse.stat.ncsu.edu/ =================================================================