I think what Roger described is the classical occupancy problem http://probabilityandstats.wordpress.com/2010/04/04/a-formula-for-the-occupancy-problem/
The structure breaks up into finite series components, which should provide an elegant example of J's capabilities. The problem will be with the factorials which may become very large. There are obviously a number of ways of rearranging the components of the terms to maintain better numerical properties. The choice of those transforms will likely have a big effect on the range over which the functions will give good results. Fraser ----- Original Message ----- From: "Roger Hui" <[email protected]> To: "Programming forum" <[email protected]> Sent: Friday, May 27, 2011 1:24 PM Subject: [Jprogramming] roll and deal > What is the least c such that the probability of ?c$n having > m distinct values is at least p? > > ref: http://www.jsoftware.com/jwiki/Essays/Birthday%20Problem > > > > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
