Hi,
I found these videos on the Internet which may be of some interest.
http://video.yahoo.com/watch/8018241/21273089
http://video.yahoo.com/watch/8075163/21368715
N.B. Darrel Dolph claims to be a retired Professor of Electrical Engineering
P.S. A better method to establish the maximum separation between the best
and worst numbers drawn, would be to use walk forward testing for say 50 draws
using varying numbers of previous draws say, from 100 to 400 until the optimum
number of previous draws was established to give the maximum percentage
separation between the best and worst numbers actually drawn.
daniel