I've deliberately discarded the recurring numbers and increased the sampling on the two high number sets:

1st Moon Phase     
 

Feb 16, 2005  Florida Lotto  6 7 8 35 48 52
 Dec 18, 2002  Illinois Lotto  2 8 35 40 48 52
 Jun 19, 2002  Florida Lotto  4 10 14 33 48 52
 Dec 6, 2000  Florida Lotto  3 10 25 37 48 52
 May 17, 2000  Florida Lotto 9 12 40 42 48 52
 Dec 13, 1997  Illinois Lotto  7 14 25 48 52 
 Mar 30, 1996  Illinois Lotto  44 48 51 52 54 
 Mar 15, 1995  Illinois Lotto  25 27 38 48 52 
       
11/12/2002  3 28 45 48 50 24
 Sep 23, 2000  11 13 30 39 48 50
11/20/2004  4 29 45 48 50 1

33 unique numbers and the frequency of hits of each within this sampling:

1-1 10-2 24-1 30-1 40-2 50-3
2-1 11-1 25-3 33-1 42-1 51-1
3-2 12-1 27-1 35-2 44-1 52-8
4-2 13-1 28-1 37-1 45-2 54-1
6-1 14-2 29-1 39-1 48-10 
7-1         
8-2

Yesterday, with a smaller sampling of the high combos, but roughly the same number of unique numbers the result was 4-1

If there's any validity to this specific approach this sampling of 33 should have them all on tonight's PB draw.

Jack