Various Lotto Programs I've Written

For anyone dropping by this website, I still haven't won "the big one" but in it's pursuit, I have now written a number of programs which perform a range of analysis functions I've trialled, including:

My previously presented Boxing Assessor tool

My Relations Assessor tool which I present in this post.

Next post I'll talk about the Probability Assessor tool I've written

This one (the Grouping Assessor) generates random game sets and assesses how many times in the history of the lotto drawings set, each particular set of numbers has been successful as a 4,5 or 6 number set.
This results in the following output:
1.5% of the 1564 random sets would have resulted in a 4# drawing
0.3% of the 1564 random sets would have resulted in a 5# drawing
0.1% of the 1564 random sets would have resulted in a 6# drawing




The number combinations which would have resulted in those drawings were:

 









From this I generate a Number Heat-map which helps me visualize which number combinations occur most frequently from different viewpoints from week-to-week
e.g.  #1 has occurred in combinations 6 times in the 4,5 & 6 number sets whereas #14 has not occurred in combination with other numbers even ONCE ... (out of  1564 random set generations in this run)
#1 has also been actually drawn 6 times and a further 3 times as a supp in the last 24 drawings (9 times out of 24)
On the other hand, #14 has been actually drawn only 3 times in the last 24 drawings but not once as a supp.
 

 
 This program also gives me much more accurate and selectable Bell-curve limits to use in my WinLotto game Generator

 

Saturday Gold Lotto Prediction 

for Sat 31 Aug 2013

What I think will be the most likely 14, 16, 18 or 20 of the 45 numbers 

to contain the majority of the winning 6+2 numbers for tonight's draw

based on statistical probability. 

Refining by recent frequency and boxing would render 14 to choose from

And the the Winning Numbers were: 

 

So the selection verdict was:

! 
 n of the 8 numbers using 20 of the 45 
The nn number-set netted ?

n of the 8 numbers using 14 of the 45 
The nn number-set netted ?

The most-to-least-frequent boxing based on frequency of appearance in the last 12 weeks
(to better view these wide arrays right-click and choose 'view image')




A snapshot of past winning distributions 
You can see from this that just by using only the top 32 of the 45, you mostly get 5 or more of the 8
I then attempt to apply several further reductions to get as few "likely" numbers to play, as possible.
I reduce based on statistical probability in each appearance position [1~6,s1,s2]
I then weight those likely numbers based on their overall appearance frequency across all positions
This usually gets me down to between 14 and 22 numbers
I then wheel those numbers and further reduce the resulting number sets by:
Game-sum bell curve value and;
Previous draw occurrence of either 4 or 5 number sets

BTW. The Game-sum Bell Curve this week should be between 104 and 275 
[and I get that correct 82% of the time]