by Jerzy » Fri Oct 26, 2007 12:56 pm
Hi Joe,
I tested your Number Positions Range in Columns filter.
I do not have any statistics that would be suitable for this filter, so as the first step I found the setttings which pass all past winning combinations. With those settings I filtered the full package of 8,145,060 combinations and removed approximately 20,000 of them. Not many combinations. To remove more of them, less gentle settings are needed.
Next, using the last draw result, I reset the data base, and found the exact location of winning numbers in rows of each column, e.g. for the column 0 in rows: 8 , 5, 4, 10, 2, 9 . I entered these into the filter, e.g. for column 0:
P1 from 8 to 8
P2 from 5 to 5
P3 from 4 to 4
P4 from 10 to 10
P5 from 2 to 2
P6 from 9 to 9
and then I filtered the full package (8,145,060).
Filtering the column 0 reduced the package to 16 tickets, after filtering the column -1 (with the setting 5, 4, 16, 5, 4, 11) 4 tickets only were left, and after using the filter for the column -2, only the jackpot ticket remained in the package.
This was done with the data base setting which was reflecting the situation before the last draw.
Then I reset the data base to the situation after the last draw to filter out the same row - position patterns from the full package to get the tickets for the next draw (assuming that these patterns would not be repeated).
I used the Back-test function for columns 0, -1, -2. There was 1 ticket only in the history which passed the filter. As you can guess, it was the last draw jackpot ticket.
I used the Preview function and no combination passed the filter. The last draw winning ticket was in the package, but its winning numbers had a different locations in the rows of columns after resetting the data base.
It was impossible to remove the last draw row - position pattern from the package that I was preparing for the next draw, because there was no such pattern.
The conclusion: It is a very sharp filter. Proper statistics are needed and more experimentation to find the best way of using this filter.
Thank you for your filter.
Good luck
Jerzy