by Jerzy » Sat Jan 26, 2008 8:14 am
Joe,
Distinct of History Differences in Columns filter is based on a very simple concept of counting the rows in the History Differences, which contain the winning numbers for the next draw, but it brings about very impressive results.
The filter counts the HD rows which contain the next draw winning numbers. It does not count the winning numbers. It does not matter how many winning numbers are there in a row, and it does not matter which number in which row is present. The same count of rows can result from several different configurations: eg, 4 rows with winning numbers: 1 + 1 +2 + 2 or 1 +1 +1 + 3.
The HD Rows with Winning Numbers would be probably a better name for this filter.
Someone asked you to create the possibility of entering a range of values into the filter. It is not necessary. This possibility already exists. Only 6 values are possible in 6/xx game and 5 in 5/xx game. And these values ( or ranges of values) can be entered easily by selecting the appropriate checkboxes.
The count of HD rows which contain the winning numbers in my 6/45 game looks similar to the data from your game:
12/01/08: 6, 5, 6, 4, 6, 4, 5, 4, 6, 5, 5 leaves 30 tickets for19/01/08
05/01/08: 5, 5, 6, 5, 6, 5, 4, 6, 6, 6, 4 leaves 158 tickets for 12/01/08
29/12/07: 6, 5, 3, 6, 6, 6, 5, 5, 5, 5, 5 leaves 10 tickets for 05/01/08
22/12/07: 5, 5, 4, 5, 5, 6, 6, 5, 5, 5, 6 leaves 81 tickets for 29/12/07
15/12/07: 5, 5, 4, 5, 5, 6, 6, 5, 5, 5, 6 leaves 294 tickets for 22/12/07
In each case the jackpot ticket is preserved. In most cases also some lower prize tickets are present.
The values: 4, 5 and 6 occur almost always. After a draw the values in the columns remain the same or change only +/- 1. But if it was 6 it can change only to the "-", and if it was 4 it will tend to change towards a higher value as the 3s are uncommon.
Occurrences of values in the columns are simillar to each other:
Values from 1 to 2 - the occurrences of fractions of the percent only.
3 approximally 2.6%
4 approximally 17.5%
5 approximally 42.5%
6 approximally 37.0 %
Two basic strategies are possible:
1. Attempting to find the exact counts of rows in the columns
2. Using the ranges of values or combinations of values to achieve a smaller but a safer reduction of the package and then using additional filters.
The statistics would prevent mistakes which are quite likely when we prepare the data "manually". But I do not know what statistical information would increase the success rate with this filter. The count of rows which contain the last draw winning numbers is the most important information we need. The next draw count most likely would look similar (but not the same).
Thank you.
Jerzy