I have been over the prediction track for some time, and I think it's time to move on into trapping...
In my many failed experiments with prediction, I have come to the conclusion that elimination is a big mistake, so rather than wheel 20 or 30 numbers, I would rather catch all of them a given number of times...
Here's what I am working on for Powerball so far...
the UberWheel...
Lines = 42 (this is the basis for the wheel, play all 42, or target 3 for elimination---riskier--- and play 39, but that remains to be seen)
Guarantee = one $3 winner... that's it... anything else is up to luck
This wheel will be for wednesday drawings only, and will be grouped as follows...
5 lines each containing ONE number from last wednesday's drawing.
5 lines each containing ONE number from two weeks ago.
5 lines each containing ONE number from FOUR weeks ago.
10 lines containing TWO numbers from last wednesday's drawing.
1 line containing THREE numbers from last weeks drawing and TWO numbers from 2 weeks ago.
21 lines containing ZERO numbers from the last 4 wednesday drawings (wheeling all of the remaining 33 to 48 non-repeats).
Total is 42 lines, powerballs played from 1 to 42...
now this is still in the planning stages, but the dvents I am tracking (repeats from 1 to 4 weeks ago) are significant enough to warrant some kind of attention, I am hoping to catch either the occurrence or NON-occurrence of these dvents. Distribution will be based on an 80% to 90% rule based on previous draw history (the numbers in each position had to hit in 80 to 90 percent of the entire history).
I am still checking the data to see if any of the sets can be modified or combined, but how does the idea sound so far?