The mid day passed the halfway point with no wins yet. Evening will do the same tonight, and also winless as of last night.

The observations from the back test point to an interesting concept... only play week 3... that would lead to a new set of combos each week.

There are two ways to do this...

1. Hold the history updates by 14 days. Seems a simple enough solution, but there is a better way...

2. Set the classify variable to 14. This would make the pick virtually 2 weeks ago AND classify the first 2 weeks of results. Why this would be better is that I do not have to modify the history file and can continue to update as normal. This solution would also be applicable to any of the other games... though I do not back test the jackpot games as rigorously as the pick N games as they only need to work once.

Why I might not... the live test is only halfway done. It is most likely to complete without any win, so maybe observations are subconsciously tinted by what I wanted to see...

So, ride the test out and see what happens in week 3... if there is a win it would mark 2 cycles in a row with a week 3 win (even if the first was paper play).

That makes the plan hinge entirely on week 3...

IF there is no win... don't change anything

IF there is a week 3 win... implement the above changes.

Next year, the pick 5 showed no such quirk, so that cycle will be business as usual.

I can't even believe that they will be taking away the Cash 4 Life in February! This changes the allocation of any hit on pick 3 to $260 towards funding next year and $240 for a shot at the C4L.

System wise, everything is finally synchronized... the method, the software, the budget... all that remains is to actually win... same as any other system or year...