At $5 a ticket and a nightly draw, that is prohibitively expensive to "experiment" with. Does not mean that I would not, just means I would not regularly. The very first 1 week test concluded with a $6 profit... so now is the time to step away with a rare return on a poorly designed spreadsheet analysis of 2 weeks worth of data.

Looming at those 2 weeks eventually yielded one combo to play, so I dropped $35 for a 7 day stretch. Second night was a 2+1 for $25. Sixth day was a 1+1 for $8 and the last day was a 2+0 for $8. $35 in, $41 out, net + $6. Of the 5 white balls, the wins involved 4 of them, one pick was never drawn. The 1/5 has a 20% chance of coming out, so that factored in twice.

With that, I do not think that the "system" I employed to get that 1 line was anything special... first, third, fourth and 5th nights matched 0+0. Though it would be interesting to run a "control" experiment using a generic Python QP, I am cheap. If the mood ever strikes again, I will run exactly that experiment. The one thing that seems to hold with the jackpot games is playing for a cycle vs a single draw, always seem to do better playing for a week than per draw or long term... maybe that is the magic time interval... play the PB the same way, pick 1 combo and play for a week (3 draws). 

That was the first time I fired up a spreadsheet in about 4 months... that used to be a daily ritual for draw updates. The draw update script turned that into a single button push whenever I need the data, so roughly once per week for the latest PB draws... I automated the time sink into the hobby, amazing how quickly it was replaced as the hobby of choice.