If the jackpot was won and the winning ticket was a QP, Thrifty's 10 QP chances of winning are compared to that total. Even if 100% of all the 20 million tickets sold in the last drawing were QPs, there was only a 11.4% chance of a jackpot winning ticket. When the jackpots get to the "must win" level, I'm assuming if the self-pick players buy extra tickets, they will probably buy QPs. Some people play favorite numbers in every drawing regardless of the size of the jackpot and the percentage of those tickets to the total is probably much higher in the start-up jackpots.
QPs probably do win more than 75% of all jackpots, but since not all drawings produce a jackpot, the overall chances of any QP winning a jackpot is still 175 million to 1. Over time when the total of all the QPs sold is added up, the average should be close to 1 out of every 175 million QPs sold won a jackpot.
Since Thrifty decided to purchase 10 QPs in one drawing as opposed to 1 QP in 10 drawings, his chances of winning are compared to the number of QPs sold for that drawing and the chances of a QP winning that drawing are based on how many QPs are sold.
IMO the overall percentage of QP purchases is probably much higher than 85%, but we know for a fact QPs don't produce a jackpot winning ticket in every drawing, so the chances of any self-pick line different than any of the QP lines in those drawings is compared to the number of lines not sold. If 20 million QPs were sold in the last drawing and if Thrifty selected ten lines and one of his lines was different than any of the 20 million QP lines, his odds were slightly reduced to 1 in 155 million because none of the QPs matched the winning numbers.
My point to Thrifty was if he is going to play someday and purchase QPs, it should be for a drawing that is more likely to produce a jackpot winner. If he is going to play once every ten drawings and in that drawing, based on sales projections, that drawing is unlikely to produce a jackpot winner, playing self-picks might be the better option. While it looks good on paper to say we have a chance of winning a jackpot 104 times a year, we're ignoring the obvious; not all drawing produce a jackpot winner so in reality we don't have 104 chances a year.
The logic of playing QPs "someday" makes sense because since PB went to $2 a ticket, there were only 10 drawings with about a 50% chance of having a winner based on the sales. Your guess is as good as mine whether is makes more sense having 104 chances a year compared to having 104 chances in one drawing that is more likely to produce a jackpot because either way, the expense is still $208 a year.