Most of them, actually, do find that the state's payout (50% of the "fair" value, in typical 3-digit or 4-digit daily-lotto games) is so many standard deviations less than the expectation from just saving the money that it is "off the chart", less than .01% chance, which is as far as they need to get (I then give them scary estimates about how outrageous the denominator is, in the real probability of being, say, 8 standard deviations from the mean).