It's actually much higher than that, for a few reasons.
* It doesn't correct for students (12th graders) who were absent on the day of the survey, or for the 12-15% of students who drop out of high school before the fall of their senior year. By using clever methodology they are actually able to measure how the differences in drug use among students who were absent the day of the survey, but they don't correct for it. (They explain the differences in the methodology though.) And as for students who dropped out, Appendix A of volume 1 of the same report explains how they concluded that dropouts were probably 1.5x more likely to have used marijuana by 12th grade. So if you take into account that the 20% of dropouts and absentees have a lifetime prevalence closer to 88%, then that would increase the total lifetime prevalence of marijuana use by about 1%, so to 83%.
* About 1.8% of those who admit marijuana use recant each year. It's likely those who recant actually did use the drug, for reasons that are explained here: http://www.erowid.org/psychoactives/statistics/statistics_ar...
The survey partially corrects for this by counting those who recant after admitting to use on at least two previous surveys (multiple years apart) as having used marijuana. However, students weren't surveyed until 12th grade, so it stands to reason that if the 1.8% linear rate of recanting held until age 13 years 11 months, the average age of first marijuana use, then you'd have to add about 5.5%. Again this is somewhat speculative, but probably roughly accurate.
* It seriously undercounts those who later become heavily addicted to drugs. So anyone who didn't use drugs through 12th grade but then became a drug addict essentially wouldn't be counted at all.
* It doesn't count people who would have used marijuana at some point after 12th grade, but who died of non drug-related causes before they got a chance to. This is significant because unlike alcohol where virtually everyone who will ever drink has already started by age 18 or so, a large percentage of eventual marijuana users are still trying the drug for the first time up until their mid 30s.
The fact is that if you make it to age 50, your chances of having smoked weed are probably closer to 88-90%. The Erowid thing I linked to is also good in general for as a quick way of grokking the validity of the numbers.