That math assumes a flat distribution of popping times, which I suspect is incorrect.

Listening to a bag of microwave popcorn, it starts off slow, gets more rapid, and then tapers off again, implying that kernels are more likely to pop near the average time, which makes it somewhat more likely for two kernels to pop simultaneously.

But yeah, whole bag at once is probably still basically zero. Unless you use one of these, of course.