Taken directly from another forum:
A complete, unsearched box yields 3 or 4 Jordans, no more, no less. For that matter, it yields 3 or 4 of every card, no more, no less. Fleer's collation was excellent during the 80s basketball years.
36 packs per box x 12 cards per pack = 432 cards per box
432 cards per box divided by 132 cards per set = 3.27 sets per box
So in addition to getting three sets per box, there's a 27% chance of finding a fourth Jordan, Barkley, Ewing, etc. This math holds true for the stickers as well. Since the mid 70s when Topps went to a single series printing through the 90s when short prints were introduced back into the hobby, it has been all about the math. You simply cannot average a Rickey Henderson rookie per 1980 Topps wax box; the math does not support it. You can't average a Griffey rookie per 1989 Upper Deck wax box (low or high); the math does not support it, no matter how bad the collation is.
Mike's description of the sequence is sort of correct. The cards fall in two parallel reverse alphabetical order sequence, alternating cards, 66 cards apart. Not counting the sticker (which is randomly paired with the 12 cards, but still follows the same mathematical insertion rate), starting from the top card (face up, next to the gum), a pack may look like this:
#58 Clark Kellogg
#123 Buck Williams
#57 Michael Jordan
#122 Gerald Wilkins
#56 Vinnie Johnson
#121 Dominique Wilkins
#55 Steve Johnson
#120 Spud Webb
#54 Marques Johnson
#119 Bill Walton
#53 Magic Johnson
#118 Jay Vincent
When you see what an uncut sheet looks like, you can see why the cards are collated this way.