Wulf,
You're both correct - to an extent.
The poster's original math ignored the order in which your failures occurred, so long as you didn't have more than the number required to actually fail.
It also, however, assumed optimal skill check numbers for all people rolling.
pemerton's point is...