If you're in a place where everyone drives a moped or bicycle and cars are rare, then mopeds are a lot safer. Stick the moped in traffic surrounded by clueless SUV drivers and now the moped is a death-trap.
Go to a place where everyone rides a bicycle and there's no cars and I guarantee the fatality-per-mile rate is ridiculously low. See downtown Copenhagen for example.
That said walking is very dangerous per mile, but cars get a lot more miles on them.
Maybe I'm just dense but how does that invalidate parent's point? Based on what you are saying (difference in trip length), it appears that fatalities per mile actually does a better job (than fatalities per trip) at representing the relative safety of different classes of vehicles...
Because cars kill a lot of pedestrians, but those count more vs people walking than vs people driving due to the relative distances involved.