[pointman]

I could be wrong, but isn't there a fundamental problem underlying the data generated trying to compare fatalities on different surfaces in that they assume that every fatality is caused by the surface?