The problem with statistics is that they are often misinterpreted: for example, if 30% of
all accidents are caused by drunken drivers, it does not imply that 70% are caused by
sober ones, who are therefore the greater danger: so you have to think about what the
figures really mean