Except that article you link does a terrible job of presenting this. They look only at a small sample (20 accidents), made up only of accidents which had both f

Except that article you link does a terrible job of presenting this. They look only at a small sample (20 accidents), made up only of accidents which had both fatalities and survivors. When what you want for a real risk assessment is to consider first the accident rate among commercial flights, then the percentage of accidents which have fatalities (and also the rate at which those accidents have any survivors).