If one incident can reverse your conclusion then the sample size isn't large enough to draw that conclusion.
With only one death, the only conclusions you could reasonably reach from statistics are either "we don't know yet" or "we've gone so many miles with only one death that we know it's safe enough." There's no reasonable way to argue from the stats that we know it's not safe.
With only one death, the only conclusions you could reasonably reach from statistics are either "we don't know yet" or "we've gone so many miles with only one death that we know it's safe enough." There's no reasonable way to argue from the stats that we know it's not safe.