It's just like in the 60-ties the US government calculated that there is a 0% risk of accidental detonation of nuclear bombs[1],because they have not experienced it yet.
I believe that electric cars are safer than gasoline cars,but the sample size for Tesla is way to small to draw such conclusions yet.
Read that book I mentioned in the original post. It mentions plenty of nuclear-related accidents. The one you probably think about happened when an armed thermonuclear bomb fell out of a bomber, yanking the arming wire out, arming every subsystem as it fell - barometric switches activated,thermal batteries ignited , and upon hitting the ground the piezoelectric crystals in the nose cone crushed activating the high explosives surrounding the core.
It would have detonated, except that the "fire-safe" switch in the cockpit was not switch to "fire". That switch was just two very low voltage wires running close together, and it was found later that pretty much anything in that bomber could have triggered it. It was extremely lucky that that bomb has not detonated that day.
There were also a few bombs which were lost in fire on the ground, couple of those had their high explosives ignite without producing a nuclear yield but spreading plutonium around, contaminating ground.
There were bombers which fell apart in the air, which crashed, most of the time the nuclear bombs they were carrying survived intact,but a few times those crashes resulted in a widespread contamination with plutonium - a famous one in Spain and another near an American Polar station that the bomber was monitoring from Soviet attack - that one is particularly interesting,because had the bombs aboard that bomber detonated and destroyed that station, it would be extremely likely that the US government would assume that it was destroyed by a Soviet nuclear strike,and not by their own machinery - and that would probably cause them to retaliate starting a nuclear war.
I believe that electric cars are safer than gasoline cars,but the sample size for Tesla is way to small to draw such conclusions yet.
[1] Eric Schlosser, "Command And Control"