Wow,
Cak that graph is an appalling abuse of data, as
Vessbot explains on his thread. The central point is that the use of FBW should, if correctly executed, have had a positive impact on risk. It hasn’t, and while we will never know which deaths were prevented by it, we know that deaths have been caused by it, IN THE WAY THAT AIRBUS INSISTED UPON ENGINEERING IT (sorry for shouting but the point needs to be made forcefully).
Their corporate mental model that it was a cure-all was equally damaging - I still have a copy, somewhere, of their letter to operators which stated that UA recovery was not a necessary syllabus item on their products.