We are remarkably safe. Across a wide range of industries fatal accident rates have been falling for decades, and accidents are less significant as a cause of death than at any time during the 20th century. This is despite the fact that we are using analysis techniques which were developed decades ago for applications that were far simpler than the complex technical and socio-technical systems we produce today. In fact, the situation is rather worse than this. Examination of industrial practice shows that these inadequate techniques are often even more inadequately applied. These concerns apply, mutatis mutandis, to safety standards. This leads us to enquire “how did systems get so safe?” If we can answer the question (or shed light on it) then this should help us to focus effort on methods that are effective, and thus to preserve safety in the face of growing complexity of, and societal dependence on, safety-related systems.