Was the Deepwater Horizon incident a “Normal” accident?

•The paper addresses the much-debated questions about accident “causes”. It seeks to explore whether “normal” variabilities in operating conditions and environments can explain unexpected outcomes and even accidents.•It uses the Functional Resonance Analysis Method (FRAM), to examine the high-profile Macondo Well incident in the Gulf of Mexico.•Hollnagel developed the FRAM approach originally to look at accidents and achieved a breakthrough by abstracting the system model to the level of functions, to show exactly the complexity and interdependability identified by Perrow in his Normal Accident Theory.•The results show that, rather than the highly unlikely simultaneous failure of a number of barriers designed to prevent Blowouts, the pattern of the events on the day can be explained as the way variabilities in the effectiveness of key functions combined to allow the disastrous consequences that emerged as an extreme, but still predictable outcome of “normal” behaviours.•It concludes that on this basis it was a “normal” accident but agrees with the Hopkins critique of Perrow’s Normal Accident Theory that such accidents are predictable extremes of normal behaviour which should be expected and are not inevitable.