Not all high-consequence events make the news and capture national attention, like the Deepwater Horizon explosion or the release of methyl isocyanate in Bhopal, but by definition they cause serious, life-altering damage, injuries and deaths, affecting not only the persons involved, but their families, friends and coworkers, and in severe cases, the local community. At the very least, businesses are likely to suffer damage to their reputations, and the consequences may be much worse. According to the Federal Emergency Management Agency (FEMA), 40 percent of businesses do not reopen after a disaster, and another 25 percent fail within one year of reopening. A large number of these events had an extremely low probability of occurring, and yet they did. Nicolas Taleb refers to these types of occurrences as "black swan" events, and describes them as outliers, with extreme impact, and retrospective (though not prospective) predictability.
Current approaches to risk assessment are based on a determination of the probability and severity of incident outcomes. In addition to evaluating normal operating conditions, more sophisticated risk assessments often also evaluate the outcome of interruptions to normal operations (upset conditions). However, simple probability and severity analysis may not be adequate in evaluating low probability, high-consequence events.
In order to start understanding low-probability/high-consequence events, we want to look first at theories of how accidents are caused. One of the earliest theories is the Domino Theory, developed by H.W. Heinrich in 1932. As depicted below, it posits that accidents occur when the linkage for a chain reaction lines up like dominoes on end. Each of the factors in the chain is dependent on the preceding factor.