The UK Military Aviation Authority, in its regulatory article 1210, states that "the severity of a Single Risk is an assessment of the worst credible outcome that could result from the hazard" (RA 1210 issue 5). The relevance of this being included in the paragraph relating to severity should not be overlooked because it encourages the risk manager to remove the idea of likelihood from their assessment of "worst credible".
In my opinion this is where the good people from Aviation Risk Management Solutions may have caused confusion. I encourage you to read their excellent paper entitled ARMS Methodology for Operational Risk Assessment in Aviation Organisations (v4.1 March 2010), and this really is my only criticism, but the event risk classification matrix poses the question: "What was the effectiveness of the remaining barriers between this event and the most credible accident outcome?" I believe the word “most”, introduces an element of likelihood into the question.
The CAA UK doesn't aid clarity of thought either: "Organisations should define whether they are assessing severity using the worst case scenario or the most credible outcome" (CAP 795 Feb 15).
Why does this matter? The other day I had a long discussion with some very knowledgeable people about the difference between worst case, worst credible and most credible accident outcomes and I realised two things: 1. how passionate people are about this and 2. the importance of nuance in the language we use.
Let's take the example of a runway excursion. The most likely outcome of a runway excursion is the aircraft departs the runway surface and comes to a halt in a safe area to the side or end of the runway. This may be the most likely outcome but in the discussion we were having it was being conflated with the most credible outcome. But that's right, isn't it? The most credible outcome is that safety barriers are in place and working and a serious accident is avoided. Straight away we have added likelihood into the assessment of the severity of the outcome. If we went down this path, why would we bother with a RESA, frangible structures etc?
We then got onto the "worst case" outcome. One suggestion that was an aircraft departs the runway surface at high speed, collides with a building full of people and both aircraft and building are destroyed in the ensuing fire. My question was whether or not this was a credible scenario and was it more than a single risk or a result of multiple failures? I feel that using “worst case” terminology engenders thinking way beyond that which is conceivable, or valid.
Credible = conceivable, valid, plausible. This will still be a subjective assessment, but it should be de-coupled from likelihood.