Critical Response #4: On the Quantitative Definition of Risk

            Risk is an interesting concept to think about when it comes to decisions. What makes us choose certain risks over another? Stanley Kaplan and B. John Garrick make an attempt to formalize the entire process.

We begin by defining what risk really is. The article makes a distinction between ‘risk’, ‘uncertainty’, and ‘hazards’. For the most part, I definitely agree with the definitions. Risk is a combination of uncertainty and a damaging consequence – something that is only uncertain wouldn’t really be considered risky. The weather is uncertain on a day-to-day basis, but it isn’t really considered risky until there is an event that requires sun (a potentially damaging consequence).

Hazards are considered sources of pain – obstacles, open circuits, etc. while risk is considered a potential or likelihood of pain. The main distinguishing factor is the probability or uncertainty involved. I don’t think risk should be exactly defined quantitatively though; my main issue with doing so is that it could put a determined limit on what should be considered risky or what shouldn’t be considered risky. In some instances, I would argue that a more qualitatively definition is much more appropriate since the definition of risk can greatly differ on a case by case situation. I don’t think it’s necessarily possible to quantitatively graph or analyze situations since that relies on data points that we may/may not know or may/may not even have control of. Sometimes, definite numbers and graphs can give us a false sense of security or false sense of safety.

In the article, it is also brought to light that in reality, risk is infinite and shouldn’t be used to necessarily deter the construction of things like nuclear reactors.

The discussion of the distinction between ‘frequency’ and ‘probability’ is also interesting. Most of the discussion is focused around the difference in wording – but that can also influence a difference in how we think of the topic. Frequency is considered a hard measurement because it is based on observations of repeated trials and is a concrete number. On the other hand, probability still involves a degree of belief, or a state of confidence. In my opinion, the distinction is meaningful but I don’t completely agree with how it is presented. I think there is still a degree of uncertainty and a degree of confidence/belief with frequency. When I say that a lighting bolt strikes a tree once in every 10,000 lighting strikes, the degree of uncertainty is whether or not the next lighting strike will hit the tree I’m standing at. While some may argue that is their definition of probability, I would say that it is all implicitly a part of the entire notion of probability/frequency.

Overall, I agree with the conclusions that the article brings in that a single number is not big enough to make a decision. The extensive lists of different types of graph are all being used to make one decision – is this decision a safe decision or a risky decision. It’s a question that requires a lot of forethought and patience – but it’s one of the most important things that must happen on a day-to-day basis for an engineer!