Four Types of Risk

I have been reading an old blog about “Epistemic, Ontological and Aleatory Risk” on Mathhew Squair’s blog, Critical Uncertainties.  I doubt I am alone in having to think hard to get my head around the definitions.  So, to help myself, I made a diagram.

There is a lot to think about in Matthew’s article and it stimulated a lot of parallel thoughts. The summaries and interpretations in this diagram are all mine, but it owes a lot to Matthew’s article.

Four Types of Risk

4 thoughts on “Four Types of Risk

  1. Glen B. Alleman

    Mike, the four risk “types” are all epistemic. You lack knowledge and you can acquire knowledge by “buying it.” “Buy it” means more research, more testing, prototype, or other means of acquiring information that will reduce the uncertainty.
    Matthew’s use the terms aleatory, epistemic etc. should start with “uncertainty.” This are classifications of uncertainty, that result in risk.
    Aleatory uncertainty results in “natural variances.” Demining Variances. The resulting risk can not be “brought down.” Only margin can protect you from that risk. Schedule, cost, and technical “margins.”
    Even in your upper right box, the Known Knowns may be epistemic. Aleatory uncertainty only has a probability distribution, but you don’t “know” the probability of on outcome. Just the distribution function. The variance that will “naturally” occur on that activity, or product, or measurement.
    One non project example is the water pump mounting bolt holes for a Toyota Camry. The machine that makes the holes, doesn’t make them perfectly. The is “slop” in them. There is an uncertainty to their “exact” location and dimensions. Do “margin” is made for the receiving positions on the engine block so if the water pump is “within tolerance” it will fit the “within tolerance” block holes.
    For a project, activity durations need “slop” to assure they fit within the planned completion date. This is an aleatory uncertainty that can be discovered with Monte Carlo simulation, but we can’t “buy down” this uncertainty without changing the underlying statistical generating function for the activity.

    1. Matthew Squair

      Risk can also exist in the Unknown, Known quadrant. For example if risk is ‘known’ in one group (say Morton Thiokol’s engineers) but not effectively communicated to another (say NASA and Thiokol management) then you can say that the resultant management decision (say to launch on a cold day well outside historical norms) was made under uncertainty (a known/unknown). Information needs to get to the right parts of an organisation to be effective.

    2. Matthew Squair

      And in response to Glenn, yes I agree that it’s all ‘Uncertainty’. However I don’t agree that it’s all epistemic. I think there’s a useful differentiation between those unknowns about which we can say something meaningful, and potentially do so mathematically, versus those uncertainties which we can’t. I’ve used the term ‘ontological’ to denote the cardinal difference between the question of what we understand reality to be (a question that is ontological in nature) versus what (and how) we ‘know’ about that reality ( a question epistemic in nature). I think it’s a useful exercise in classification and maps well into the four quadrant model of knowledge.

  2. Matthew Squair

    Hi Mike, I wrote a followup piece on the concept of four quadrant risk (and the philosophy of Don Rumsfeld) which can be found here As an interesting note, I’m aware that to reduce ontological uncertainty (unknown, unknown)in climate change studies it’s common practice to run different models of the same climate processes and cross compare the results. I like the diagram by the way 🙂


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s