In this chapter, we highlight some core principles in the design of risk models. Many of these are similar to those in traditional static modelling, but there are, of course, some differences.
The extent of similarity is often overlooked, resulting in modellers who are competent in traditional static approaches being apprehensive to engage in potential risk modelling, or overwhelmed when initially trying to do so. Key similarities include:
- The importance of planning the model and its decision-support role, bearing in mind the context, overall objectives, organisational processes and management culture.
- The importance of keeping things as simple as possible, yet insightful, useful, but not simplistic.
- The importance of distinguishing the effect of controllable items (i.e. decisions) from that of non-controllable variables, both in terms of model building and in results presentation.
- The use of sensitivity thought processes as a model design tool.
- The need to build formulae that contain logic that is flexible enough to capture the nature of the sensitivity or uncertainty.
- The use of switches to control the running of different cases or scenarios.
- The need to capture general dependency relationships, both between inputs and calculations, and between calculated items.
- Very often, from the perspective of the formulae required in a model, it is often irrelevant whether the process to vary an input value is a manual one, or one that ...