Modelling a risk event | Vose Software

# Modelling a risk event

The figure below illustrates a model to estimate the impact of a set of risks that may impinge on a project.

In this model the total cost of a project is being estimated. Seven uncertain elements have been modelled:

• The base project cost;

• The potential impact of five identified risks: Health and Safety Executive intervention; a strike; bad weather sub-contractor insolvency and a change in the ruling political party;

• The rate of inflation

The base project cost is modelled by a simple Triangle distribution in Cell C10. The inflation rate is modelled in Cell C23 with a PERT distribution. The selection of a Triangle or PERT to express uncertainty given a three point estimate (minimum, most_likely, maximum) is discussed elsewhere.

The point of this model is really to illustrate a way of modelling inter-related risk events. H&S, bad weather, and political change risks have 10%, 30% and 2% probability of occurring. The risk of strike, however, has a 15% chance of occurring unless the H&S risk occurs, when it is considered the probability increases to 30%. The insolvency probability is 5%, but goes up to 75% if the H&S and the strike risks both occur. We can use conditional logic with Excel's IF function, depending on whether or not the F column (see below) contains a zero, to alter the probability of these two risks accordingly.

Column E models the impact of the risk: a range of 80% to 150% of the most likely risk impact is modelled using a Triangle distribution object (80% and 150% is for the convenience of illustration: we recommend that you review each risk separately). Column F uses the VoseRiskEvent function that returns a random value from the impact distribution if the risk occurs, and a zero otherwise.

The effect of this model is to recognise that the H&S risk has a much more significant impact than one might suppose when reviewing it in isolation. It is extremely common for risks to be inter-connected: for example, a certain risk occurring might draw resources to manage it that are no longer available to prevent another risk. The occurrence of a risk might also affect the size of an impact of another risk. We haven't shown it here, but it is simply modelled by using the same IF logic on the Most Likely (M L) value column.

The spreadsheet of this model, which also includes the Triangle and Pert distributions, is provided here: risk portfolio