Computing risk drivers

What are risk drivers?

Risk drivers provide a way to understand which activities drive a given risk position for a selected activity (e.g., P80), and the extend to which they do as a percentage of that risk position. 

Risk drivers are organised into two categories - duration uncertainty and risk events. Risk drivers under duration uncertainty are essentially activitities whose uncertainty impact the risk position of the activity you have selected. Risk drivers under risk events are activities whose risk events impact the risk position of that activity.  

Once you select a target activity and the p position you care about it, you will see at most, theTop 20 activitities that drive it's P-position in the form of an interactive tornado chart. For each activity, you will see a repsective bar that indicates the level of impact that activity has, both with respect to uncertainty and risk events. 

How do you compute risk drivers?

We analyse the results of Monte Carlo simulations to determine risk drivers. In particular, we correlate the target activity’s finish delays that arise from the the Monte Carlo simulations, with the driving activity's change in duration from that same simulation. In other words, we find the delay that occurs on the target activity, and correlate the change in that activity's duration with the duration changes that took pace in all the other activitites. 

We then normalise the correlations of all potential driving activities to 100%, and show you the Top 20. The rest are aggregated and presented separately as an additional percentage. 

Each driving activity has contributions from duration uncertainty and/or risk events. We correlate each contribution independently with the target activity delay. Summing up the correlations of all contributions we find by which fraction they impact the total correlation for this specific driving activity, information that is presented on the tornado plot.