Addressing the Failure of Risk Quantification: Moving from Speculating on Project Complexity to Actually Quantifying It

This workshop is based on the author’s recent book ‘Modern Risk Quantification in Complex Projects: Non-linear Monte Carlo and System Dynamics Methodologies’ published by Oxford University Press this year.

Project practitioners and decision makers complain that both Monte Carlo and especially parametric methods fail to produce accurate project duration and cost contingencies in the majority of cases. Apparently, these methods have unacceptably high systematic errors as they miss out critically important components of project risk exposure. Namely, the components associated with structural and delivery complexity are usually overlooked.

Hence, this report zeroes in on the most crucial but systematically overlooked characteristics of complex projects. Any mismatches between two fundamental interacting subsystems - a project structure subsystem (PSS) and a project delivery subsystem (PDS) - result in non-linear interactions of project risks. Three kinds of the interactions are distinguished - internal risk amplifications stemming from long-term ('chronic') project system issues, knock-on interactions, and risk compounding. Affinities of interacting risks compose dynamic risk patterns supported by a given project system. A new methodology to factor these patterns into Monte Carlo modelling is referred to as non-linear Monte Carlo methodology. (An integrated non-linear schedule and cost risk analysis (N-SCRA) version of the non-linear Monte Carlo methodology is discussed.) It allows adequately predict outcomes of even most notorious project-outliers comparable with Boston Big Dig Artery/ Tunnel project (1,400% overspending) or Sydney Opera House project (220% overspending).

Actually, all projects could be qualified as simple, complicated, complex or chaotic depending on their PSS and PDS characteristics. The traditional Monte Carlo methodology could be adequate enough but only in the case of simple projects when risk interactions are negligible. In all other cases of complicated, complex, and chaotic projects the proposed non-linear Monte Carlo methodology is a must. In addition, a power of project system dynamics methodology is uncovered based on so-called project rework cycles when project risks are consistently added. It can be adopted as an accurate risk quantification methodology in complex projects as the results produced by the system dynamics and the non-linear Monte Carlo methodologies are comparable.