Overview of Unconventional Risk Quantification Methods
Unconventional risk quantification methods focused on risk assessments only are reviewed. It is shown that all unconventional methods are actually not risk quantification methods being aimed at evaluation of overall project durations and costs. Three methods are reviewed—parametric, artificial neural networks (ANN), and system dynamics. Reasons for high systematic errors and low accuracy of parametric methods are uncovered. They stem from using biased sampling—convenience and judgement samples—as well as from mathematical shortcomings of parametric methods. ANN methodology is positioned as a possible fundamental upgrade for parametric methods. It is shown that the ANN method also depends on quality of used samples. The system dynamics method is introduced to reveal non-linear interactions in project systems based on implications of project rework cycles. It is pointed out that if identified and addressed project risks are factored into the system dynamics modelling the latter could be requalified as a conventional methodology.