Conclusion

This chapter provides a high-level overview of concepts, models, and results discussed in this book. It is accentuated that risk quantification used in decision-making can be tagged as ‘political mathematics’. The political aspects are defined by various realizations of bias from strategic misrepresentation to the Hiding Hand principle. It is pointed out that the mathematical aspects may succumb to the similar realizations of bias as political ones. Recommendations to balance the political and mathematical aspects are proposed. A role of the non-linear Monte Carlo N-SCRA methodology supported by the project system dynamics modelling is established as a role of the Revealing Hand. It should provide ammunition for decision-making before the project approval, not after in a form of lessons learned as the Hiding Hand does. This is a way to decline the invitation to attend a ‘banquette of consequences’ referring to project failures as ‘unpleasant unsurprises’.

Non-linear Monte Carlo Modelling Requirements

The purpose of this chapter is to finalize requirements to undertake non-linear Monte Carlo modelling. Besides mathematical aspects (Chapter 13), additional requirements to identification and addressing of risk interactions should be put forward. All relevant instances of risk interactions should be identified. Identification of ‘chronic’ project system issues that serve as additional causes of risks is required to pin down internal risk amplifications. Identification of cross-risk interactions can be undertaken by visualization of dynamic risk patterns (cross-risk interaction mapping). Principles of risk interaction calibration are introduced keeping in mind two challenges. First, a calibration of aggregated risk interactions at the project level. Second, evaluation of individual instances of risk interactions based on the overall calibration. Methods to address identified risk interactions are discussed. In the case of intra-risk interactions, the ‘chronic’ issues should be addressed. In the case of cross-risk interactions, the dynamic risk patterns should be disrupted.

A Project Structure Subsystem (PSS)

This chapter provides a high-level discussion on the project structure subsystem (PSS) through description of main typical ‘project parts’ and their interactions. Project deliverables, labour requirements, locations, technologies, interface points, and stakeholders are reviewed as ‘project parts’. A discussion on interactions in complex systems including risk interactions is reiterated to tell apart intra-risk interactions (internal risk amplifications) and cross-risk interactions (knock-on and compounding interactions). A role of interface points for propagation of cross-risk interactions is accentuated. High-level PSS description for simple, complicated, and complex projects is provided. Chaotic projects are considered complex projects that have been pushed to the edge of chaos and beyond due to overwhelming risk interactions.

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.

A Project Delivery Subsystem (PDS)

This chapter comes up with a high-level description of a project delivery subsystem (PDS) that is supposed to match the project structure subsystem (PSS). If PDS has long-term ‘chronic’ system issues that serve as causes of project risks, the static PSS–PDS mismatch is inevitable leading to internal risk amplifications of corresponding risks. Additional dynamic PSS–PDS mismatches can result from risk interactions and even from standalone unamplified project risks. Two primary PDS factors are considered: adopted project development procedures and consistency of their implementation. Procedures related to opportunity shaping, project team design/shaping, engineering, contracting, construction, stakeholder management, and change and risk management are overviewed. Factors influencing procedure’s implementation—available skills and realizations of bias are discussed. The importance of risk quantification, scheduling, and estimating procedures and their implementation is accentuated. Various realizations of relevant bias—from strategic misrepresentation, to the Hiding Hand, to 10% syndrome—are revealed.

Project System Dynamics (SD)

This chapter demonstrates project system dynamics (SD) models that mirror the Monte Carlo L-SCRA models developed in Chapter 5 (project Zemblanity). Historic overview of project system dynamics is undertaken: from claims and dispute resolutions to modelling of project rework cycles and project changes. Approaches to model extra risk-related works, out-of-sequence works, impacts on productivity of works, etc. are introduced. Three SD models are built that combine engineering and construction rework cycles (the two-rework-cycle models). The first is a risk-free model to become aligned with the project Zemblanity schedule. The second is a model that takes into account only quality-related risks. The third factors in all risks of project Zemblanity. Results of the SD modelling demonstrate good alignment with L-SCRA Monte Carlo outcomes and explain reasons for a non-linear behaviour of the project Zemblanity system. All three workable SD models are available on the book’s companion website.

Project Zemblanity Business Case (II)

This chapter is devoted to the second business case of project Zemblanity. Based on the developed risk quantification principles for complex projects, two non-linear Monte Carlo schedule and cost risk analysis (N-SCRA) models are developed. These models factor in all relevant risk interactions before and after addressing. Modified ‘non-linear’ project risk registers that take into account the risk interactions are developed as inputs to the Monte Carlo models. It is shown that before risk interaction addressing the forecast project duration and cost are unacceptably high due to unaddressed risk interactions. Agreed risk interaction addressing measures factored to the models result in the acceptable project duration and cost. A joint confidence level (JCL) concept is used to amend the N-SCRA results at the P70 confidence level to distinguish stretched targets and management reserves using JCL70. The two workable N-SCRA models are available on the book’s companion website.

Introduction to Project Complexity

This chapter provides a literature review of the complexity concept and selects practical ideas that can be used for risk quantification in complex projects. This topic has two angles—complexity of project dealing with practical methods and complexity in project that reflects on general complexity ideas coming mostly from physics. A classification of projects—simple, complicated, complex, and chaotic—is adopted. Three types of interactions in project systems are distinguished—pooled, sequential, and reciprocal. As these interactions exist even in ideally developed project systems, only deviations from the ideal case are of interest that are interpreted as project risks and their interactions. Three analogues types of risk interactions—internal risk amplifications, knock-on interactions, and risk compounding are distinguished. Two fundamental interacting subsystems of a project system are introduced—a project structure subsystem (PSS) and a project delivery subsystem (PDS). Project risks and their interactions are associated with PSS–PDS mismatches.

Conventional Project Risk Management (PRM) Methodologies

Overview of the conventional PRM methodology is based on the introduction of a PRM system that comprises a PRM context, a PRM process, and PRM tools. Typical PRM process steps to identify, address, assess, monitor, etc risks are reviewed to comply with PRM context (i.e. features and challenges of a project). Main sorts of uncertainties related to risks as possible impacts on project objectives—uncertainties of identification, manageability, assessments of impacts and probabilities, and favourability—are discussed. Key PRM tools to support the PRM context and the process—a PRM plan, a bowtie diagram, a risk breakdown structure (RBS), a risk assessment matrix (RAM), and a risk register—are introduced. A role of psychological and organizational bias as a systematic error in the PRM process is elaborated.

High-level Overview of Project Risk Management (PRM)

Selection of adequate project risk management (PRM) methodologies should be based on deep understanding of PRM context of a project to avoid PRM context—PRM method mismatch. Various realizations of bias can impede the selection. All PRM methodologies can be grouped as either conventional (they follow traditional PRM process steps to identify, evaluate, address, monitor, etc.) and unconventional (they are focused mostly at risk assessments). All PRM methodologies—deterministic (scoring), probabilistic (Monte Carlo), parametric, etc—have limitations. Deterministic (scoring) methods can help with development of risk addressing although they are useless for development of project contingencies. Monte Carlo methods can be used for development of project contingencies only if they take into account all relevant components of project risk exposure consistently. Parametric methods suffer from using biased sampling—convenience and judgement sampling—that undermine their accuracy. Two emerging methodologies—system dynamics and artificial neural networks (ANN)—can be considered unconventional.