Probabilistic Chain Analysis (PCA): A Framework for Running Projects in a Probabilistic World

This paper introduces Probabilistic Chain Analysis (PCA), a quantitative risk methodology for infrastructure and EPC project delivery that models project risks as causally dependent chains rather than independent events. Traditional risk registers and standard Monte Carlo (MC) simulation assume risk independence, systematically underestimating tail exposure by treating cascade amplification as zero. PCA addresses this structural limitation through three integrated components: (1) cascade chain mapping via directed acyclic graphs (DAGs), (2) Bayesian Network-coupled Monte Carlo simulation with Latin Hypercube Sampling and Gaussian Copula correlation, and (3) an Intelligence Log framework that carries Bayesian-updated posteriors across project cycles, building institutional risk memory. The methodology is validated through a calibrated UK highways Survey and Investigation scenario, constructed from a cross-referenced dataset of 16,000 infrastructure projects spanning eight sectors. Results demonstrate that cascade lift factors of 3.3–4.1 exist on six risk nodes that static registers treat as independent. A single targeted pre-mobilisation intervention — identified through mitigation what-if analysis — reduced P90 cost exposure by £250,000 at a management cost of £15,000, representing a 16.7x return on risk management effort. CVaR90 was reduced by £101,025 against the standard Monte Carlo baseline. The paper presents the full theoretical foundation, empirical calibration methodology, case study results, and a six-step practitioner framework for applying PCA within a standard project risk review cycle. Limitations, including subjective probability inputs and schedule-cost decoupling, are documented. PCA is positioned as a practitioner-accessible bridge between academic Bayesian risk theory and live infrastructure project delivery. Keywords: probabilistic risk analysis, Bayesian networks, Monte Carlo simulation, cascade risk, infrastructure project management, cost overrun, CVaR, EPC, risk register, project controls