Monte Carlo simulation of construction costs using subjective data: comment

This note supplements papers by Chau in an attempt to further the consideration of risk measures in forecasts of construction project costs. The note reviews the application of PERT-based aspects of analysis and simple portfolio theory. Consideration is given to subjective factors in cost forecasting which arise from the natures and experiences of the forecasters as well as pressures to which they (may) feel subjected. Alternative measures of risk exposure are examined and supplementary analyses of the data contained in Chau's paper are reported. Although the log-triangular distribution is confirmed as a superior means of measuring risk, the observed risk measures still exceed those calculated.