Credibility Evaluation of Project Duration Forecast Using Forecast Sensitivity and Forecast-Risk Compatibility

Abstract

This paper presents a credibility evaluation framework (CEF) that analytically evaluates deterministic project duration forecasts to detect false early warnings and misleading trends. The CEF consists of forecast sensitivity evaluation, forecast-risk compatibility check, and independent sanity checks using probabilistic models. New concepts of the forecast sensitivity index and threshold are defined and used to establish an unrealistically sensitive prediction zone as a graphical tool for quick and quantitative credibility evaluation. The forecast-risk compatibility check assesses the consistency of schedule variability predicted by deterministic forecasts and by standard risk assessment techniques. A hierarchical decision structure based on the compatibility check is also presented. Practical implementation and effectiveness of the CEF are demonstrated using both a notional and a real project. The framework would positively contribute to improving the effectiveness of deterministic schedule forecasts and would be best employed at both project and program levels prior to taking interruptive and burdensome control actions in order to identify false warning signals.