Realistic and Useful: Toward Better Estimates

The PERT technique of three estimates is a traditional, well-known approach to expert estimation. Estimators are asked for their optimistic, most likely and pessimistic estimation; then, a beta probability distribution is built to fit the data. Having a probability distribution for each activity allows a Montecarlo simulation to estimate the overall project effort with better accuracy. Still, experiments and practice suggests that estimators are often giving very narrow ranges, even in front of significant uncertainty, leading to unrealistic estimate. Magne Jørgensen has developed an alternative approach that has been proven to increased realism. However, Jørgensen’s technique is not directly based on a probability distribution, and therefore it is not directly possible to apply a Montecarlo simulation. Here a novel approach, unifying Jørgensen’s technique and PERT 3 estimates, is proposed and empirically validated using published data. Empirical validation shows that both realism and accuracy improve significantly using the unified approach.