Estimation Uncertainty and Quantifying Risk

Estimation uncertainty refers to the inherent lack of precision in predicting how long a software task will take. It acknowledges that unforeseen issues, changing requirements, or dependencies can impact duration. Three-point estimation (O, P, M) makes this uncertainty explicit by defining the potential estimate range (from O to P).

PERT provides a tool to quantify this uncertainty: the Standard Deviation (SD). It measures the dispersion or spread of possible outcomes around the Expected Duration (E). The simplified formula is:

SD = (P - O) / 6

Think of SD as a numerical indicator of confidence or risk:

• A small SD means the Optimistic (O) and Pessimistic (P) estimates are close together. This implies lower uncertainty and higher confidence in the Expected Duration (E). The task is likely well-understood.
• A large SD means O and P are far apart. This signifies high uncertainty, lower confidence, and greater risk associated with the task. There might be significant unknowns or dependencies.

Calculating and communicating the SD alongside the E and the O-P range provides crucial context. It allows stakeholders to see not just the likely duration but also the level of risk associated with that estimate, facilitating better risk management discussions and more informed decisions.