Capacity-aware Parr model improves agile project forecasting

Classical effort distribution models like the PNR family and the Parr alternative curve assume a variable staffing pattern over time. This assumption breaks down in agile environments where team capacity is fixed, partially fixed, or externally constrained. Pedro Colla’s paper introduces a capacity-aware refactoring of the Parr model that adapts to real-world agile constraints. By combining a normalized latent effort demand (the classical Parr curve) with an observed or planned capacity trajectory, the model forecasts progress, completion time, capacity deficit, and slack without requiring the same internal activity path as the original model.

The model uses only a handful of parameters—total effort K, a Parr shape parameter, an origin constant c (to match nonzero initial staffing), and the capacity trajectory itself. A discrete sprint formulation enables direct application in Scrum. The paper also provides a calibration method from ordinary Scrum records and a rolling origin validation protocol against simple management baselines. This makes it practical for teams that want to detect schedule risks early without complex causal models.