PERT Distribution

The PERT (Programme Evaluation and Review Technique) distribution is a modified beta distribution parameterised by minimum, most likely, and maximum values — the same three inputs as the triangular distribution — but with a different shape. Rather than rising and falling linearly to the mode, the PERT distribution forms a smooth bell-like curve that gives significantly more weight to the most likely value and tapers more gradually toward the extremes. The PERT mean is calculated as (minimum + 4 × most likely + maximum) ÷ 6, which gives four times the weight to the mode compared to the simple arithmetic mean of the three values.

The PERT distribution is generally considered more realistic than the triangular distribution for project cost and duration estimates. In reality, activities are more likely to complete near the most likely value than near the extremes — the probability does not increase linearly up to the mode and then drop off a cliff. The PERT distribution captures this concentration around the central estimate while still allowing for tail events. It is widely used in both schedule risk analysis tools and cost risk models.

The distinction between PERT and BetaPERT is subtle: the original PERT distribution uses a fixed weighting factor of 4 for the mode, while BetaPERT allows the weighting factor to be varied (typically with a lambda parameter of 4 as the default). For most project risk analysis purposes, standard PERT and BetaPERT with default settings produce very similar results. The choice between them is less important than ensuring the three input values are well-calibrated. The most common error with PERT inputs — as with triangular — is setting the minimum and maximum too close to the most likely, producing a distribution that underestimates variability.