Project Duration Estimator (PERT)
Estimate realistic task or project durations using the PERT three-point technique, which blends optimistic, most-likely, and pessimistic scenarios into a single weighted average. Use it whenever a single-point estimate feels too uncertain.
About this calculator
The Program Evaluation and Review Technique (PERT) was developed by the U.S. Navy in the 1950s to handle scheduling uncertainty in complex projects. Instead of relying on a single guess, PERT collects three estimates: the Optimistic duration (O, best case), the Most Likely duration (M, realistic case), and the Pessimistic duration (P, worst case). The weighted average formula is: PERT Estimate = (O + 4M + P) / 6. The factor of 4 on the most-likely estimate reflects the assumption that a beta distribution best models task durations, giving the central scenario four times the weight of either extreme. The standard deviation of the estimate is σ = (P − O) / 6, which can be used to build confidence intervals: the estimate ± σ covers roughly 68% of outcomes, while ± 2σ covers 95%. This transforms a subjective guess into a statistically grounded planning figure.
How to use
A development task has an Optimistic estimate of 4 days, a Most Likely estimate of 7 days, and a Pessimistic estimate of 16 days. Step 1 — apply the formula: PERT = (4 + 4 × 7 + 16) / 6 = (4 + 28 + 16) / 6 = 48 / 6 = 8 days. Step 2 — calculate standard deviation: σ = (16 − 4) / 6 = 2 days. Step 3 — 95% confidence range: 8 ± (2 × 2) = 4 to 12 days. Plan for 8 days but communicate a 4–12 day range to stakeholders who need to understand the uncertainty.
Frequently asked questions
How is the PERT estimate different from a simple average of the three durations?
A simple average weights each of the three estimates equally: (O + M + P) / 3. PERT deliberately over-weights the most-likely estimate by a factor of four — (O + 4M + P) / 6 — because experienced estimators generally have the best intuition about the realistic scenario, while extreme cases are harder to predict precisely. In practice this means the PERT estimate is pulled closer to the mode of the distribution rather than the midpoint of the range. For a symmetric distribution the two formulas agree, but for skewed tasks (e.g., a long pessimistic tail) PERT gives a more conservative and more accurate forecast.
What confidence level does the standard PERT formula represent?
The standard PERT formula (O + 4M + P) / 6 produces a point estimate corresponding roughly to the mean of a beta distribution fitted to the three points — it is not inherently a confidence interval. To obtain confidence intervals, you calculate the standard deviation σ = (P − O) / 6. The estimate ± 1σ covers approximately 68% of outcomes, ± 2σ covers about 95%, and ± 3σ covers 99.7%, assuming the underlying distribution is approximately normal. For critical-path tasks, project managers often use the +2σ value as a schedule buffer to reduce the probability of overrun to around 5%.
When should I use PERT estimation instead of story points or analogous estimation?
PERT is best suited for tasks where you can identify plausible best, worst, and expected cases but lack historical data for statistical modelling — common in novel engineering, construction, or research work. Story points work better for agile software teams with established velocity data because they capture relative complexity rather than calendar time. Analogous estimation (scaling a past similar project) is faster when you have directly comparable historical work. Use PERT when a task is unfamiliar enough that a single estimate feels risky but structured enough that a subject-matter expert can reason about the three scenarios meaningfully.