PERT Estimator

How it works

1. 1List the activities you are estimating. Three sample rows are filled in — overwrite them.
2. 2For each one give an optimistic, most likely and pessimistic figure in whatever unit you work in.
3. 3The expected duration is (optimistic + 4 x most likely + pessimistic) / 6, and the standard deviation is (pessimistic - optimistic) / 6.
4. 4Across activities the variances are added, not the standard deviations — which is why the combined range is tighter than adding the individual ranges.

Common questions

Why is the total range narrower than the sum of each activity’s range?

Because the activities are treated as independent. For every task to hit its pessimistic figure at once, every risk has to land together — unlikely. Statistically the variances add and the standard deviation is the square root of that sum, which grows more slowly than a straight total.

What does the confidence range actually mean?

A 90% range means that if your three-point estimates are honest, roughly nine times in ten the real duration lands inside that band. It is not a guarantee, and it says nothing about risks you did not estimate.

Should I commit to the expected duration?

The expected value carries about a 50% chance of being met, so committing to it means missing half the time. Most teams commit somewhere near the 85–90% figure and hold the difference as schedule contingency.

Why did my activity get a warning?

Either the three values are out of order — optimistic should be the smallest and pessimistic the largest — or there is no gap between optimistic and pessimistic, which means you have declared the activity risk-free.

Does this assume a normal distribution?

Each activity uses the beta distribution behind classic PERT. The rolled-up total and the probability figure use a normal approximation, which is reasonable once you have several independent activities.