PERT 3-Point Estimator
Three point estimation done right. Enter your optimistic, most likely, and pessimistic durations, get the PERT expected duration, standard deviation, and 68% and 95% confidence ranges. Multi-task mode gives the combined estimate for a whole work package.
Probability distribution
Enter each task. Totals roll up using PERT variance addition, which is more accurate than just summing.
Combined distribution
How PERT works and why it beats single point estimates
Single point estimates pretend the future is certain. PERT does not. It asks for three numbers and weights them as (O + 4M + P) / 6, then derives a standard deviation from (P - O) / 6. The output is a range with confidence, not a point.
Standard deviation gives you the 68% and 95% confidence ranges. If you commit to a deadline at the expected duration, you have roughly a coin flip of hitting it. Commit at expected + 2σ and you hit 97.5% of the time.
For multi-task plans, variances add, not standard deviations. The combined StdDev is the square root of the sum of variances. That is why a 10-task package with a 5 day expected duration each is almost never 50 days when one of them goes sideways.