how to estimate project completion probability using PERT
The Program Evaluation and Review Technique (PERT) helps estimate the likelihood that a project finishes by a target date using three time estimates per task. This method translates uncertainty into a probability, letting project managers set realistic deadlines and prepare contingency plans.
What is the PERT project completion probability and when to use it
PERT completion probability is the chance that the total project duration will be less than or equal to a given deadline. Use PERT when tasks have uncertain durations, tasks are linked in a network (dependencies), and you need a probabilistic forecast rather than a single-point estimate.
Core formulas: expected time, variance and Z-score
PERT uses three estimates for each activity:
- Optimistic time (a): shortest plausible duration
- Most likely time (m): best estimate under normal conditions
- Pessimistic time (b): longest plausible duration
From these, compute:
- Expected time (TE) for an activity: TE = (a + 4m + b) / 6
- Variance for an activity: Var = ((b - a) / 6)^2
For the critical path (sum of TE across critical activities), total expected duration (TE_total) equals the sum of each activity TE. Total variance (Var_total) is the sum of each activity variance along the critical path. Standard deviation for the critical path is SD_total = sqrt(Var_total).
To estimate the probability that the project finishes by deadline D, compute the Z-score:
Z = (D - TE_total) / SD_total
Then use the standard normal distribution (phi) to find the probability: Probability = Phi(Z).
Step-by-step: manual calculation
- List all project activities and dependencies; build a network diagram and identify the critical path(s).
- For each activity on the critical path, record optimistic (a), most likely (m) and pessimistic (b) durations.
- Compute TE and Var for each critical activity using the formulas above.
- Sum TEs to get TE_total and sum variances to get Var_total for the critical path.
- Compute SD_total = sqrt(Var_total).
- Choose a target deadline D (project due date) and compute Z = (D - TE_total) / SD_total.
- Find Phi(Z), the cumulative probability from the standard normal distribution table. Phi(Z) gives the probability of finishing by D.
Practical example with numbers
Imagine a simple project with 3 activities on the critical path (A, B, C):
| Activity | a (days) | m (days) | b (days) |
|---|---|---|---|
| A | 4 | 6 | 10 |
| B | 3 | 5 | 9 |
| C | 2 | 4 | 8 |
Compute TE for each:
- TE_A = (4 + 4*6 + 10) / 6 = (4 + 24 + 10) / 6 = 38 / 6 ≈ 6.33 days
- TE_B = (3 + 4*5 + 9) / 6 = (3 + 20 + 9) / 6 = 32 / 6 ≈ 5.33 days
- TE_C = (2 + 4*4 + 8) / 6 = (2 + 16 + 8) / 6 = 26 / 6 ≈ 4.33 days
Sum TE_total = 6.33 + 5.33 + 4.33 = 16.0 days (rounded).
Compute variances:
- Var_A = ((10 - 4)/6)^2 = (6/6)^2 = 1^2 = 1.00
- Var_B = ((9 - 3)/6)^2 = (6/6)^2 = 1.00
- Var_C = ((8 - 2)/6)^2 = (6/6)^2 = 1.00
Var_total = 1 + 1 + 1 = 3. SD_total = sqrt(3) ≈ 1.732 days.
Suppose the deadline is D = 18 days. Z = (18 - 16) / 1.732 ≈ 1.155.
Phi(1.155) ≈ 0.875 (87.5%). So there's about an 87.5% chance the project finishes within 18 days.
How to use an online PERT calculator
Online PERT calculators speed up steps: enter tasks, dependencies and the three time estimates. The tool computes the critical path, TE_total, Var_total, SD_total and gives the completion probability for any deadline.
Use the calculator on Calculatorr to:
- Quickly test different deadlines and see how probability changes.
- Identify which activities contribute most to variance (useful for risk mitigation).
- Export results for stakeholder reports.
Tip: when using an online tool, double-check the critical path it identifies against your project network — incorrect dependencies can yield wrong probabilities.
Interpreting results: what the probability means and how to act
- High probability (e.g., >90%): deadline is likely realistic, but still review critical tasks for hidden risks.
- Medium probability (50%–90%): consider contingency time, resource buffers, or accelerating critical tasks.
- Low probability (<50%): deadline is risky; re-plan scope, add resources, or negotiate deadline.
Remember: PERT assumes task durations are statistically independent and the sum of many activities approximates a normal distribution. If tasks are correlated or durations have fat-tailed distributions, PERT probabilities can be misleading.
Common errors and how to avoid them
- Using inaccurate estimates: get input from people who will perform the work and use historical data when possible.
- Ignoring non-critical path risks: a near-critical path can become critical if delays occur; reassess network regularly.
- Overlooking dependencies: incorrect precedence relationships yield wrong critical path and probabilities.
- Assuming normality blindly: for small projects or skewed tasks, consider Monte Carlo simulation for a more robust probability estimate.
- Rounding too early: keep decimals through intermediate steps and round only final outputs to avoid cumulative rounding error.
Extensions: Monte Carlo vs PERT
PERT gives a fast, closed-form probability using the normal approximation. Monte Carlo simulation samples task durations from specified distributions (triangular, beta, lognormal) many times to build an empirical completion-time distribution. Use Monte Carlo when:
- Task durations have strong skew or known non-normal behavior.
- There are correlations between task durations.
- You need percentiles beyond what normal approximation reliably provides.
Both approaches can be complementary: run PERT for a quick estimate and Monte Carlo for a detailed risk analysis. Calculatorr offers tools and guides to compare both methods and pick what fits your project complexity.
Quick checklist before trusting the PERT probability
- Verify activity estimates (a, m, b) with team members.
- Confirm the critical path and adjacency of tasks in your network diagram.
- Check variance contributors—focus on activities with large (b - a) spreads.
- Consider running a sensitivity or Monte Carlo analysis if results are borderline.
- Use the probability together with contingency plans, not as the sole decision criterion.
Use Calculatorr's project and probability tools to enter your tasks and get instant TE, SD and completion probabilities. Link results to your calendar and stakeholder reports for transparent planning and risk communication.
Example: adjusting the deadline
From the earlier example (TE_total = 16, SD_total ≈ 1.732):
- Deadline 17 days: Z = (17 - 16)/1.732 ≈ 0.577 → Phi ≈ 0.72 (72%)
- Deadline 16 days: Z = 0 → Phi = 0.5 (50%)
- Deadline 15 days: Z = (15 - 16)/1.732 ≈ -0.577 → Phi ≈ 0.28 (28%)
This quick table shows how a one-day change can shift probability significantly when SD is small—use such sensitivity checks to decide on buffers.
Where to go next
After getting a PERT probability, prioritize actions: reduce variance on high-impact activities, add parallel resources where feasible, or negotiate scope/deadlines. For complex projects, combine PERT with Monte Carlo simulations and use tools on https://calculatorr.com/ to automate calculations and produce visual reports.
