Three-point estimating uses optimistic, most-likely and pessimistic task outcomes. Classic PERT calculates (O + 4M + P) / 6 as a modeled expected value. It can make assumptions visible, but neither the mean nor an approximate spread is a promised completion time or a confidence interval.
Terms used in this guide. fee: the amount charged to the client; cost: the expense included in the stated project model; margin: revenue left after the included costs; margin percentage = that amount / revenue × 100; ROM: rough order of magnitude: an early estimate for a planning decision.
In plain words
Three-point estimating asks for three numbers instead of one: a best case, a most likely case and a worst case. The PERT formula combines them as (best + 4 x likely + worst) divided by 6. It shows where the risk is and gives a more balanced figure than a single guess.
Who this helps: Project managers and team leads estimating uncertain tasks, and anyone studying project management exams.
Questions this guide answers:
- What is the PERT formula?
- How do you calculate a three-point estimate?
- When should you use PERT?
Read first: Project estimation: process, methods and what to consider shows where this method fits in the whole estimation process.
Words used in this guide
| Word | Plain meaning |
|---|---|
| Optimistic (O), most likely (M), pessimistic (P) | Best plausible, typical and bad-but-realistic effort for one task |
| Expected value (E) | The weighted average used for planning: (O + 4M + P) ÷ 6 |
| Spread (σ) | A rough measure of how uncertain the task is: (P − O) ÷ 6 |
| Variance | The spread squared; variances of independent tasks can be added |
| Triangular estimate | A simpler average of the three figures: (O + M + P) ÷ 3 |
What is three-point estimating?
Three-point estimating records optimistic (O), most-likely (M) and pessimistic (P) outcomes for a defined task. Classic PERT uses the modeling convention E = (O + 4M + P) / 6. The weight of four is an assumption, not a measured probability that the task will take M hours.
A triangular model uses the mean (O + M + P) / 3. Neither mean is inherently superior: choose the distribution and bounds that reasonably represent the task’s evidence. Both are estimates under assumptions, not guarantees or confidence intervals.
Why does three-point estimating help IT project estimates?
A single most-likely-only estimate hides risk by presenting a guess as if it were a certainty. A weighted three-point estimate does two useful things at once: it produces a number that is defensible when negotiating with a client who wants "your best guess," and it makes the uncertainty itself visible and reviewable, rather than buried inside one unlabeled figure that looks precise but isn't.
When should you use PERT, and when not?
Use three-point estimating when you can explain plausible low, most-likely and high outcomes for a defined task. Routine work can still be uncertain: a calibrated historical distribution may provide stronger evidence than three guesses. A detailed task plan and uncertainty analysis can be used together. Choose the method from the available evidence and the decision, and avoid inventing ranges merely to fill a template.
How do you get good best, likely and worst-case numbers?
- Ask the person who would actually do the work, not a manager estimating on their behalf.
- Define the optimistic outcome using explicit favorable but plausible conditions. Define the pessimistic outcome with adverse conditions and exclusions; it is not necessarily the worst imaginable outcome.
- Anchor the pessimistic figure to a realistic bad day — a dependency slips, a defect surfaces — not a worst-case catastrophe that would also derail every other task.
- Sanity-check that P is greater than M and M is greater than O, and that the gap between them is proportionate to genuine uncertainty, not an arbitrary percentage padded on top of M.
PERT formula: worked example for an uncertain IT task
For an integration task, assume O = 30, M = 45 and P = 90 hours. The PERT mean is (30 + 4 × 45 + 90) / 6 = 50 hours. At USD 110 per hour, expected labor cost is USD 5,500 under the constant-rate assumption. The triangular mean is 55 hours and USD 6,050; it reflects a different distribution choice, not an arithmetic error.
The traditional approximation σ ≈ (P − O) / 6 gives 10 hours. It is a rough dispersion input under the model, not proof of normal variation or a reliable 40–60-hour completion interval. A decision requiring an 80% confidence estimate needs an explicit, justified distribution and its relevant quantile, not the mean plus an unexplained buffer.
Carry the selected effort into the resource plan, then test the adverse outcome separately. If the firm adds contingency, identify the exposure and avoid charging the same uncertainty through both padded task hours and an allowance.
Scenario: rolling three uncertain tasks into a plan
Priya, a delivery lead, has three uncertain tasks in an integration project. She asks each task owner for an optimistic (O), most-likely (M) and pessimistic (P) figure, then applies E = (O + 4M + P) ÷ 6 and an approximate spread σ = (P − O) ÷ 6.
| Task | O | M | P | Expected hours | Spread (σ) |
|---|---|---|---|---|---|
| Connect to the client's old system | 30 | 45 | 90 | 50.00 | 10.00 |
| Map the customer data | 16 | 24 | 48 | 26.67 | 5.33 |
| Build the screens | 20 | 30 | 50 | 31.67 | 5.00 |
| Total | 108.33 | about 12.4 |
- Add the expected hours: 50 + 26.67 + 31.67 = 108.33 hours. (The table shows rounded values; the unrounded expected hours are 50, 26.6667 and 31.6667, which add to 108.3333, and the cost below uses that unrounded total.)
- If the tasks are independent, add the variances (the squares of the spreads): 100 + 28.44 + 25 = 153.44; the combined spread is its square root, about 12.4 hours.
- At USD 110 an hour the expected labor cost is 108.33 × 110 = USD 11,916.67.
- Priya plans 108 hours and tells the project manager that the first task drives most of the risk: it should be investigated first.
Adding the expected hours is valid even when the tasks are related; relatedness changes the spread, not the expected total. If the old system's poor documentation hurts both the connection and the data mapping, the spread is larger than 12.4 hours. The figures are planning expectations under stated inputs, not promises, and not a confidence interval.
How does wide uncertainty differ from tight uncertainty?
Task A has O = 38, M = 40 and P = 46 hours: PERT mean 40.67 hours and approximate σ = 1.33 hours. Task B has O = 20, M = 40 and P = 120: mean 50 hours and approximate σ = 16.67 hours. The larger spread flags a different risk discussion; neither task is guaranteed to finish within one standard deviation of its mean.
Expected task costs can be summed on a consistent basis. If task costs are independent, total variance is the sum of their variances; standard deviation is the square root of that sum. Shared dependencies require covariance terms. Project duration also depends on sequencing and the critical path, so summing task means does not establish a reliable completion date.
How do you use a three-point estimate in a resource plan?
For a synthetic specialist task with optimistic 8, most-likely 16 and pessimistic 40 hours, the weighted estimate is (8 + 4 × 16 + 40) / 6 = 18.67 hours when rounded. At an illustrative USD 80 cost rate, the unrounded expected labor cost is approximately USD 1,493.33. This is a planning expectation under the stated inputs, not proof that the task will complete in that time.
Managed Margin does not run PERT for you. Review the input assumptions outside the application, then translate the selected planning effort into stage or weekly role loading and check the resulting rounded hours. Create a different strategy if the 40-hour adverse task outcome is important to the commitment; adding a ROM price range to the first strategy does not change its effort.
Sum the task with the remaining project work, add the applicable cost rules, and calculate a target fee or assess the quoted fee. Avoid adding contingency for the same uncertainty without explaining what additional exposure remains. Finalize the approved strategy and compare later task evidence with its assumptions. The resource estimation workflow turns the reviewed hours into a cost and margin model.
- Software Project ROM Estimate From a Short Brief
- How Much Contingency to Add to a Project Estimate
- Cone of Uncertainty in Project Estimation, Explained
Real example: a team lead estimates a risky task
An illustrative story with invented figures, to show how the idea works in a real office.
Imani, a team lead, is asked how long it will take to connect to a client’s old system. She asks the engineer for three numbers: best case 40 hours, most likely 60, worst case 110.
Expected effort is (40 + 4 x 60 + 110) divided by 6, which is 390 / 6 = 65 hours. The spread, (110 minus 40) divided by 6, is about 12 hours.
She plans 65 hours and tells the project manager the task could run to 110 if the old system behaves badly. That is more useful than “about 60”.
Where does three-point estimating fit in an IT project?
| Step | Owner | Checkpoint |
|---|---|---|
| Pick the uncertain tasks | Delivery lead | Only tasks where unfamiliar technology or dependencies create a wide gap; routine work is skipped |
| Collect O, M and P | The person who will do the task | P is a realistic bad day, not a catastrophe; P > M > O |
| Calculate and roll up | Presales consultant | Expected hours and spread per task; the formula is stated |
| Review | Delivery lead or peer | Inputs have evidence; related tasks are checked for shared causes |
| Feed the plan | Project manager | Expected effort goes into the resource plan; the widest tasks are scheduled for early investigation |
Three-point estimating and PERT: frequently asked questions
Why weight the most likely figure by four?
The weight of four is a modeling convention, not an observed probability. Use it only when the assumed distribution is reasonable; a triangular distribution or analysis of historical outcomes may fit better.
How do I gather good O/M/P inputs?
Ask for the best case, the worst case and the most likely, each tied to a scenario, ideally from more than one estimator.
How does it relate to contingency?
The spread shows where risk sits; use it to size a contingency allowance for the riskiest tasks.
What is three-point estimating?
Estimating a task with three numbers, optimistic, most likely and pessimistic, then combining them. The PERT formula weights the most likely figure four times: expected = (O + 4M + P) ÷ 6. For O = 30, M = 45 and P = 90 hours, expected = 50 hours.
Continue the sequence: guides to read next
- Project Estimation: Process, Methods and What to Consider: where three-point sits among methods
Sources and methodology
The calculations and scenarios are original illustrations, not customer results or market benchmarks. Product guidance was checked against the current pricing and delivery implementation on 13 October 2026.
- GAO Cost Estimating and Assessment Guide: evidence, estimating methods and uncertainty analysis.
- AACE International recommended practice 18R-97 (not linked: the freely available copy is a sample): the process-industry classification of estimates and the limits of generic accuracy ranges.
- PMI: Project schedule risk analysis: three-point models and the limits of simple schedule aggregation.