A three point duration estimate defines a credible range for an activity using optimistic, most likely and pessimistic durations. Instead of assuming that one deterministic duration will occur, the three estimates describe the uncertainty surrounding the work.
Schedulers use these values to improve duration discussions, document assumptions and provide inputs to a schedule risk analysis (SRA). However, the three points are not three separate project schedules. They are parameters that describe the possible duration of an activity under defined conditions.
The optimistic value should represent a credible short outcome. The most likely value represents the mode, or the duration considered most probable. The pessimistic value represents a credible long outcome. None of the three should be selected merely to produce a desired program finish date.
What a Three Point Duration Estimate Represents
A three-point estimate expresses uncertainty that a single duration cannot show. For example, entering 15 working days for an engineering analysis implies more precision than the team may actually have. A range of 11, 15 and 24 working days provides management with better information about the uncertainty.
The three values usually support one or both of these applications:
- Calculating a single expected duration for use in a deterministic schedule.
- Defining a probability distribution for a probabilistic schedule risk analysis.
The second application provides more insight because the analysis evaluates the duration ranges within the entire logic network. The GAO Schedule Assessment Guide describes three-point durations as one method for collecting schedule risk data and emphasizes that risk analysis should account for uncertainty across the schedule, not just on the current critical path.
Optimistic duration
The optimistic duration is the shortest credible time needed to complete the defined scope. It assumes favorable conditions, but it should not depend on an impossible level of productivity, unavailable resources or the removal of required work.
For example, the optimistic case may assume that an experienced engineer performs the work, inputs arrive when needed and the first review finds only minor comments. It should still include every step required by the activity’s completion criteria.
Most likely duration
The most likely duration is the single outcome considered most probable. It is not automatically the midpoint between the optimistic and pessimistic values.
Develop this estimate from the planned approach, resource availability, historical performance and expected working conditions. In addition, consider normal review cycles, anticipated coordination and the productivity of the assigned organization.
Pessimistic duration
The pessimistic duration is the longest credible outcome under unfavorable but plausible conditions. It may reflect lower productivity, difficult coordination, extensive comments or normal technical troubleshooting.
However, avoid turning the pessimistic estimate into an unlimited disaster scenario. If an identifiable event has a probability of occurrence below 100 percent, such as a test article failure requiring a formal retest, consider modeling it as a discrete risk rather than burying its entire effect in the pessimistic duration.
Three Point Duration Estimate Formulas
A three-point estimate does not prescribe one universal calculation. The selected formula or probability distribution should match the purpose of the analysis and the assumptions behind the inputs.
Triangular expected duration
The arithmetic mean of a triangular distribution is:
Expected duration = (O + M + P) / 3
In the formula, O is optimistic, M is most likely and P is pessimistic. This calculation gives equal weight to all three values. NASA scheduling guidance has described this simple average as one way to derive an expected duration while retaining the three inputs for probabilistic analysis in its Schedule Management Handbook.
Traditional PERT expected duration
The traditional Program Evaluation and Review Technique (PERT) formula gives more weight to the most likely estimate:
PERT expected duration = (O + 4M + P) / 6
For example, assume an activity has an optimistic duration of 8 days, a most likely duration of 12 days and a pessimistic duration of 21 days. The triangular mean is 13.67 days. The traditional PERT result is 12.83 days because the formula gives four times as much weight to the most likely value.
Microsoft provides an example of the traditional weighted calculation in its archived discussion of three-estimate project analysis. However, neither formula is automatically the correct contractual or program method. The team should document which calculation it used and why.
An expected value is not a confidence date
A formula that produces one expected duration does not perform a schedule risk analysis. It reduces a range to another single-point estimate. As a result, it does not show the probability of meeting a milestone or account for alternate critical paths.
A Monte Carlo analysis instead samples values from activity distributions during repeated iterations. Each iteration recalculates the network. Therefore, an activity on a near-critical path may become critical when sampled durations change. See how Monte Carlo schedule risk analysis works for the broader simulation process.
How to Develop Defensible Duration Ranges
Strong three-point estimates come from structured discussions, not standard percentage factors. Applying minus 10 percent and plus 20 percent to every activity may be convenient, but it rarely reflects the technical differences among engineering, procurement, software and test work.
- Confirm the activity scope. Define the work, deliverable and objective completion criteria before discussing duration.
- Identify the execution approach. Confirm the resources, calendars, work sequence, facilities, tools and planned review process.
- Determine whether the estimate covers total or remaining duration. For an in-progress activity, estimate only the remaining work. Do not apply a new range to completed effort.
- Review relevant history. Use actual durations from comparable work when available. Adjust for differences in complexity, staffing, maturity and approval requirements.
- Develop the most likely estimate. Start with the planned conditions and most probable outcome.
- Develop credible bounds. Ask what favorable conditions could shorten the work and what plausible conditions could extend it.
- Separate uncertainty from discrete risk. Identify specific risk events that should be modeled with occurrence probabilities and conditional impacts.
- Document the estimate. Record assumptions, data sources, estimating method, responsible subject matter experts and the rationale for the range.
This documentation should become part of the activity’s schedule basis of estimate. A useful schedule duration basis of estimate allows reviewers to understand why the range exists and what conditions could invalidate it.
Fictional Program Example: Qualification Testing
Consider a fictional defense electronics program preparing an environmental qualification test. The integrated master schedule (IMS) contains an activity named “Conduct Thermal-Vacuum Qualification Test.” The activity starts after test readiness approval and finishes when the approved test report confirms completion.
The test lead develops these remaining-duration estimates:
- Optimistic: 9 working days. Chamber setup proceeds without interruption, instrumentation performs correctly and the test generates few review comments.
- Most likely: 13 working days. The estimate includes normal setup adjustments, planned thermal cycles, data reduction and one comment-resolution cycle.
- Pessimistic: 20 working days. The chamber requires additional troubleshooting, data reduction takes longer and the report receives extensive comments.
The team also identifies a separate risk: a test article anomaly could require a partial rerun. The event has less than a 100 percent probability of occurring and could add 6 to 15 working days if realized. Therefore, the scheduler does not automatically add the full retest impact to the pessimistic duration. Instead, the SRA model assigns the risk to the affected test and report activities.
This distinction helps management understand two different exposures. The 9-to-20-day range represents normal duration uncertainty. Meanwhile, the anomaly represents a specific event that the program can monitor and mitigate.
Using the Estimates in Schedule Risk Analysis
Before loading ranges into an SRA tool, assess the health of the underlying network. Missing logic, unjustified constraints, excessive lags and unrealistic calendars can distort simulation results. Probabilistic software does not repair a weak deterministic schedule.
Next, select distributions that reflect the available evidence. A triangular distribution is intuitive and uses the three inputs directly. A beta-PERT distribution places more probability near the most likely value and creates smoother tails. Other distributions may be appropriate when historical data supports them.
Also consider correlation. Activities performed by the same team, supplier or facility may experience related duration changes. Treating every duration as statistically independent can understate overall schedule exposure. NASA’s program planning and control glossary identifies triangular and beta-PERT distributions as examples used to characterize uncertainty from three-point schedule parameters.
Finally, review the resulting confidence dates, criticality indices and sensitivity measures. The analysis should help management decide where mitigation, schedule margin or an alternate execution strategy provides the most value. The complete schedule risk analysis guide explains those outputs in more detail.
Application in Proposals, the IMS and EVMS
During proposal development, three-point estimating can expose aggressive assumptions before the team commits to contractual dates. It works particularly well for uncertain engineering, long-lead procurement, software integration and qualification testing.
However, the proposal scheduler should not conduct the exercise alone. Control account managers (CAMs), technical leads, supply chain personnel and test managers should provide the technical assumptions. The scheduler should challenge the logic and ensure that each estimate uses the correct working calendar and scope definition. For additional proposal context, see how to develop proposal schedule durations.
On an earned value management system (EVMS) program, a risk-adjusted forecast may show that a contractual or program milestone has a low probability of success. That result supports management action, but it does not by itself authorize a baseline change. The program must control changes to the Performance Measurement Baseline through its approved change-control process.
Similarly, do not overwrite current IMS durations with pessimistic values merely because an SRA identified exposure. The IMS should remain an executable plan. Risk analysis should inform mitigation, forecast discussion and management decisions rather than become an uncontrolled method of adding contingency to every activity.
Contractual Requirements Versus Recommended Practice
Three-point estimating is a recognized planning and risk-analysis technique. However, it is not a universal Federal Acquisition Regulation or Defense Federal Acquisition Regulation Supplement requirement for every contract or every IMS activity.
A requirement becomes contractually applicable when the solicitation or contract incorporates it through a statement of work, data item description, contract data requirements list, special clause or other binding direction. Agency guidance may also establish internal expectations, but guidance and best-practice publications do not automatically create a contractor deliverable.
Therefore, review the specific contract, applicable data requirements and customer instructions before defining the estimating process. Requirements may vary by agency, program, contract type and negotiated tailoring.
Common Three-Point Estimating Mistakes
- Using arbitrary percentage ranges. Standard factors conceal real differences in technical uncertainty.
- Making the optimistic value a crash plan. An estimate that requires unapproved overtime or unavailable resources is not credible.
- Treating the most likely value as the midpoint. The most probable outcome may sit closer to either bound.
- Using the pessimistic estimate as a risk register. This approach hides discrete risks and weakens mitigation analysis.
- Estimating summary tasks. Develop ranges at the level where scope, logic and technical ownership can be evaluated.
- Ignoring calendars. Ten working days can produce different finish dates under different workweeks and holiday calendars.
- Estimating only the deterministic critical path. Uncertainty can cause another path to become critical during simulation.
- Changing estimates to meet a target date. Compress scope, add resources or revise the execution strategy openly instead of manipulating durations.
- Failing to retain assumptions. Undocumented ranges are difficult to validate, update or defend during customer review.
Recording Three-Point Estimates in Microsoft Project
Microsoft Project schedules use the active Duration field to calculate task dates. Organizations can store optimistic, most likely and pessimistic values in custom duration fields before exporting the data to an SRA tool or applying an approved internal calculation.
For example, a scheduler can rename Duration1, Duration2 and Duration3 for the three estimates. Microsoft confirms that its Duration1 through Duration10 custom fields can store additional duration information. However, those custom values do not affect schedule calculations unless the organization implements a controlled formula, macro or external integration.
Standardize the field names, units and calculation method before collecting estimates. In addition, preserve the original inputs instead of replacing all three values with the calculated result. That approach supports traceability and later SRA updates.
Frequently Asked Questions
Does every IMS activity need three-point estimates?
Not necessarily. The required level of application depends on the purpose of the analysis, tool capability, available data and any contractual direction. However, excluding activities solely because they are not on the current critical path can understate schedule risk.
Should the baseline use the most likely or expected duration?
Programs use both approaches. The choice should follow the program’s scheduling procedure and remain consistent. More importantly, the team should document the method and retain all three estimates for risk analysis.
Can the pessimistic duration exceed a contractual milestone?
Yes. An activity range describes possible execution outcomes; it does not promise that every sampled result will remain inside the contract period. If the SRA shows inadequate confidence in a contractual date, management should evaluate mitigation and execution alternatives.
How often should the estimates be updated?
Update them when remaining scope, performance, resource assumptions or risk exposure changes materially. For in-progress work, replace total-duration ranges with defensible remaining-duration estimates and preserve actual progress.
A well-developed three point duration estimate does more than produce a formula. It records what the team knows, exposes what remains uncertain and gives management a defensible basis for evaluating schedule realism.