HKSM

Key Points

• Uses optimistic (O), most likely (M), and pessimistic (P) values to model uncertainty.
• Common formulas: triangular mean (O + M + P) / 3 and beta/PERT mean (O + 4M + P) / 6.
• Standard deviation often approximated as (P - O) / 6 for beta/PERT to express variability.
• Applies to duration, cost, or effort at activity, work package, or feature level.
• Improves stakeholder communication by providing an expected value with a range and confidence.
• Relies on expert judgment and historical data; assumptions must be documented.

When to Use

• When scope is understood but uncertainty in effort, cost, or duration remains.
• When risks could realistically push outcomes higher or lower than the most likely value.
• When stakeholders need a range and confidence level rather than a single point.
• During early planning to set realistic expectations and during re-estimation as new data emerges.
• When combining with quantitative risk analysis or Monte Carlo simulations.

How to Estimate

• Define the item to estimate and the unit of measure (e.g., days, hours, currency).
• Elicit O, M, and P from qualified estimators, using historical data and clear assumptions.
• Select a distribution: triangular for simple averaging; beta/PERT for a weighted mean toward M.
• Calculate expected value: triangular = (O + M + P) / 3; beta/PERT = (O + 4M + P) / 6.
• Estimate variability: standard deviation ≈ (P - O) / 6 (beta/PERT); compute variance as SD².
• Derive a range or confidence interval (e.g., expected ± 1 or 2 standard deviations).
• Review with the team, validate assumptions, and adjust for known correlations or constraints.
• Document the basis of estimate and update the assumptions log and risk register.

Inputs Needed

• Defined scope, WBS, and activity or work package descriptions.
• Historical data or reference class data for similar work.
• Expert judgment from people who will perform or have performed the work.
• Risk register entries relevant to the item being estimated.
• Resource calendars, availability, and productivity assumptions.
• Organizational estimating policies, units, and calendars.
• Assumptions and constraints that affect optimistic or pessimistic conditions.

Outputs Produced

• Expected value estimate for duration or cost.
• Range expressed via optimistic and pessimistic bounds and, optionally, a confidence interval.
• Standard deviation and variance for the estimate, where applicable.
• Basis of estimates documenting data sources, assumptions, and methods used.
• Updates to the assumptions log and risk register based on insights from O and P scenarios.

Assumptions

• Optimistic and pessimistic values are feasible and reflect realistic conditions.
• The chosen distribution shape reasonably represents uncertainty for the item.
• Units, calendars, and productivity rates are consistent across O, M, and P.
• Correlations with other estimates are recognized and handled during aggregation.
• O and P consider identified risks and opportunities but exclude unknown unknowns.

Example

You are estimating the duration of an activity with O = 8 days, M = 10 days, and P = 16 days.

• Triangular mean = (8 + 10 + 16) / 3 = 11.33 days.
• Beta/PERT mean = (8 + 4 × 10 + 16) / 6 = 10.67 days.
• Standard deviation (beta/PERT) ≈ (16 - 8) / 6 = 1.33 days; variance ≈ 1.78 days².
• Approximate 95% range ≈ 10.67 ± 2 × 1.33 = 8.01 to 13.33 days.

Report the expected value with the range and document the assumptions behind O and P.

Pitfalls

• Using unrealistic O or P values that are not actually achievable.
• Anchoring on the most likely value and ignoring the calculated range.
• Mixing units or calendars across estimates, which distorts results.
• Ignoring dependencies and correlations when aggregating multiple estimates.
• Overstating precision by reporting too many decimal places.
• Failing to update estimates as new information and performance data emerge.

PMP Example Question

A team wants to reflect uncertainty in activity duration estimates and provide stakeholders with an expected value and a realistic range. Which technique should the project manager recommend?

1. Analogous estimating
2. Bottom-up estimating
3. Three-point estimating
4. Parametric estimating

Correct Answer: C — Three-point estimating.

Explanation: Three-point estimating uses optimistic, most likely, and pessimistic values to produce an expected value and a range, making uncertainty explicit.