HKSM

Key Points

• Uses a model and probability distributions to test thousands of possible outcomes.
• Commonly applied with Monte Carlo for schedule and cost risk analysis.
• Requires sound input data, realistic assumptions, and attention to correlations.
• Produces ranges, percentiles (for example, P50, P80), and probabilities of meeting targets.
• Supports decisions on contingency, buffers, and risk response effectiveness.
• Results are only as good as the model and inputs; document assumptions for transparency.

Purpose of Analysis

To quantify the impact of uncertainty and variability on project outcomes and to estimate the likelihood of hitting schedule or cost goals. Simulations turn uncertain inputs into probability-based outputs that inform realistic plans. They guide the setting of reserves, selection of risk responses, and communication of confidence levels to stakeholders.

Method Steps

• Define the model: scope of analysis, key outputs (schedule, cost, performance), and boundaries.
• Identify uncertain inputs and assign appropriate probability distributions (for example, triangular, normal, beta-PERT).
• Capture dependencies and correlations between inputs where they logically exist.
• Select the number of iterations and run the simulation (for example, Monte Carlo).
• Review outputs: histograms, cumulative distributions, percentiles, and sensitivity results.
• Validate results with experts, adjust assumptions if needed, and document findings and decisions.

Inputs Needed

• Deterministic model structure (for example, cost breakdown, schedule network, or performance model).
• Three-point estimates or statistical parameters for uncertain inputs.
• Assumptions and constraints, including calendars, resource limits, and funding caps.
• Risk events with probability and impact ranges, including potential opportunities.
• Correlation assumptions and dependency logic between variables and activities.
• Historical data or benchmarks to calibrate distributions and validate realism.

Outputs Produced

• Probability distributions of total cost, finish dates, or performance metrics.
• Percentile values (for example, P50, P80) and confidence levels for targets.
• Recommended contingency or management reserve based on risk exposure.
• Sensitivity indicators showing which inputs drive the most variation.
• Scenario comparisons (baseline vs. mitigated cases) to evaluate risk responses.
• Visuals such as histograms and cumulative S-curves for stakeholder communication.

Interpretation Tips

• Focus on ranges and percentiles, not single-point estimates.
• Check whether key percentile values stabilize when increasing iterations.
• Confirm that assumptions and correlations reflect how the real system behaves.
• Use sensitivity results to target the few drivers that matter most.
• Compare unmitigated vs. mitigated simulations to quantify response benefits.
• Communicate confidence levels clearly (for example, P80 is not a guarantee).

Example

A team models project duration by assigning three-point estimates to critical activities and correlating tasks affected by the same supplier risk. Running 10,000 iterations yields a P50 finish in 24 weeks and a P80 finish in 27 weeks. The team commits to a 27-week target with a buffer and plans responses for the top two drivers from the sensitivity results.

Pitfalls

• Using arbitrary distributions without data or expert validation.
• Ignoring correlations, leading to underestimation of overall risk.
• Too few iterations, producing unstable or misleading percentiles.
• Overcomplicating the model beyond available data quality.
• Presenting results with false precision or as guarantees.
• Failure to document and review assumptions with stakeholders.

PMP Example Question

A project manager needs to set a realistic cost contingency based on uncertain estimates and risk events. Which technique best provides a probability-based range for total project cost?

1. Analogous estimating.
2. Deterministic critical path analysis.
3. Monte Carlo simulation.
4. Three-point estimating only.

Correct Answer: C — Monte Carlo simulation.

Explanation: Simulation models uncertainty across many iterations to produce probability distributions and percentiles for total cost. Three-point estimating alone does not quantify overall project-level probability.