A project plan ends on one date, but several activities still have uncertain durations. Presenting that finish as an exact prediction hides the uncertainty rather than resolving it. Monte Carlo schedule analysis explores many possible combinations of those durations within a defined model.
The US Government Accountability Office’s Schedule Assessment Guide treats schedule risk analysis as part of building a reliable plan. The activity network, duration estimates and risks need credible foundations before a simulation can produce useful information.
The model runs many versions of the plan
Instead of assigning only one duration to every uncertain activity, the analysis uses specified ranges or distributions and repeatedly calculates the resulting schedule. Dependencies determine how delays travel through the work. The output can show a spread of completion dates rather than a single answer.
For an invented example, imagine a design review followed by a supplier build, while permit work runs in parallel. The project cannot proceed until both relevant paths are ready. Merely adding the average duration of a few tasks misses how the paths interact.
Ask what a percentile actually means
If a report labels a date P80, it generally means 80% of the modelled outcomes finish by that date under the stated assumptions. It is not an unconditional promise that reality has an 80% chance of obeying the model. Missing work, optimistic inputs or overlooked relationships can make the output misleading.
Ask which uncertainties were included, who supplied the estimates and how shared risks were represented. Two activities affected by the same supplier problem should not casually be treated as unrelated.
Use the result to improve a decision
Compare possible responses: earlier review, alternative sequencing, additional contingency or a different commitment date. Keep the assumptions and analysis date with the chart so a later team can understand what changed.
A simulation is most useful when it changes the questions asked about the plan. Thousands of calculated scenarios do not compensate for a schedule that leaves out a necessary approval. Validate the work and dependencies first, then use the range of outcomes to discuss a commitment honestly.