Somewhere on most capital programs, a slide says the project will finish on a given date at P50. Leadership reads that as the date. The board hears it as a commitment. The lender writes it into a covenant. But a P50 date is a coin flip. Half of the simulated futures finish later. Treating it as a promise is how a perfectly good risk analysis becomes the most expensive number on the program.
What a QSRA actually tells you
A quantitative schedule risk analysis runs the schedule thousands of times in a Monte Carlo simulation. Each run draws durations from ranges and fires risks according to their probability. The output is not a date. It is a distribution of dates, and the P-values are points on that curve.
- P50 is the date the project beats half the time. It is a fair estimate of the middle, not a target to commit to.
- P80 is the date the project beats four times out of five. Most owners who have been burned once start committing here.
- The spread between them is the real message. A narrow gap means the plan is well understood. A wide gap means the team does not yet know what it is building, how or when.
The P50 tells you where the middle is. The gap between P50 and P80 tells you how much you do not know. Leaders should spend their time on the gap.
Risk advisory lead, KonradBeck
Garbage in, confidence out
A QSRA is only as honest as the schedule and the risk register behind it. Run a Monte Carlo on a schedule full of hard constraints, open ends and missing logic, and the simulation cannot move the dates that matter. The curve comes back tight and reassuring, and completely wrong. Before any risk run, the schedule should pass a health check. If it does not, fix the schedule first.
The ranges matter too. Teams anchor on their own plan, so three-point estimates are often far too narrow. Good risk workshops ask for the bad day, not the plausible one, and challenge every range that looks suspiciously tidy.
Contingency is a price on uncertainty
The same logic applies to cost. Contingency should not be a percentage someone copied from the last job. It should be the difference between the base estimate and the confidence level the owner chooses, calculated from the risks the team has named. That gives every contingency dollar a reason to exist, and every drawdown a risk to retire.
- Choose the confidence level on purpose. A public agency with a fixed appropriation may need P80 or higher. A developer with flexible financing may accept less. Either is fine, as long as it is chosen.
- Tie contingency to risks. When a risk is retired, its share of contingency should be reviewed, not quietly absorbed into scope.
- Track drawdown against a curve. Contingency spent faster than risk is retired is an early warning, usually months before the estimate at completion moves.
Say the number, and say what it means
None of this means leaders should stop asking for a date. It means every date should carry its confidence level, out loud, every time. "Substantial completion in June at P50, September at P80" is an honest sentence. It invites the right conversation: what would it take to pull September toward June, and is it worth paying for?
That conversation is the point of risk analysis. Not a prettier chart. A better decision. If your program reports a single date with no confidence level attached, it is worth asking what is hiding behind it.
