Understanding Estimate Ranging

What estimate ranging is

Where expected value asks "what's each risk worth on average?", ranging asks "how much could each major cost item realistically vary, and what does that do to the total?" It's a step up in rigor: instead of a single expected number, you get a distribution of possible project costs, and you can pick the confidence level you want to fund to. RP 41R-08 (originally the Curran "range estimating" method) is the classic statement of this approach.

Range the critical variables, not everything

The defining idea of range estimating is selectivity. A project may have hundreds of cost line items, but only a handful actually drive the overall uncertainty. Ranging focuses effort on those critical variables — the Pareto few — and leaves the trivial many alone.

How estimate ranging works

• Identify the critical variables — Find the cost items (and risk drivers) whose uncertainty has the biggest effect on the total — by value, by uncertainty, or both.
• Assign a three-point range to each — For every critical variable, estimate a low, most likely, and high value — and crucially, the probability of not exceeding the target.
• Simulate the combined effect — Run a Monte Carlo simulation that samples each range many times and adds them up, building a distribution of total project cost.
• Read contingency at your confidence level — From the resulting distribution, pick the confidence you want (e.g. 80%) and the contingency is the gap from the base estimate to that point.

Using estimate ranging well

Use ranging when the project is large or uncertain enough that a single expected-value number isn't enough and you need to fund to a stated confidence level. Invest your effort in finding the true critical variables and ranging them honestly — wide enough to be realistic, informed by data and experience rather than optimism. Present the result as a confidence curve and let the decision-makers choose the funding level deliberately.

Nine things to remember

1. Estimate ranging puts a low–likely–high range on uncertain cost items.
2. It turns the cost from a point into a distribution of possible outcomes.
3. Range only the critical variables — the Pareto few that drive the total.
4. Ranging everything backfires — it buries drivers and can understate risk.
5. Each critical variable gets a three-point range (low, most likely, high).
6. Simulation combines the ranges into a total-cost distribution.
7. Contingency is read at a chosen confidence level (e.g. P80).
8. The output is an S-shaped confidence curve — the funding trade-off made explicit.
9. It bridges expected value and full Monte Carlo (Module 5C).