Interpret percentiles and set a contingency policy

For percentile q, the model reports the empirical outturn Pq. The raw modeled delta is Pq(outturn) − unrisked baseline; required contingency or reserve is the larger of zero and that delta. Percentage of baseline is undefined when the baseline is zero. Existing cost contingency and schedule reserve are shown separately as a gap or surplus and are never added into the baseline before simulation.

P50 is the median modeled outturn, not necessarily the most likely single value. P80 means 80 percent of simulated outcomes are at or below the displayed value under the current inputs; it does not mean an 80 percent real-world promise. Higher cost or duration percentiles are worse outturns. Confidence at an approved budget or deadline is the empirical fraction at or below that human-entered boundary.

Choose policy explicitly

The target percentile reflects governance, risk appetite, funding flexibility, consequences of overrun, portfolio context, and stage of maturity. The model must not choose it silently. Cost-P80 and schedule-P80 are marginal measures; calculate joint budget-and-deadline confidence directly when integrated data and both boundaries exist. Never infer joint confidence by multiplying or equating separate percentiles.

Distinguish estimate contingency from escalation, known scope allowances, management reserve, contractual risk allocation, and organizational reserve. A percentile supports deliberation; only the authorized decision owner can adopt and control funds or time. Continue to decision review and governance. ## Interpretation discipline

Keep inputs, derived results, and human decisions visibly separate. Every material number needs linked evidence or an explicit assumption flag and basis. Unknown is not zero. A percentile is conditional on this model, scenario, seed, engine version, and source version; it is neither a guarantee nor an approval. Any material model edit makes an earlier result stale.