Deferral-fraction frontier ========================== The 20% scope target was chosen as an example. This frontier sweeps risk-ranked first-year deferral from 0% to 25% and reports the trade between calendar completion, fixed-future residual risk, recovery, and five-year cost. Completion uses the assumed 75/105-day weather/outage calendars. Recovery uses fixed future factors; it is a screening frontier, not out-of-sample validation for every point. Results ------- .. list-table:: :header-rows: 1 * - Target - Actual deferred - Completion 75 - Completion 105 - Fixed final loss - Recovery final loss - Reopt cost * - 0% - 0.0% - 13.2% - 61.9% - 0.0000 - 0.000003 - 298,723 * - 5% - 6.0% - 19.7% - 67.7% - 0.0109 - 0.000003 - 295,027 * - 10% - 11.1% - 27.9% - 70.0% - 0.0209 - 0.000003 - 292,078 * - 15% - 16.1% - 39.3% - 79.7% - 0.0307 - 0.000000 - 289,012 * - 20% - 21.3% - 47.1% - 87.4% - 0.0413 - 0.000003 - 285,825 * - 25% - 26.0% - 53.3% - 91.9% - 0.0509 - 0.000000 - 283,082 All targets are first-year budget-feasible in 100% of scenarios. Higher deferral improves calendar completion and lowers five-year reoptimized cost, but increases fixed-future residual risk. Under fixed-factor recovery, the final risk gap is essentially closed at every tested fraction. Limits ------ - Fixed-factor recovery is a screening result; only the 20% point has out-of-sample future validation. - Risk ranking is first-year only and whole actions are deferred; actual fractions overshoot. - No single utility or risk appetite is optimized; the table is a frontier, not a recommendation. - Weather and outage calendars are prototype assumptions. - No recurrence, safety, asset-owner priority, or field calibration.