Answers are grounded in this paper and cite their sources.
Fire-tenability surrogate errors matter when they change an evacuation decision. This study links sampled available safe egress time (ASET), missed unsafe classifications and finite-horizon failure probability. It combines an 18-case Fire Dynamics Simulator (FDS) library, 5000 surrogate scenarios, matched statistical experiments and 48 additional fires evaluated with a frozen predictor. A common-horizon formulation keeps missed-failure loss observable despite censored ASET. In the retrospective calculation, shifting censored endpoints explained 6.24 of a 6.38-percentage-point increase in flags after a 20 s deduction. Visibility governed all 32 observed breaches in the additional-fire batch; 25 predictions were late, with mean absolute error 7.97 s across breached cases. A 20 s deduction removed 0.0057 percentage points of joint misses under full-descent timing and 0.2009 points under corridor-segment timing, while adding 6.1551 and 1.1948 points of false-unsafe classifications, respectively. Under equal door weights, the corridor changes became 0.3947 and 1.8542 points. These are classification changes, not reductions in physical fire risk. The contribution is an auditable assessment of censoring, scenario support and decision cost. Numerical findings remain conditional on one geometry, prescribed fires and assumed evacuation laws; no universal deduction is established.
No takes yet. Share an insight, caveat, or question.
Tan et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: