The backtest applies the identical FSS formula (Modification 3) and FX Stress Indicator (Modification 5) used in the current ARGOS engine to historical data. No parameters were tuned to fit past crises. The FSS formula weights are fixed: BIS credit gap (40%), IMF debt-to-GDP (35%), IMF fiscal balance (25%), with adaptive re-weighting when a component is unavailable.

Quarterly resolution advantage: Using BIS quarterly credit gap data directly (rather than annual averages) improves crisis timing detection. The GFC is detected at FSS=57.9 with 2-quarter lead time (Q1 2008, before the September Lehman collapse). The quarterly approach also captures the rapid credit gap build-up in Ireland (80.7) and Iceland (77.0) that annual data smooths away.

Remaining 4 misses: (1) Argentina 2001 - the peso peg masked FX stress until Q1 2002 when the peg broke; (2) Nigeria 2023 - no BIS EER data available; (3) Bolivia 2020 - no BIS data and IMF indicators were just below threshold (FSS=28.8); (4) Japan 1997 - a domestic banking crisis with no yen depreciation and low government debt at the time. These represent genuine epistemological limits of structural indicators.

BIS coverage gap: 43 of 87 ARGOS nations have BIS credit gap data; 60 have BIS EER data. For nations outside BIS coverage, the system relies solely on IMF indicators, reducing sensitivity but maintaining specificity.