Test whether triggering-core concentration improves prospective forecasting

Determine whether conditioning earthquake forecasts on the finite-horizon triggering-core concentration regime, in addition to the fitted ETAS conditional intensity, yields measurable prospective forecast gains using out-of-sample likelihood, calibration, and information-gain tests.

Background

The finite-horizon triggering-core size is derived from the same fitted ETAS components that determine the conditional intensity, so it does not necessarily add information under a correctly specified model. Its proposed value is diagnostic: it stratifies forecasts according to whether triggering responsibility is concentrated in a few earthquakes or distributed across many sources.

The paper explicitly leaves unresolved whether this concentration-based stratification produces practical forecasting benefits. Resolving the question requires out-of-sample evaluation rather than analysis of the current catalog alone.

References

Whether conditioning on these regimes yields measurable prospective forecast gains is an empirical question that should be assessed by out-of-sample likelihood, calibration, and information-gain tests.

— Finite-Horizon Triggering Cores of Earthquake Sequences  (2609.37213 - Sornette et al., 29 Sep 2026) in Section 7, Discussion, subsection “Implications for seismicity analysis, forecasting, and limitations”