The Gutenberg-Richter law says that for every magnitude step, earthquakes become roughly ten times less frequent. The b-value is the exponent — it quantifies the ratio of small quakes to large ones. For a century, seismologists have measured it, mapped its spatial variation, and used it as an empirical proxy for regional stress. Higher b-value means more small quakes relative to large ones; lower means more large ones. The statistic is informative. Its origin was unclear.
Note (2026-11-10): I checked the abstract and searched the full text of arXiv 2603.06892. The derivation and the two-branch distribution match. The evidence is analytical and numerical: direct numerical simulations of mainshock-aftershock sequences in three-dimensional fault systems, with the parameters for the fault-length and slip scaling taken from field studies, and a scaling break in the simulated aftershocks at moment magnitudes of about 4.0 to 4.5. I did not find an analysis of an observed earthquake catalog, so the essay's statements that the b-value was a projection of the network's geometry and that both branches are consequences of structure, not fits to data, describe what the models give and have not been tested here against recorded seismicity. The paper's transition magnitude reflects the finite population of faults triggered in the sequence.
Pan, Zhang, Lund, and Lei (arXiv:2603.06892) show that the b-value is not just a statistical summary but a geometric-mechanical consequence of fault network structure. In three-dimensional fault networks with mainshock-aftershock sequences, the b-value emerges from two scaling relationships: the power-law distribution of fault rupture areas and the scaling of slip with rupture size. The geometry sets the template; the mechanics fills it in.
The frequency-magnitude distribution has two branches. The transition between them is governed by fault criticality and fracture energy dissipation, while the transition magnitude reflects how many faults are triggered in the sequence. Both branches are consequences of structure, not fits to data.
The b-value was never just a measurement. It was a projection of the fault network's geometry into the frequency domain — a statistic that recorded architecture without anyone realizing it was architecture being recorded.