Daily gamma rate variation with variance 0.05 raises a Ridgecrest day 1-30 count ratio tail fraction to 0.144

We tested whether a fixed mixture of zero-offset power-law rates can explain faster late decay in a fixed USGS Ridgecrest catalog. In three equal log-time bins, any such positive mixture requires the ratio of mean counts mu1 squared divided by mu0 times mu2 to be at most one. At preferred magnitude at least 2.5 within the source radius of 100 km, days 1-30 have counts 283, 342, and 306. The observed ratio is 1.3507. An exact conditional test gives p=0.01704 under independent Poisson counts. The 95% conditional interval for the mean ratio is 1.0222 to 1.7740. Smaller radius, one magnitude type, and a later start weaken the Poisson evidence. A labeled sensitivity experiment, added after initial results, keeps a fitted single-power mean curve and adds daily gamma rate multipliers with mean one. With variance 0.05, the simulated tail fraction is 0.1444, compared with 0.0147 at variance zero. Their Monte Carlo standard errors are 0.00157 and 0.00054. These are plug-in model diagnostics. The daily variance is assumed, not estimated. The deterministic mixture bound is useful, but the catalog does not securely reject this mean-rate explanation once the tested daily intensity variation is allowed. One sequence, retrospective selection, mixed magnitudes, uncertain detection, and restricted dependence models limit inference. The test does not identify a physical cause or reject secondary triggering. The package includes proofs, exact source fields and hashes, code, all planned checks and simulation counts, figures, an independent arithmetic audit, and an offline repeat.

Ridgecrest · aftershock decay · power-law mixture · log convexity · exact conditional test · Poisson assumptions · gamma-Poisson sensitivity · catalog completeness · null result · reproducibility