O-39 - Cosmic-ray data imply 1e28 iron-iron collisions at AGS-equivalent energies on the Moon and 1e47 in our past light cone
These are the exposure numbers all of the paper’s empirical bounds rest on. Inputs beyond the flux data: lunar surface area (8pi^2 R^2 factor, eq. 4), lunar iron abundance f_Fe = 0.012 (Apollo samples, conservative low end of 4.2-17.2% FeO by weight), geometric cross section sigma = 0.18 A^(2/3) barn, and for the light-cone number, integration over the volume and age of the universe following Hut-Rees (whose flux inputs were subsequently confirmed). Energy scaling of every rate follows the measured power law (e.g. E^-1.7 integrated flux, E^-3.4 for two-flux collision rates).
Link to original
Why this is evidence
Not diagnostic as a bare flux fact — both members predict the same cosmic-ray record. What discriminates is the conjunction the observation makes precise: an enormous count of heavy-ion collisions capable of leaving near-rest products (lunar Fe-Fe at AGS-equivalent energies; interstellar collisions with all CM-velocity components small) has coexisted with the Moon and the stars remaining ordinary matter. Under H-29 every such collision was a trial that, on producing a stoppable negative strangelet, ends in runaway conversion of the host body — so the larger the exposure, the more improbable the ordinary-matter outcome we in fact observe. Under H-40 the exposure count carries no such tension. Direction: toward H-40, with strength scaling with the per-collision production-and-stopping probability (the rapidity-suppression and metastability escapes are the attached arguments’ subject).
Likelihood
# E-31 (HC-8) — lone edge, one observation O-39 "cosmic-ray exposure counts: ~1e28 lunar Fe-Fe at
# AGS-equivalent energies, interstellar at-rest Fe-Fe rate, ~1e47 in the past light cone". Members in
# HC-8.hypotheses order [H-29 runaway, H-40 no-runaway], no residual. As a bare flux fact O-39 is
# non-diagnostic (both members predict the same cosmic-ray record); it discriminates only as the TRIAL
# COUNT underlying the persistence bounds. Anchor H-40 = 1 (rule 7). Deliberately WEAK to avoid
# triple-counting: the persistence force is already priced through E-30 (lunar) and E-32 (stellar) —
# this edge's honest own increment is only that N is enormous and well-measured, plus the light-cone scale.
lik_counts_H40 = 1.0 # anchor: the exposure record is exactly as observed under no-runaway
lik_counts_H29 = 0.5 # 0.5x: under runaway a larger measured N sharpens how improbable the observed
# ordinary-matter outcome is (A-59 lunar p<2e-11; A-60 approved, interstellar->stellar
# supernova bound 2e-8 tightening to ~1e-21). But those bounds' persistence content
# lives on O-23 / O-22 (priced in E-30 / E-32); O-39 supplies only the count, so its
# marginal HC-8 decrement beyond them is modest — kept near 1. A-55 (corrected) is
# light-cone VACUUM decay (O-40), a different cluster, so it adds nothing to HC-8 here.
t_counts = 0.75 # cap = trust_score(O-39 source S-3) = 0.85; docked to 0.75 because O-39 is a DERIVED
# quantity — fluxes folded with lunar Fe abundance (Apollo f_Fe=0.012), geometric
# cross sections, and Hut-Rees light-cone integration — a pipeline with real
# extrapolation, though order-of-magnitude and conservative
evidence("HC-8", ["O-39"], [lik_counts_H29, lik_counts_H40], t=t_counts)