What the analysis says
The cluster asks: granting the extreme premise of a non-evaporating hole captured inside Earth, what follows? Three members — H-8 - An LHC-produced metastable black hole could accrete at the Eddington limit and become catastrophic (catastrophic, Plaga’s Eddington-radiation scenario), H-22 - Even hypothetically stable LHC black holes pose no risk to Earth on timescales below the Earth’s natural lifetime (Giddings–Mangano safe), and H-36 - A stable trapped black hole would lead to some outcome not listed here (residual intermediate). The prior, deliberately built from Earth-accretion physics alone, sat at [0.2727, 0.5795, 0.1477] — H-8 held a real 0.27 because Earth-internal calculation leaves fast-accretion windows open (A-48 - For crossover radius above about 200 Angstrom Earth accretion could be fast, so Earth-internal calculation alone cannot close the safety case). Five evidence blocks — the compact-star survival record — then collapse it to [0.0072, 0.9464, 0.0463]. The movers are the “clean” astronomical edges: white-dwarf survival E-39 - O-33 × HC-4 — old low-field massive white dwarfs survive (lik_H8 = 0.1, t = 0.72) and neutron-star survival CG-1 - HC-4 joint over O-34+O-2 (lik_H8 = 0.12, t = 0.82), reinforced by E-42 - O-22 × HC-4 — exposed astronomical bodies persist with no anomalies (0.15). Earth/Sun survival E-43 - O-35 × HC-4 — Earth and Sun themselves have survived is deliberately weak (lik_H8 = 0.55) and the torsion-balance window E-44 - O-7 × HC-4 — torsion-balance window leaves the catastrophic warping range open mildly favours H-8 but is nearly inert (t = 0.3). This is the empirical closure of the safety case: physics alone leaves danger live, the survival of dense stars is what puts it to rest.
What the model may not capture
The entire 0.27→0.007 collapse flows through one argument chain — A-49 - Solar-mass white dwarfs provide enough column density to stop cosmic-ray-produced neutral black holes for D up to 7, A-51 - White-dwarf survival excludes every scenario in which Earth accretion would be dangerous - the accretion-time ratio is about 2e4, A-52 - Neutron-star survival bounds the D of 8 and higher scenarios via binary beam dumps and cosmogenic neutrinos, A-53 - No Eddington limit throttles micro-black-hole accretion in Earth, white dwarfs, or neutron stars — all from a single paper (S-37), each approved as “trusted/checked” but explicitly not retraceable (“a specialist computation not retraceable here”). The one independent corroborator, Koch–Bleicher–Stöcker, sits in sources/non-curated (S-47). So the load-bearing empirical result of the whole KB is one un-replicated calculation the curated set does not independently verify. Beneath it, the accretion law itself (Bondi + subatomic Coulomb capture) is G–M’s own, with no prior primary (flagged at step 1, HC-5) — both named members’ timescales inherit it.
The carving is also questionable at the corner that matters. The exclusion is a relay: A-49 covers D≤7 (the D=7 sub-case leaning on Earth-time, not the star), A-52 covers D≥8 but by its own statement fails “unless cosmic rays are almost purely heavy AND the cosmogenic neutrino flux is absent or non-reactive.” Add the sub-M_min quantum-stopping gap (A-8 - The white-dwarf stopping argument presumes semiclassical validity and fails for black holes below three times the Planck mass scale). In that named corner neither empirical bound bites — and a hole that evades stopping then accretes dangerously in Earth is an H-8 outcome, not an intermediate one. Yet the model routes surviving danger mostly into H-36 (posterior 0.046) over H-8 (0.007). My read: the residual is absorbing catastrophe-flavoured mass that belongs to H-8, and the 0.007 headline reads “put to rest” over a parameter space where a conditional, un-closeable corner still exists. Is the true answer on the list? For most of parameter space yes; but the leaked corner is worse than any listed member and its risk there is not 0.007.
What would help
- An independent re-derivation of the G–M stopping/accretion chain (A-49–A-53) — exists, uncurated (S-47, plus any post-2009 replication). 2. A first-principles primary for the Coulomb-capture/Bondi accretion law — unclear (possibly does not exist). 3. Direct measurement of the GZK cosmogenic-neutrino flux, which would close A-52’s escape hatch — exists, unread (IceCube/Auber limits post-2008). 4. Cross-section behaviour of sub-M_min quantum-regime holes in dense matter (A-8) — does not exist.
Confusions and contradictions
E-42 (O-22) is scored as an independent lone edge at lik_H8 = 0.15, but its discriminating content is largely the same white-dwarf and neutron-star survival already priced in E-39 and CG-1; the block defends only “absence of an anomalous BH population” as incremental, yet its value sits right at E-39’s 0.1 — suggesting the shared survival fact is being counted twice, sharpening the H-8 collapse beyond what the independent evidence supports. Separately, E-44’s t = 0.3 is a provenance artifact: O-7 is a rock-solid 95%-CL torsion-balance limit (Kapner et al. 2007), docked to 0.3 only because it reaches the model via Plaga (S-30). The block itself concedes “no measurement/selection weakness.” The asymmetry is uncomfortable — the sole H-8-supporting edge is throttled by who cited it, while the danger-excluding G–M chain carries t ≈ 0.82.
External consensus
LSAG/CERN, the 2011 LHC Safety review, and the mainstream treat this question as closed exactly on the H-22 line — the posterior 0.946 matches the field. The divergence worth flagging is not the landing point but the confidence: the consensus rests, as this cluster does, on Giddings–Mangano being right, and no independent group re-did the compact-star accretion integration. The analysis lands where the field lands, for the same single-source reason.