{"id":"92b4f436-513d-4eb9-891d-18ff025c0d41","arxiv_id":"2504.13380","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"high","formal_verification":"none","parameter_count":5,"one_line_summary":"Deep LTB void models with two or three distance-dependent jumps in the supernova absolute magnitude fit Pantheon+ and Planck data well and push the inferred local Hubble constant toward the SH0ES value.","lead":"This paper claims that a large, deep local void plus step-like jumps in the supernova absolute magnitude can resolve the Hubble tension. The authors fit these models to Pantheon+ supernova and Planck CMB data and report strong model-selection support.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claimed strong preference for ΛLTB is not established: Tables III/V compare against ΛCDM baselines with a single constant M, so the reported ΔBIC/ΔAIC/lnB may be driven by the added M-step flexibility rather than by the void.","rationale":"The reader's weakest assumption is exactly the load-bearing concern I find: the fiducial baselines lack the M-step flexibility that the ΛLTB models are given, so the model-selection statistics in Tables III and V do not isolate the contribution of the void. The paper is internally coherent and transparent, and it acknowledges the over-fitting risk in Sec. V, but it does not close the gap. The proposed check is computational and inexpensive given the existing pipeline: refit ΛCDM with the same step structure and compare information criteria. If the preference survives that fair comparison, the central claim would be substantially strengthened; if not, the conclusion would need to be weakened to 'ΛLTB with M steps is preferred over constant-M ΛCDM,' which is not the advertised claim. I therefore do not change the reader's CONDITIONAL verdict; the concern confirms that conditionality rather than undermining it.","tokens_in":24604,"tokens_out":6556,"duration_ms":65304,"concrete_test":"Fit flat ΛCDM with the same two- and three-step M transitions of Eqs. (26)–(27), with free H0, Ωm, M0...M3 and free dcr1, dcr2, dcr3, to the same Pantheon+ likelihood (Eqs. 22–24) and the same CMB distance-prior likelihood (Eqs. 28–31), using the same Cobaya/GetDist pipeline. Recompute the Table V columns for ΛLTB3M and ΛLTB4M relative to this ΛCDM+M-steps baseline. If the ΔBIC values remain below about -10, the void's preference is genuine; if they fall to -6..0 or become positive, the M-steps, not the void, carry the preference. A minimal version appropriate for the ΛLTB2M comparison is the same test with one M step and Planck-fixed H0 and Ωm, matching ΛCDM P18 with M flexibility.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The model-comparison claim rests on an asymmetry in model complexity. In Sec. III B and Tables III and V, the ΛLTB2M/3M/4M models are compared only against ΛCDM F and ΛCDM P18, both with a single constant absolute magnitude M. But the M-step structure of Eqs. (25)–(27) is itself a substantial addition: three extra magnitudes and three fitted step locations, motivated in the introduction by [63,64]. Since no ΛCDM baseline with the same M-step freedom is fitted, the reported Δχ², ΔBIC, ΔAIC and lnB values conflate the benefit of the void with the benefit of the extra M flexibility. The paper states in Sec. V that the risk of over-fitting is quantified by AIC/BIC, but those criteria are computed relative to baselines that do not carry the same flexibility, so the quantification is incomplete. In the strongest case, Table V, z<2.0, ΛLTB3M vs ΛCDM F, Δχ²=-71.4 and ΔBIC=-34.2; the BIC penalty difference is only about 5×ln(1700)≈37.2. A ΛCDM model with two M transitions has essentially the same number of free parameters as ΛLTB3M and could plausibly recover a large part of the χ² improvement, since the M steps directly target the low-redshift SNIa residuals. If it recovers more than roughly 64 units of χ², the reported 'very strong' preference reverses. The paper explicitly cites [63,64] showing that such M transitions are favored in ΛCDM, which makes the missing baseline a direct internal tension rather than an outside-consensus objection. This does not prove the void model wrong; it means the central preference claim is not yet supported by the presented model selection.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper tests Lemaitre-Tolman-Bondi void models with a cosmological constant (ΛLTB), supplemented by piecewise-constant transitions in the Type Ia supernova absolute magnitude M, against Pantheon+ SNIa data alone and combined with Planck 2018 distance priors. The authors report that the ΛLTB models with two, three, or four M segments yield a central expansion rate H0(r=0) near 73 km/s/Mpc while matching the Planck value H0,out = 67.36 km/s/Mpc outside the void, and they claim strong or very strong preference over flat ΛCDM baselines in terms of Δχ², ΔAIC, ΔBIC, and Bayes factors (Tables III and V). The analysis is built on MCMC fits using the Cobaya/GetDist pipeline, with constraints reported in several figures and tables.","tokens_in":25007,"tokens_out":4721,"duration_ms":45547,"significance":"If the model-comparison claim were robust, the paper would offer a notable phenomenological resolution of the Hubble tension: a single framework in which local and CMB determinations of H0 agree within 2σ while retaining strong statistical preference over ΛCDM. The paper is also useful in that it confirms and extends the earlier finding of an M transition at ~20 Mpc to additional transitions near 129 Mpc and 860–960 Mpc, and it makes the fitted void parameters and information criteria available in tabular form. However, the central preference claim rests on a comparison in which the ΛLTB models carry the extra flexibility of the M-step structure while the ΛCDM baselines do not, and the SN-only comparisons impose external Planck values without a corresponding CMB likelihood. These issues are load-bearing, so the significance of the claimed result is not yet established as stated.","major_comments":[{"comment":"The claimed strong preference for the ΛLTB models is computed against ΛCDM baselines with a single constant absolute magnitude M, while the ΛLTB2M/3M/4M models of Eqs. (25)–(27) include extra magnitudes and freely fitted step locations. The paper itself cites Refs. [63,64] showing that such M transitions are favored in ΛCDM alone, so the comparison conflates the benefit of the void with the benefit of the extra M flexibility. A concrete test is to fit ΛCDM with the same M-step structure (one, two, or three transitions) and compare ΔBIC/ΔAIC/lnB against the ΛLTB models; in the strongest case in Table V (z<2.0, ΛLTB3M vs ΛCDM F), Δχ² = -71.4 and ΔBIC = -34.2, and a ΛCDM model with two M transitions has essentially the same number of free parameters and could plausibly recover a large part of the χ² improvement, reversing the reported preference. Because this missing baseline is directly relevant to the central claim, the model-comparison statistics in Tables III and V do not currently establish that the void, rather than the M-step flexibility, is what the data prefer.","section":"Sec. III B and Tables III and V"},{"comment":"The SN-only fits in Table III impose the Planck values Ωm,out = 0.3153 and H0,out = 67.36 km/s/Mpc on the ΛLTB models without including any CMB likelihood in the fit. This is acknowledged in Sec. IV, where the authors state that it is 'unfair to impose these conditions without using the CMB data' and that when these conditions are imposed, the CMB data must also be taken into account jointly. Yet the paper still presents the SN-only comparisons of Table III as evidence, including a 'very strong' ΔBIC_P18 claim. Since the ΛCDM F baseline fits H0 and Ωm freely, the comparison embeds external Planck information into one model but not the other, and the information criteria do not account for this. The SN-only claims should either be removed, or the SN-only fits of the ΛLTB models should be repeated with H0,out and Ωm,out as free parameters.","section":"Sec. III and Sec. IV"},{"comment":"The central output H0,in ≈ 73 km/s/Mpc is not an independent prediction in the sense implied by the abstract: the Pantheon+ likelihood used in Eq. (22) includes the 77 SH0ES Cepheid-calibrated host-galaxy distance moduli in the first line of Eq. (23), which directly anchor the local distance ladder. The void depth δV and the M transitions are fitted to the same likelihood, so the agreement of H0(r=0) with the SH0ES value is partly a consequence of the calibration embedded in the data rather than a falsifiable prediction of the void model. To assess how much of the 'alleviation' is driven by this anchoring, the authors should show results with the Cepheid-host term removed (i.e., using only the SNIa distance moduli with M marginalized or fitted), or with the SH0ES anchors treated as a separate dataset whose consistency with the void model is explicitly tested.","section":"Sec. III, Eq. (23), and Figs. 4 and 8"}],"minor_comments":[{"comment":"The text in the conclusion acknowledges the increased model complexity and the risk of over-fitting, but the statement that this risk is 'quantified by the Bayesian evidence, AIC and BIC' is only accurate for comparisons against baselines that carry the same M-step flexibility; the quantified evidence in Tables III and V does not include such baselines.","section":"Sec. V"},{"comment":"The label 'SH0SE' in the figure captions appears to be a typo for 'SH0ES'; this should be corrected.","section":"Fig. 4 and Fig. 8 captions"},{"comment":"The entries for rV and Δr in Table II are reported as lower or upper limits (e.g., '> 5403.66' and 'none'), which is appropriate, but the text in Sec. III states that the constraints on the free parameters are 'fairly tight (except rV and Δr)'; it would be clearer to state explicitly that two of the six or seven void parameters are effectively unconstrained and that the reported H0(r=0) values therefore depend on the assumed prior ranges for these parameters.","section":"Sec. III, Table II"},{"comment":"The distance-prior covariance matrix is given to high precision, but the paper does not state how the ΛLTB luminosity distance is used to compute dA(z*) for the shift parameter R in Eq. (30); since the light path to z* passes through the void, a brief clarification of whether R is evaluated with the full ΛLTB dL(z*) or with the background ΛCDM expression would improve reproducibility.","section":"Sec. IV, Eq. (29)"}],"recommendation":"major_revision","confidential_remarks":"The paper contains an internal acknowledgment in Sec. IV that the SN-only comparison of Sec. III is 'unfair' because it imposes Planck boundary values without a CMB likelihood; this self-identified inconsistency, together with the absence of a ΛCDM baseline with the same M-step flexibility, is the core of my major concerns. The manuscript should not be accepted in its current form, but the issues are addressable within the paper's scope by redoing the model comparison with symmetric baselines and by either removing or reframing the SN-only claims."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe thing to know: this paper does not establish what the abstract claims. It fits ΛLTB void models with two, three, or four step-like SNIa absolute magnitudes to Pantheon+ (and Pantheon+ plus Planck distance priors), and reports H0(r=0) near 73 km/s/Mpc with 'very strong' BIC and Bayes-factor preference over ΛCDM. The catch: the ΛCDM baselines are fit with a single constant M. The M-steps at ~20, ~129, and ~860–960 Mpc are fitted freely and are motivated by earlier papers showing such steps are favored even in ΛCDM. So the reported Δχ², ΔAIC, ΔBIC, and lnB conflate the benefit of the void with the benefit of the added M flexibility. A ΛCDM model with the same step structure has roughly the same number of parameters as ΛLTB3M or ΛLTB4M and could plausibly absorb much of the improvement. No such baseline is presented. That is the load-bearing issue.\n\nWhat the paper does well: the LTB machinery is standard and clearly presented; the MCMC details are reproducible; they compute AIC, BIC, and Bayes factors, and report χ². They independently reproduce the known ~20 Mpc M transition and identify two new candidate transition distances (~129 Mpc and ~860–960 Mpc). The authors are also transparent about over-fitting risk—the final version explicitly mentions a referee's concern and attempts to address it with model-selection criteria.\n\nThe soft spots beyond the missing baseline: (1) the central H0 value is not an independent prediction—the void depth is fit to Pantheon+, which includes the SH0ES Cepheid host distances that anchor the local distance ladder. (2) The deep void (δV ~ −0.4 to −0.5) needs external checks (kSZ, ISW, full CMB); the paper cites a loose cluster constraint but not the strongest probes. (3) The BIC comparisons penalize the ΛLTB models for their extra parameters but not the ΛCDM baselines for the flexibility they are denied.\n\nWho is it for: cosmologists working on the Hubble tension or SNIa absolute-magnitude systematics. It is a well-documented attempt that needs a proper control, not a definitive answer. A serious referee should send it back with a request for a ΛCDM + M-step baseline and external void constraints.\n\nMy vote: conditional at best, but the paper deserves peer review rather than desk rejection—the question is timely and the analysis is transparent enough to be fixed.\n\nBest,","headline":"The claimed 'strong preference' for ΛLTB rests on an unfair baseline—ΛCDM never gets the same M-step freedom—so the central model-selection result is not yet established.","tokens_in":25579,"tokens_out":2668,"would_cite":false,"duration_ms":25872,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["98.80.Es","98.65.Dx","98.80.-k"],"model":"deepseek-v4-flash","headline":"The paper argues that the Hubble tension disappears if we live in a deep local void and supernova brightness changes in steps with distance, and that this scenario is strongly preferred by the Pantheon+ supernova and Planck CMB data.","keywords":["Hubble tension","local void","LTB void model","absolute magnitude transition","Type Ia supernovae","Pantheon+","information criteria","cosmological constant"],"falsifier":"Re-fit the same Pantheon+ and Planck 2018 distance-prior data to flat $\\Lambda$CDM with the same number of $M$-steps at the same fitted transition distances; if the BIC gap between $\\Lambda$CDM and the $\\Lambda$LTB models no longer meets the paper's 'very strong' threshold, the claimed preference for a void collapses. Separately, a galaxy-cluster or peculiar-velocity measurement inside roughly 100 Mpc that rules out $\\Omega_m \\approx 0.2$ would falsify the specific deep void fitted here.","tokens_in":24352,"feed_emoji":"🌌","tokens_out":20662,"duration_ms":159085,"temperature":0.7,"pith_summary":"The paper argues that the 5-$\\sigma$ gap between local and early-universe Hubble-constant measurements disappears if we live inside a deep, underdense void and the absolute brightness of Type Ia supernovae is allowed to change in steps with distance. Fitting $\\Lambda$LTB void cosmologies to the Pantheon+ supernova sample, alone or with Planck 2018 CMB distance priors, yields $H_0 \\approx 73$ km/s/Mpc at the void center and $67.36$ km/s/Mpc outside the void, so both sides of the tension can be correct at once. The authors report that these void models are strongly preferred over flat $\\Lambda$CDM baselines when CMB data are included, with statistically significant $M$ transitions at roughly 20, 129, and 860–960 Mpc. A reader should care because the proposal offers a local, non-exotic resolution of the Hubble tension, at the cost of a deliberately inhomogeneous universe and a non-universal supernova brightness.","feed_headline":"Deep local void plus step-changing supernovae eases Hubble tension","feed_subtitle":"Inside the void the expansion rate is about 73; outside it matches the CMB value 67.4. Both can be right.","key_machinery":"The central object is the $\\Lambda$LTB void: a spherically symmetric, radially inhomogeneous cosmology with a cosmological constant, in which the matter density dips to $\\Omega_m(r) \\approx 0.2$ near the center through a constrained hyperbolic-tangent deficit profile and returns to the flat $\\Lambda$CDM background values $\\Omega_{m,\\mathrm{out}}=0.3153$ and $H_{0,\\mathrm{out}}=67.36$ km/s/Mpc far outside. It is paired with a piecewise-constant absolute magnitude $M$ whose transitions at $d_{\\mathrm{cr}} \\approx 20$, $\\approx 129$, and $\\approx 860$–960 Mpc are free parameters. The void changes the luminosity distance $d_L(z) = (1+z)^2 R(r(z), t(z))$ along radial null geodesics, producing a high local $H_0$ while preserving the CMB-calibrated exterior; the $M$-steps absorb distance-modulus jumps that would otherwise make the void fit worse. The model comparison is carried by $\\Delta\\chi^2$, $\\Delta$AIC, $\\Delta$BIC, and the Bayes factor.","core_discovery":"The claim is that the Hubble tension is not a contradiction between datasets but a signature of our location: the Solar System sits near the center of a deep, radially inhomogeneous void described by the $\\Lambda$LTB metric with a cosmological constant. Inside the void the expansion rate $H_0(r=0)$ comes out near 73 km/s/Mpc, matching the local distance-ladder measurement, while far outside the void the same model matches the Planck 2018 CMB value $67.36$ km/s/Mpc. The extra ingredient that makes this work is allowing the SNIa absolute magnitude $M$ to take several discrete values, with fitted transitions at about 20, 129, and 860–960 Mpc; the void depth is driven to $\\delta_V \\approx -38\\%$ to $-50\\%$, so the local matter density falls to $\\Omega_m \\approx 0.2$. With the Pantheon+ sample alone, the $\\Lambda$LTB models with $M$ transitions are strongly preferred over a $\\Lambda$CDM model with $\\Omega_m$ and $H_0$ fixed by Planck, though not always over a fully free $\\Lambda$CDM when judged by BIC. Adding Planck 2018 CMB distance priors makes the preference very strong by all criteria used ($\\Delta\\chi^2$, AIC, BIC, Bayes factor), with $\\Delta$BIC between roughly $-9$ and $-46$.","pith_inferences":["Inference: a clean stress test would give flat $\\Lambda$CDM the same number of $M$-steps at the same fitted distances; if its $\\Delta$BIC versus the void models becomes weak, the paper's preference is mostly about model flexibility rather than a physical void.","Inference: because part of the 20 Mpc transition is attributed in earlier work to a volumetric redshift-scatter bias, the new steps near 129 and 860–960 Mpc should be checked against the same systematic corrections before being read as new physics.","Inference: if the $M$-steps come from changed physics of white-dwarf explosions, such as a varying effective gravitational constant, supernova light-curve properties like stretch or color should show correlated jumps at the same distances.","Inference: the deep-void picture predicts a coherent local expansion flow that could be looked for in bulk-flow and redshift-space distortion data independent of supernova magnitudes."],"forward_implications":["If the claim holds, the local distance-ladder value and the CMB-inferred value are both the true expansion rate, measured at different places; no new early-universe physics is required.","Distance-ladder analyses that assume a single absolute magnitude for all Type Ia supernovae would need to allow distance-dependent steps, since the data used here show transitions at about 20, 129, and 860–960 Mpc.","The fitted void is deep, with local matter density $\\Omega_m \\approx 0.2$, which is testable by galaxy-cluster and peculiar-velocity surveys at scales of tens to hundreds of megaparsecs.","The earlier failure of void models against the older Pantheon sample is attributed to missing $M$-step flexibility and the $H_0$–$M$ degeneracy of pre-Pantheon+ samples, not to the void idea itself."],"supporting_citations":[{"why":"Fixes the outside-the-void background expansion rate and matter density (67.36 km/s/Mpc and 0.3153) used as boundary conditions and defines the CMB side of the tension.","marker":"[11]"},{"why":"Provides the local Cepheid-calibrated value of about 73.04 km/s/Mpc that the void center is designed to match.","marker":"[12]"},{"why":"Establishes the ΛLTB void setup and the z-cut subsets, and shows that a void alone fails against the older Pantheon sample; the present work follows its void-depth criterion.","marker":"[48]"},{"why":"Releases the Pantheon+ sample and Cepheid-calibrated host distance moduli that break the degeneracy between expansion rate and absolute magnitude, allowing M to be fitted as separate parameters.","marker":"[50–52]"},{"why":"Reports the roughly 20 Mpc transition of the supernova absolute magnitude in Pantheon+, which this paper confirms and extends.","marker":"[63]"},{"why":"Tests sudden and linear M-transition models in ΛCDM and finds they fail under BIC, motivating the combination of M transitions with a void.","marker":"[64]"},{"why":"Supplies the Planck 2018 CMB distance priors (shift parameter and acoustic scale) used in the joint supernova+CMB fits.","marker":"[74]"}],"fun_headline_variants":["Local void plus supernova magnitude steps ease Hubble tension","Void and discrete supernova brightness resolve Hubble mismatch","Deep local void plus step-changing supernovae ease Hubble tension","Hubble tension eased by local void and supernova magnitude shifts","Our cosmic void and stepwise supernova brightness fix Hubble"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument stands on allowing the void models step transitions in the supernova absolute magnitude that the flat-$\\Lambda$CDM baselines do not get; the paper never fits a $\\Lambda$CDM baseline with the same step flexibility, so part of the claimed preference could be a reward for extra model flexibility.","fun_headline_variants_meta":{"raw":{"variants":["Local void plus supernova magnitude steps ease Hubble tension","Void and discrete supernova brightness resolve Hubble mismatch","Deep local void plus step-changing supernovae ease Hubble tension","Hubble tension eased by local void and supernova magnitude shifts","Our cosmic void and stepwise supernova brightness fix Hubble"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000701,"raw_usage":{"total_tokens":3221,"prompt_tokens":1058,"completion_tokens":2163,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":674,"completion_tokens_details":{"reasoning_tokens":2083}},"tokens_in":674,"tokens_out":2163,"duration_ms":16825,"temperature":1.0,"reasoning_tokens":2083,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T12:10:41.614398+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Re-fit the same Pantheon+ and Planck 2018 distance-prior data to flat $\\Lambda$CDM with the same number of $M$-steps at the same fitted transition distances; if the BIC gap between $\\Lambda$CDM and the $\\Lambda$LTB models no longer meets the paper's 'very strong' threshold, the claimed preference for a void collapses. Separately, a galaxy-cluster or peculiar-velocity measurement inside roughly 100 Mpc that rules out $\\Omega_m \\approx 0.2$ would falsify the specific deep void fitted here.","supporting_citations":[],"review_version":1}