Pith. sign in

REVIEW 3 major objections 5 minor 27 references

Deus ex $H_0$ -- Is evidence for dynamical dark energy conditioned on early cosmology?

T0 review · 3 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read This paper argues that the evidence for dynamical dark energy is conditioned on early-universe cosmology: Hubble-tension solutions shift the inferred $H_0 r_d$ and $\Omega_m$ into a region where the preference drops from roughly 2.5σ to…

desk verdict A clean, likely-correct argument that the DESI dynamical dark energy preference is conditioned on early-universe assumptions; the main weakness is an unvalidated Gaussianity assumption in the significance map. read the letter →

arxiv 2608.07654 v1 pith:VJMPIRMV submitted 2026-08-07 astro-ph.CO

Nils Schöneberg , Rodrigo Calderón , Julien Lesgourgues This is my paper · ORCID
classification astro-ph.CO
keywords dynamicaldarkenergyHubbletensionbaryonacousticoscillationssupernovaeequationofstateCPLparametrizationsoundhorizonearlyuniversecosmology
topics Dark EnergyThe Hubble Tension
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper argues that the widely quoted preference for dynamical dark energy from baryon acoustic oscillations and supernovae is not a model-independent fact: it depends on what one assumes about the early universe, specifically on the inferred product of the Hubble constant and the sound horizon, $H_0 r_d$, and the matter density $\Omega_m$. Using a grid of fixed values for these two parameters, the authors fit a two-parameter dark energy equation of state (the Chevallier-Polarski-Linder form) to BAO and supernova data and convert the distance from a cosmological constant into a significance. The significance is smallest near a region they call the 'nexus,' centered near $h r_d \approx 103\,\mathrm{Mpc}$ and $\Omega_m \approx 0.285$ for the DES Dovekie sample. Models that resolve the Hubble tension with early-universe physics generically shift the inferred parameters toward this nexus, weakening the preference for dynamical dark energy from roughly 2.5σ to 1.3σ while slightly increasing a residual tension with supernovae. If correct, claims of phantom crossing or thawing dark energy are premature until early cosmology is pinned down.

What carries the argument

The central object is the geometric degeneracy encoded in $H(z)=H_0 E(z)$: BAO data measure $\beta(H_0 r_d)/f(z_i)$ and $(H_0 r_d)E(z_i)$, while uncalibrated supernovae measure $(1+z)^2/f(z)$, so these probes constrain only the product $H_0 r_d$ and the shape of $E(z)$. The CMB angular scale contributes the nearly model-independent band $\theta_{\rm CMB}=\beta_{\rm drag}\beta(H_0 r_d)/f(z_{\rm CMB})$. The paper scans a grid in ($H_0 r_d,\Omega_m$); at each point it fits the Chevallier-Polarski-Linder parameters $w_0,w_a$ with $w(a)=w_0+w_a(1-a)$, and converts the covariance-weighted Mahalanobis distance from the cosmological-constant point ($w_0=-1,w_a=0$) into a $\sigma$ level through a $\chi^2_2$ distribution. The 'nexus' is the local minimum of this $\sigma$ map. A fourth-order Chebyshev expansion of the dark energy density serves as an independent reconstruction check.

What would settle it

A reader could test the paper's central claim by computing the full parameter-shift metric or nested-sampling evidence ratio for CPL versus $\Lambda$CDM at every grid point instead of using the Gaussian Mahalanobis conversion; if the significance at the nexus rises above about 1σ, or the minimum moves noticeably, the claimed conditioning on early cosmology weakens.

Watch

Extended reading notes

Core claim

The central discovery is that the level of evidence for dynamical dark energy, measured by fitting the CPL parametrization $w(a)=w_0+w_a(1-a)$ to BAO and uncalibrated supernova data, is not a single number but a field over the early-universe parameters ($H_0 r_d,\Omega_m$). Only the product $H_0 r_d$ and the expansion shape $E(z)$ are accessible to these late-time probes, so the same data can support 'cosmological constant' or 'dynamical dark energy' depending on where the early universe places the model. The CMB angular scale provides an almost model-independent ridge in this plane, and the minimum of the dynamical-dark-energy preference, the 'nexus,' sits on that ridge without having used CMB data. Early-universe solutions to the Hubble tension, which move the joint fit to higher $H_0 r_d$ and lower $\Omega_m$, push the fit toward that nexus and reduce the significance from about 2.5σ to 1.3σ, making the apparent preference depend on the assumed early cosmology. A fourth-order Chebyshev reconstruction of the dark energy density confirms the same weakening near the nexus.

Load-bearing premise

The result assumes the dark-energy fit contours are nearly bell-shaped at every grid point, so the Mahalanobis distance can be read as a sigma level; the paper states this is a good approximation but does not verify it at every point, and a failure there would shift the reported significance and the exact nexus location.

Editorial extensions

If this is right

  • The reported preference for dynamical dark energy carries an implicit early-universe prior: with early-universe Hubble-tension solutions, it drops from roughly 2.5σ to 1.3σ.
  • A phantom crossing or thawing behavior inferred from BAO and supernovae is not evidence for new physics until the early-universe parameters $H_0 r_d$ and $\Omega_m$ are independently fixed.
  • The nexus lies on the CMB angular constraint even though no CMB data enter its construction, linking the weak-dark-energy region to the geometric CMB ridge.
  • The residual low-redshift tension with supernovae grows by about 0.5σ when early-universe models move to the nexus, so the Hubble-tension fix and the dark-energy preference trade against each other.
  • Future measurements of $\Omega_m$ from galaxy full-shape clustering and Lyman-$\alpha$ data will identify where the true cosmology sits on the CMB constraint, and therefore how much dark-energy evidence survives.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A natural extension is that any late-time reconstruction of dark energy, not just CPL, will have its evidence reweighted by early-universe assumptions, since BAO and supernovae only fix $H_0 r_d$ and $E(z)$.
  • The paper's logic implies that an independent, model-free measurement of $\Omega_m$, for example from galaxy clustering or CMB lensing, would be more decisive for settling dynamical dark energy than more low-redshift supernovae.
  • If future data push $\Omega_m$ down toward the nexus, the same geometric tension now read as 'dynamical dark energy' could instead be read as early-universe physics plus a residual supernova offset.
  • A stress test the authors did not run is to compute full Bayesian evidence at each grid point; doing so would show whether the quantitative sigma levels, not just the qualitative nexus, survive non-Gaussian contours.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. This letter argues that the apparent evidence for dynamical dark energy (DDE) from BAO and supernova data is not model-independent but is conditioned on the assumed early-universe parameters H0 rd and Ωm. The authors show geometrically that BAO and uncalibrated SN data constrain only the expansion history E(z) and the product H0 rd (Eq. 1), while the CMB angular scale provides a nearly model-independent degeneracy line in (Ωm, H0 rd). They then divide the (Ωm, H0 rd) plane into a grid, fit a CPL dark-energy model to BAO+SN at each point, and convert the distance of the posterior from w0=-1, wa=0 into a significance using the Mahalanobis distance (Eqs. A1-A2). The resulting map has a minimum ('nexus') at high H0 rd and low Ωm, close to the region preferred by early-universe Hubble-tension solutions such as varying me, WZDR, and NEDE. They conclude that early-universe solutions generically reduce the DDE preference from >2σ to about 1.3σ, and that claims of phantom crossing or thawing DDE are premature.

Significance. If the central claim survives scrutiny, it has a significant impact on the interpretation of DESI and SN results, reframing DDE evidence as conditional on early cosmology. The geometric decomposition in Eqs. (1)-(2) is elegant, internally consistent, and largely assumption-light; the use of three SN catalogs and an independent Chebyshev reconstruction are commendable robustness checks. The paper also makes a concrete, falsifiable prediction: future SN and BAO data should show a lower Ωm if the true model has high H0 rd. The main quantitative result, however, rests on an unvalidated Gaussian approximation, so the significance values and the exact nexus location are not yet established to the standard required.

major comments (3)
  1. [Supplementary material, Eqs. (A1)-(A2)] The significance map in Fig. 1 is computed from the Mahalanobis distance of the CPL posterior to (-1,0), assuming that each (w0, wa) posterior is a two-dimensional Gaussian. The supplementary text asserts that contours are 'typically close to Gaussian' and 'near-Gaussian nature ... throughout the parameter space', but no diagnostic is shown. This is load-bearing because the paper's headline numbers (reduction from about 2.5σ to 1.3σ, and the nexus position at hrd ≈ 103 Mpc, Ωm ≈ 0.285) are read from this map. At grid points far from the BAO+SN best fit, or where the w0-wa degeneracy is strong, the posterior can be skewed or heavy-tailed, in which case the Mahalanobis distance to a single point is not the evidence ratio against a cosmological constant. I request a validation of the Gaussian assumption at several representative grid points, for example by comparing the Mahalanobis significance with the full parameter-shift metric of Ref. [11] or with an evidence-ratio calculation, and by over-plotting the actual 2D posterior contours.
  2. [Fig. 2 and Fig. 6] The Chebyshev-polynomial check is presented as confirming the main claim, but it is performed at a single point in parameter space: hrd is fixed to 102.4 Mpc (the yellow star in Fig. 1) and only Union3 SN data are used in Fig. 2, while Fig. 6 compares only this fixed hrd with a free hrd. This does not test the shape of the green iso-contours across the (Ωm, hrd) plane, nor the location of the nexus. To support the statement 'We confirm this with a full reconstruction', a grid of Chebyshev fits (or at least a handful of representative points) would be needed. Without that, the claim of robustness to the CPL parametrization is only demonstrated at the nexus itself.
  3. [Conclusions, final paragraph] The conclusion that 'Models that solve the Hubble tension necessarily push the investigation to trade a >2σ preference for DDE for an 0.5σ increase in the SN tension' is stronger than what the analysis demonstrates. The evidence consists of three specific model classes (varying me, WZDR, NEDE) whose best-fit positions are taken from Ref. [11], plus a heuristic argument citing Ref. [17] that typical early-universe solutions increase hrd and lower Ωm. This supports 'generically' or 'the models considered here', but not 'necessarily'. If there exist early-universe Hubble-tension solutions that do not move toward the nexus (e.g., models that simultaneously raise Ωm h²), the claim would be false. Please either provide a more rigorous proof of the universality or soften the wording.
minor comments (5)
  1. [Introduction, first sentence] There is a typo: 'redhsift' should be 'redshift'.
  2. [Supplementary material, first paragraph] 'Mahalabonis' is a misspelling of 'Mahalanobis'.
  3. [Fig. 1 and surrounding text] The text claims the preference is 'below 0.5σ' in the nexus, but the green iso-contours in Fig. 1 only label 1σ, 2σ, 3σ, 4σ, and 5σ; please add a 0.5σ contour or state the numerical minimum significance from the grid.
  4. [Eq. (A2)] The notation for the chi-squared cumulative distribution is unclear ('Cχ2d' in one place, 'F χ2 d' in another); please use a single, unambiguous symbol such as F_{χ²_d} throughout.
  5. [Ref. [24]] The reference for Brieden, Gil-Marín, and Verde appears malformed ('JCAP12(12)'); please correct the journal volume, issue, or article number.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the evidence map is computed from BAO+SN data at fixed (Ωm, hrd), and the early-model positions are external inputs, not consequences of the claimed conclusion.

full rationale

The paper's central claim is that the preference for dynamical dark energy is conditioned on early-universe parameters (Ωm, hrd). The derivation chain is a conditional analysis: for each grid point in (Ωm, hrd), the authors fit a CPL dark energy model to BAO and SN data, compute the Mahalanobis distance of the best-fit (w0, wa) from (−1, 0), and convert it to a significance via Eqs. (A1)–(A2). This map is an output of the data, not an input; the conclusion that the evidence weakens at high hrd and low Ωm is data-driven rather than built into the construction. The early-model star positions are taken from the authors' own Ref. [11], but they serve only as illustrations of where Hubble-tension solutions sit; the geometric argument that such a region has weaker DDE preference stands independently of that citation. The Gaussian-Mahalanobis approximation in the supplementary material is a stated modeling assumption that could affect the quantitative significance levels, but it is not circular: it is an approximation about the shape of posteriors, not a restatement of the target claim. The Chebyshev reconstruction provides an independent cross-check at a single fixed point. No equation in the paper defines the target conclusion in terms of itself, and no fitted parameter is renamed as a prediction. The self-citations present are not load-bearing in a circular sense.

Assumptions & free parameters 3 free parameters · 6 assumptions · 0 invented entities

The analysis is a data-driven mapping rather than a first-principles derivation. It relies on standard geometric assumptions, the robustness of the CMB angular scale to early-universe physics, and a near-Gaussian statistical approximation. No new physical entities are introduced; the nexus is a named region, not a postulate.

free parameters (3)
  • CPL equation-of-state parameters w0, wa = per-grid-point fit outputs, not single values
    These two parameters are fitted to BAO and supernova data at each fixed (Omega_m, h_rd) grid point; their covariance is used to compute the dynamical dark energy preference via the Mahalanobis distance.
  • Chebyshev dark-energy density coefficients = order N=1..4 coefficients, not quoted
    Used in the free-form reconstruction to show that the preference weakens when h_rd is fixed at the nexus value.
  • Phase-shift factor beta = assumed approximately 1
    Appears in Eqs. (1) and (2); the authors state it is very close to one and varies at the 0.1 percent level, but it is an input assumption rather than a fitted quantity.
assumptions (6)
  • domain assumption The universe is spatially flat, with f(z) the integral of dz'/E(z').
    Used in Eq. (1) to relate BAO angular and radial scales to the expansion history; no curvature parameter is included.
  • domain assumption For a late-time Lambda CDM expansion history, BAO and uncalibrated supernova constraints are fully summarized by (Omega_m, h_rd).
    Central to building the black and orange contours in Figure 1 and to defining the plane in which the nexus is located.
  • domain assumption General relativity holds, so uncalibrated supernova brightness measures (1+z)^2/f(z).
    Used to convert supernova magnitudes into geometric distance information with only minimal assumptions.
  • domain assumption The CMB angular scale is nearly independent of early-universe physics, with f(z_CMB), beta_drag, and beta varying by only about 0.1 to 0.5 percent for reasonable parameter changes.
    Used to draw the thin red CMB band in Figure 1 and to argue that the nexus lies on the CMB degeneracy line.
  • standard math CPL contours in (w0, wa) are near-Gaussian, so the Mahalanobis distance follows a chi-squared distribution and converts to sigma via Eq. (A2).
    The authors assert near-Gaussianity throughout the parameter space and use it to justify the significance map; this is not independently demonstrated at every grid point.
  • ad hoc to paper The CPL parametrization and Chebyshev expansions of order N=4 adequately represent plausible dynamical dark energy behaviors.
    The significance map and the reconstruction in Figure 2 depend on these specific functional forms; other parameterizations could in principle yield different evidence levels.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Deus ex $H_0$ -- Is evidence for dynamical dark energy conditioned on early cosmology?." pith.science (2026). https://pith.science/paper/VJMPIRMV

@misc{pith2026260807654,
  author       = {Pith},
  title        = {Pith review of: Deus ex $H_0$ -- Is evidence for dynamical dark energy conditioned on early cosmology?},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VJMPIRMV}},
  note         = {Machine review of arXiv:2608.07654}
}
abstract

Some free-form reconstructions of the dark energy equation of state suggest that dynamical dark energy is the only explanation for the observed data. In this letter we argue that early Universe solutions to the Hubble tension (around or before recombination) generically cause the evidence for this claim to strongly reduce, establishing a tight connection between the early and late cosmology. In particular, the level of evidence for dynamical dark energy depends on the parameters $H_0 r_\mathrm{d}$ and $\Omega_\mathrm{m}$ and early universe solutions typically push towards higher values of $H_0 r_\mathrm{d}$ and lower values of $\Omega_\mathrm{m}$, where such evidence is reduced.

Figures

Figures reproduced from arXiv: 2608.07654 by the authors.

Figure 1
Figure 1. FIG. 1. The colored regions show the 68% and 95% credible intervals for (Ω [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2. Reconstruction of the dark energy density based on [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3. Mean and standard deviation of Ω [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: FIG. 4. Same as Figure [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: FIG. 5. Same as Figure [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: FIG. 6. Improvement in best fitting [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

27 extracted references · 4 canonical work pages

  1. [11]

    The $H_0$ world cup. II. A comprehensive competition between proposed Hubble tension solutions

    N. Sch¨ oneberg, V. Poulin, A. G. Ferrari, F. Finelli, J. Les- gourgues, L. Morelli, M. R. Mosbech, R. K. Sharma, and T. Simon, arXiv e-prints (2026), arXiv:2607.13283 [astro-ph.CO]

    work page Pith review arXiv 2026

  2. [17]

    Poulin, T

    V. Poulin, T. L. Smith, R. Calder´ on, and T. Simon, Phys. Rev. D111, 083552 (2025), arXiv:2407.18292 [astro- ph.CO]

    arXiv 2025

  3. [1]

    Lodhaet al.(DESI), Phys

    K. Lodhaet al.(DESI), Phys. Rev. D112, 083511 (2025), arXiv:2503.14743 [astro-ph.CO]

    arXiv 2025

  4. [2]

    Calderonet al.(DESI), JCAP2024(10), 048, arXiv:2405.04216 [astro-ph.CO]

    R. Calderonet al.(DESI), JCAP2024(10), 048, arXiv:2405.04216 [astro-ph.CO]

    arXiv

  5. [3]

    Guet al.(DESI), Nature Astron.9, 1879 (2025), [Erratum: Nature Astron

    G. Guet al.(DESI), Nature Astron.9, 1879 (2025), [Erratum: Nature Astron. 9, 1898–1898 (2025)], arXiv:2504.06118 [astro-ph.CO]

    arXiv 2025

  6. [4]

    R. K. Sharma and J. Lesgourgues, JCAP2026(02), 034, arXiv:2510.15835 [astro-ph.CO]

    arXiv

  7. [5]

    Model-Independent Indication for a Localized Anomaly in the Late-Time Expansion History

    S. Gupta Choudhury, P. Mukherjee, E. Di Valentino, and A. A. Sen, arXiv e-prints (2026), arXiv:2607.13009 [astro-ph.CO]

    work page Pith review arXiv 2026

  8. [6]

    G. P. Lynch, L. Knox, and J. Chluba, Phys. Rev. D110, 083538 (2024), arXiv:2406.10202 [astro-ph.CO]

    arXiv 2024

Show all 27 references
  1. [7]

    S. H. Mirpoorian, K. Jedamzik, and L. Pogosian, JCAP 2025(12), 050, arXiv:2504.15274 [astro-ph.CO]

    2025

  2. [8]

    Chaussidonet al., Phys

    E. Chaussidonet al., Phys. Rev. D112, 063548 (2025), arXiv:2503.24343 [astro-ph.CO]

    2025 arXiv

  3. [9]

    A. R. Khalifeet al.(SPT-3G), Phys. Rev. D113, 103546 (2026), arXiv:2507.23355 [astro-ph.CO]

    2026 arXiv

  4. [10]

    Jhaveri, T

    T. Jhaveri, T. Karwal, T. Crawford, W. Hu, A. R. Khal- ife, L. Balkenhol, and F. Ge, arXiv e-prints (2026), arXiv:2604.08530 [astro-ph.CO]

    2026 arXiv

  5. [12]

    Asensio-Rivera, N

    F. Asensio-Rivera, N. Sch¨ oneberg, H. Gil-Mar ´ ın, and L. Verde, JCAP2026(07), 046, arXiv:2603.03443 [astro- ph.CO]

    arXiv

  6. [13]

    Abdul Karimet al.(DESI), Phys

    M. Abdul Karimet al.(DESI), Phys. Rev. D112, 083515 (2025), arXiv:2503.14738 [astro-ph.CO]

    2025 arXiv

  7. [14]

    Shlivko and V

    D. Shlivko and V. Poulin, arXiv e-prints (2026), arXiv:2603.22406 [astro-ph.CO]

    2026

  8. [15]

    Sch¨ oneberg, G

    N. Sch¨ oneberg, G. Franco Abell´ an, A. P´ erez S´ anchez, S. J. Witte, V. Poulin, and J. Lesgourgues, Phys. Rept. 984, 1 (2022), arXiv:2107.10291 [astro-ph.CO]

    2022 arXiv

  9. [16]

    A. R. Khalife, M. B. Zanjani, S. Galli, S. G¨ unther, J. Lesgourgues, and K. Benabed, JCAP2024(04), 059, arXiv:2312.09814 [astro-ph.CO]

    arXiv

  10. [18]

    Chevallier and D

    M. Chevallier and D. Polarski, Int. J. Mod. Phys. D10, 213 (2001), arXiv:gr-qc/0009008

    2001 arXiv

  11. [19]

    E. V. Linder, Phys. Rev. Lett.90, 091301 (2003), arXiv:astro-ph/0208512

    2003 arXiv

  12. [20]

    P. C. Mahalanobis, Proceedings of the National Institute of Sciences of India2, 49 (1936)

    1936

  13. [21]

    Exploring New Frontiers in Cosmology

    dataset. Equivalent results for Pantheon+ [22] and Union3 [23] are shown in Appendix A. 3 This region is inferred from a global BAO+SN fit, but appears visually as a better fit to BAO than SN data alone, due to the stronger constraining power of the for- mer dataset. Additiona...

  14. [22]

    Popovicet al.(DES), Mon

    B. Popovicet al.(DES), Mon. Not. Roy. Astron. Soc. 548, stag632 (2026), arXiv:2511.07517 [astro-ph.CO]

    2026 arXiv

  15. [23]

    Broutet al., Astrophys

    D. Broutet al., Astrophys. J.938, 110 (2022), arXiv:2202.04077 [astro-ph.CO]

    2022 arXiv

  16. [24]

    Rubinet al., Astrophys

    D. Rubinet al., Astrophys. J.986, 231 (2025), arXiv:2311.12098 [astro-ph.CO]

    2025 arXiv

  17. [25]

    Brieden, H

    S. Brieden, H. Gil-Mar ´ ın, and L. Verde, JCAP12(12), 054, arXiv:2106.07641 [astro-ph.CO]

    arXiv

  18. [26]

    Laiet al., JCAP2025(01), 139, arXiv:2404.07283 [astro-ph.CO]

    Y. Laiet al., JCAP2025(01), 139, arXiv:2404.07283 [astro-ph.CO]

  19. [27]

    D. M. Scolnicet al.(Pan-STARRS1), Astrophys. J.859, 101 (2018), arXiv:1710.00845 [astro-ph.CO]. 5 Supplementary material The main conclusions of thisLetterare robust against the choice of a particular supernova dataset. In Fig- ures 4 and 5 we reproduce Figure 1 for Pantheon+ ...

    2018 arXiv

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.