arxiv: 2605.13546 · v1 · submitted 2026-05-13 · 🌌 astro-ph.CO · gr-qc

Recognition: 1 theorem link

· Lean Theorem

No evidence for phantom crossing: local goodness-of-fit improvements do not persist under global Bayesian model comparison

Bikash R. Dinda , Roy Maartens , Shun Saito

Authors on Pith no claims yet

Pith reviewed 2026-05-14 19:06 UTC · model grok-4.3

classification 🌌 astro-ph.CO gr-qc

keywords dark energyphantom crossingBayesian evidencecosmological modelsw0wa parametrizationLambda CDMmodel comparison

0 comments

The pith

Global Bayesian comparison finds no evidence for phantom crossing or dynamical dark energy over standard Lambda CDM.

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

The paper examines recent cosmological data that some analyses interpret as hints of phantom crossing and dynamical dark energy within the w0wa parametrization. It compares Lambda CDM against w0wa and thawing quintessence models using both the Deviance Information Criterion for local fit quality and the Bayesian evidence ln Z for overall model preference. While w0wa sometimes yields a modestly better local fit in limited parameter regions, this gain disappears once the full prior volume is accounted for in the global evidence. The result is that all three models remain statistically indistinguishable, with no consistent support across datasets for either dynamical dark energy or phantom crossing.

Core claim

When DIC and Bayesian evidence are applied to multiple datasets, any apparent local improvements for w0wa are confined to narrow parameter regions and do not produce positive Delta ln Z once the entire prior volume is included; consequently, Lambda CDM, w0wa, and thawing quintessence are statistically equivalent with no robust evidence for phantom crossing.

What carries the argument

The combination of the Deviance Information Criterion for local goodness-of-fit and the Bayesian evidence ln Z for global model selection, applied to Lambda CDM versus w0wa and quintessence parametrizations.

If this is right

Cases with negative Delta DIC but negative Delta ln Z show that local fit gains are not statistically significant.
All models are statistically indistinguishable once global evidence is used.
No dataset combination yields consistent evidence for dynamical dark energy or phantom crossing.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

Future tighter data could still favor one model if it concentrates probability in the narrow region where local fits improve.
Parametrization-dependent hints require global evidence checks before being treated as detections.
Similar local-versus-global discrepancies may appear in other cosmological extensions that add parameters with large prior volumes.

Load-bearing premise

The chosen priors for the w0wa and quintessence parameters represent the full theory space without being too wide or mis-centered.

What would settle it

A dataset combination in which Delta ln Z becomes clearly positive for w0wa or quintessence while remaining consistent with negative Delta DIC, and this pattern repeats across independent analyses.

Figures

Figures reproduced from arXiv: 2605.13546 by Bikash R. Dinda, Roy Maartens, Shun Saito.

**Figure 3.** Figure 3: FIG. 3: Comparison of [PITH_FULL_IMAGE:figures/full_fig_p002_3.png] view at source ↗

**Figure 2.** Figure 2: FIG. 2 [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗

**Figure 4.** Figure 4: FIG. 4: Best-fit [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗

read the original abstract

Recent cosmological data have been interpreted as indicating deviations from $\Lambda$CDM within the standard $w_0w_a$ parametrization, including hints of phantom crossing and dynamical dark energy. However, such inferences can be parametrization-dependent and need not imply a statistically robust detection. We test these claims by comparing $\Lambda$CDM, $w_0w_a$, and thawing quintessence models, using the Deviance Information Criterion (DIC) and the Bayesian evidence $\ln \mathcal{Z}$. We find that $w_0w_a$ can provide a slightly improved local fit, but this improvement is confined to a limited region of parameter space. The global Bayesian evidence does not support it once the full prior volume is taken into account. In particular, cases with $\Delta{\rm DIC}<0$ but $\Delta \ln \mathcal{Z}<0$ indicate that these improvements are not statistically significant. We show that all models are statistically indistinguishable, and that there is no statistically consistent evidence across different datasets for either dynamical dark energy or phantom crossing.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper shows that local fit gains for w0wa models vanish under Bayesian evidence because the extra parameters are penalized over the full prior volume, leaving LambdaCDM, w0wa, and thawing quintessence statistically indistinguishable on current data.

read the letter

The main message is straightforward: recent claims of phantom crossing or dynamical dark energy from w0wa fits do not survive when you switch from local goodness-of-fit to global Bayesian evidence. The paper compares LambdaCDM against w0wa and a thawing quintessence model using both DIC and the marginal likelihood lnZ across several datasets. It finds that while w0wa can improve the fit in a restricted part of parameter space, that improvement disappears once the evidence integrates over the entire prior. Cases with negative Delta DIC but negative Delta lnZ illustrate the point cleanly. This is useful because it applies standard tools to an active debate without introducing new data or parametrizations. The work is careful to note that the evidence ratio is prior-dependent, which is the right caveat to flag. The soft spot is exactly the one the stress-test note raises: the size of the Occam penalty hinges on how the priors for w0 and wa are chosen. If those ranges are wider than the data actually constrain or than physically motivated bounds, the evidence against w0wa can be overstated. The abstract does not give the exact prior intervals or the precise data-selection cuts, so a referee would need to verify those choices against the original papers that reported the phantom-crossing hints. The math itself looks standard and the logic holds internally. This paper is for cosmologists who interpret dark-energy constraints from DESI, Euclid, or LSST. It does not claim to close the debate but supplies a clear check on how to read local improvements. It deserves a serious referee because the question is timely and the method is reproducible once the priors and likelihoods are specified.

Referee Report

2 major / 2 minor

Summary. The manuscript compares the flat ΛCDM model against the two-parameter w0wa dark-energy parametrization and a thawing quintessence model on multiple cosmological datasets. Using both the Deviance Information Criterion (DIC) and the Bayesian evidence ln Z, the authors report that w0wa occasionally yields a modestly better local fit (negative ΔDIC), but this improvement is confined to a limited region of parameter space; once the full prior volume is integrated in the marginal likelihood, Δln Z is negative, rendering the models statistically indistinguishable and providing no consistent evidence for dynamical dark energy or phantom crossing.

Significance. If the adopted priors are representative, the result supplies a useful methodological caution: local goodness-of-fit gains (DIC) need not survive global Bayesian model comparison once the Occam penalty from prior volume is included. The explicit side-by-side use of DIC and ln Z is a strength, as it directly illustrates when apparent parametrization-dependent hints fail to constitute model preference. The work therefore reinforces the view that current data do not yet distinguish these dark-energy descriptions at a statistically robust level.

major comments (2)

[Bayesian evidence calculation] The central claim that Δln Z < 0 demonstrates lack of support for w0wa (and hence no evidence for phantom crossing) is load-bearing on the chosen prior ranges for w0 and wa. The manuscript does not report a sensitivity test to prior width or centering; if the priors are substantially wider than the posterior support or than physically motivated bounds, the evidence penalty is inflated by construction and the conclusion that the models are indistinguishable becomes prior-dependent rather than data-driven.
[Results and dataset description] The statement that 'there is no statistically consistent evidence across different datasets' requires explicit documentation of the exact dataset combination, selection cuts, and likelihood implementations. Without these details the robustness of the Δln Z sign across subsets cannot be verified, weakening the cross-dataset claim.

minor comments (2)

Add a compact table listing ΔDIC and Δln Z for every model–dataset pair so that the magnitude of the local-versus-global discrepancy is immediately visible.
Clarify whether the thawing-quintessence model uses the same parameter count as w0wa or an effective one-parameter reduction; this affects the direct comparability of the evidence ratios.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful and constructive review. The comments highlight important aspects of prior sensitivity and reproducibility that we will address in the revision. Below we respond point by point to the major comments.

read point-by-point responses

Referee: The central claim that Δln Z < 0 demonstrates lack of support for w0wa (and hence no evidence for phantom crossing) is load-bearing on the chosen prior ranges for w0 and wa. The manuscript does not report a sensitivity test to prior width or centering; if the priors are substantially wider than the posterior support or than physically motivated bounds, the evidence penalty is inflated by construction and the conclusion that the models are indistinguishable becomes prior-dependent rather than data-driven.

Authors: We agree that the interpretation of Δln Z depends on the prior volume and that an explicit sensitivity test strengthens the result. The priors adopted in the manuscript are the standard ranges used in the recent literature for the w0wa parametrization (w0 ∈ [-3,1], wa ∈ [-3,3]), chosen to be broad enough to contain the posterior support while remaining within physically plausible bounds. To address the referee’s concern directly, the revised manuscript will include a new subsection reporting a sensitivity analysis in which the prior widths are varied by factors of two and recentered on the posterior means. We will demonstrate that the sign of Δln Z remains negative and the models stay statistically indistinguishable across these choices, confirming that the conclusion is not an artifact of the specific prior ranges. revision: yes
Referee: The statement that 'there is no statistically consistent evidence across different datasets' requires explicit documentation of the exact dataset combination, selection cuts, and likelihood implementations. Without these details the robustness of the Δln Z sign across subsets cannot be verified, weakening the cross-dataset claim.

Authors: We concur that full documentation of the data combinations is required for independent verification. While the manuscript references the standard public likelihoods (Planck 2018, DES Y3, BAO from SDSS/eBOSS, etc.), a consolidated description was not provided in a single location. In the revised manuscript we will add a dedicated table in the methods section that lists every dataset combination analyzed, the precise likelihood implementations and versions employed, and any selection cuts or additional priors applied to nuisance parameters. This addition will allow readers to reproduce and assess the cross-dataset consistency of the Δln Z results. revision: yes

Circularity Check

0 steps flagged

No circularity: standard Bayesian evidence and DIC applied without reduction to fitted inputs

full rationale

The paper applies the standard definitions of DIC and Bayesian evidence (marginal likelihood) to compare ΛCDM, w0wa, and thawing quintessence models on cosmological datasets. The key observation that local fit improvements (ΔDIC < 0) do not persist under global evidence (ΔlnZ < 0) follows directly from the marginalization integral over the full prior volume, which is an external statistical property and does not reduce by the paper's own equations to any quantity already fitted to the same data. No self-citations, uniqueness theorems, or ansatzes from prior author work are invoked as load-bearing steps; the analysis remains self-contained against external benchmarks and standard tools.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The analysis rests on standard cosmological assumptions and the usual definition of Bayesian evidence; no new entities are introduced.

free parameters (1)

w0 and wa
Two parameters of the w0wa parametrization that are varied and marginalized over in the evidence integral.

axioms (1)

standard math FLRW metric and general relativity as background
Invoked implicitly when defining the expansion history for all three models.

pith-pipeline@v0.9.0 · 5494 in / 1122 out tokens · 30508 ms · 2026-05-14T19:06:23.262788+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

38 extracted references · 38 canonical work pages · 15 internal anchors

[1]

A. G. Adame et al. (DESI), JCAP02, 021 (2025), arXiv:2404.03002 [astro-ph.CO]

work page internal anchor Pith review Pith/arXiv arXiv 2025
[2]

Calderon and others (DESI), JCAP 10, 048, arXiv:2405.04216 [astro-ph.CO]

R. Calderon et al. (DESI), JCAP10, 048 (2024), arXiv:2405.04216 [astro-ph.CO]

work page arXiv 2024
[3]

Lodha, A

K. Lodha et al. (DESI), Phys. Rev. D111, 023532 (2025), arXiv:2405.13588 [astro-ph.CO]

work page arXiv 2025
[4]

DESI DR2 Results II: Measurements of Baryon Acoustic Oscillations and Cosmological Constraints

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

work page internal anchor Pith review Pith/arXiv arXiv 2025
[5]

Extended Dark Energy analysis using DESI DR2 BAO measurements

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

work page internal anchor Pith review arXiv 2025
[6]

Planck 2018 results. VI. Cosmological parameters

N. Aghanim et al. (Planck), Astron. Astrophys.641, A6 (2020), [Erratum: Astron.Astrophys. 652, C4 (2021)], arXiv:1807.06209 [astro-ph.CO]

work page internal anchor Pith review Pith/arXiv arXiv 2020
[7]

M. S. Madhavacheril et al. (ACT), Astrophys. J.962, 113 (2024), arXiv:2304.05203 [astro-ph.CO]

work page arXiv 2024
[8]

The Pantheon+ Analysis: Cosmological Constraints

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

work page internal anchor Pith review Pith/arXiv arXiv 2022
[9]

Rubin et al.,Union Through UNITY: Cosmology with 2,000 SNe Using a Unified Bayesian Framework,Astrophys

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

work page arXiv 2025
[10]

T. M. C. Abbott et al. (DES), Astrophys. J. Lett.973, L14 (2024), arXiv:2401.02929 [astro-ph.CO]

work page arXiv 2024
[11]

The Dark Energy Survey Supernova Program: A Reanalysis Of Cosmology Results And Evidence For Evolving Dark Energy With An Updated Type Ia Supernova Calibration

B. Popovic et al. (DES), (2025), arXiv:2511.07517 [astro- ph.CO]

work page internal anchor Pith review Pith/arXiv arXiv 2025
[12]

Shlivko and V

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

work page arXiv 2026
[13]

R. E. Keeley, A. Shafieloo, and W. L. Matthewson, (2025), arXiv:2506.15091 [astro-ph.CO]

work page arXiv 2025
[14]

G´ omez-Valent and A

A. G´ omez-Valent and A. Gonz´ alez-Fuentes, Phys. Lett. B872, 140096 (2026), arXiv:2508.00621 [astro-ph.CO]

work page arXiv 2026
[15]

S. L. Guedezounme, B. R. Dinda, and R. Maartens, (2025), arXiv:2507.18274 [astro-ph.CO]

work page arXiv 2025
[16]

Coupled Dark Energy and Dark Matter for DESI: An Effective Guide to the Phantom Divide

S. Antusch, S. F. King, and X. Wang, (2026), arXiv:2604.08449 [astro-ph.CO]

work page internal anchor Pith review Pith/arXiv arXiv 2026
[17]

Roy and S

N. Roy and S. Chakrabarti, (2025), arXiv:2508.13740 [astro-ph.CO]

work page arXiv 2025
[18]

Silva and R

E. Silva and R. C. Nunes, JCAP11, 078 (2025), arXiv:2507.13989 [astro-ph.CO]

work page arXiv 2025
[19]

I. D. Gialamas, G. H¨ utsi, M. Raidal, J. Urrutia, M. Vasar, and H. Veerm¨ ae, Phys. Rev. D112, 063551 (2025), arXiv:2506.21542 [astro-ph.CO]

work page arXiv 2025
[20]

Challenges to the cosmological constant model following results from the Dark Energy Survey

S. Avila, J. Mena-Fern´ andez, and M. Vincenzi, Nature Astron.9, 1129 (2025), arXiv:2605.09560 [astro-ph.CO]

work page internal anchor Pith review Pith/arXiv arXiv 2025
[21]

Wang and K

Y. Wang and K. Freese, JCAP02, 023 (2026), arXiv:2505.17415 [astro-ph.CO]

work page arXiv 2026
[22]

Model-Independent Analysis of Type Ia Supernova Datasets and Implications for Dark Energy

Z. Wang and Y. Wang, (2026), arXiv:2604.11883 [astro- ph.CO]

work page internal anchor Pith review Pith/arXiv arXiv 2026
[23]

F. J. Qu et al. (ACT), Astrophys. J.962, 112 (2024), arXiv:2304.05202 [astro-ph.CO]

work page arXiv 2024
[24]

Distance Priors from Planck Final Release

L. Chen, Q.-G. Huang, and K. Wang, JCAP02, 028 (2019), arXiv:1808.05724 [astro-ph.CO]

work page internal anchor Pith review Pith/arXiv arXiv 2019
[25]

Zhai, C.-G

Z. Zhai, C.-G. Park, Y. Wang, and B. Ratra, JCAP07, 009 (2020), arXiv:1912.04921 [astro-ph.CO]

work page arXiv 2020
[26]

B. R. Dinda, JCAP09, 062 (2024), arXiv:2405.06618 [astro-ph.CO]

work page arXiv 2024
[27]

B. R. Dinda, R. Maartens, S. Saito, and C. Clarkson, JCAP08, 018 (2025), arXiv:2504.09681 [astro-ph.CO]

work page arXiv 2025
[28]

Do equation of state parametrizations of dark energy faithfully capture the dynamics of the late universe?

¨O. Akarsu, M. Caruana, K. F. Dialektopoulos, L. A. Escamilla, E. O. Kahya, and J. Levi Said, (2026), arXiv:2604.12987 [astro-ph.CO]

work page internal anchor Pith review Pith/arXiv arXiv 2026
[29]

Montefalcone and R

G. Montefalcone and R. Stiskalek, (2026), arXiv:2603.25735 [astro-ph.CO]

work page arXiv 2026
[30]

Nesseris, Y

S. Nesseris, Y. Akrami, and G. D. Starkman, (2025), arXiv:2503.22529 [astro-ph.CO]

work page arXiv 2025
[31]

T. Xu, S. Kumar, Y. Chen, A. J. S. Capistrano, and ¨O. Akarsu, (2026), arXiv:2602.11936 [astro-ph.CO]

work page arXiv 2026
[32]

D. D. Y. Ong, D. Yallup, and W. Handley, (2026), arXiv:2603.05472 [astro-ph.CO]

work page arXiv 2026
[33]

L. T. Hergt, S. Henrot-Versill´ e, M. Tristram, and D. Scott, (2026), arXiv:2602.06115 [astro-ph.CO]

work page arXiv 2026
[34]

emcee: The MCMC Hammer

D. Foreman-Mackey, D. W. Hogg, D. Lang, and J. Goodman, Publ. Astron. Soc. Pac.125, 306 (2013), arXiv:1202.3665 [astro-ph.IM]

work page internal anchor Pith review Pith/arXiv arXiv 2013
[35]

MultiNest: an efficient and robust Bayesian inference tool for cosmology and particle physics

F. Feroz, M. P. Hobson, and M. Bridges, Mon. Not. Roy. Astron. Soc.398, 1601 (2009), arXiv:0809.3437 [astro- ph]

work page internal anchor Pith review Pith/arXiv arXiv 2009
[36]

B. R. Dinda and R. Maartens, Mon. Not. Roy. Astron. Soc.542, L31 (2025), arXiv:2504.15190 [astro-ph.CO]

work page arXiv 2025
[37]

Cosmological Dynamics of Exponential Quintessence Constrained by BAO, Cosmic Chronometers, and DES-SN5YR/Pantheon+ Data

S. Sultana and S. Chattopadhyay, (2026), arXiv:2604.06941 [astro-ph.CO]

work page internal anchor Pith review Pith/arXiv arXiv 2026
[38]

R. E. Ouardi, A. Bouali, A. Errahmani, R. E. Keeley, A. Shafieloo, and T. Ouali, (2026), arXiv:2604.04056 [astro-ph.CO]

work page internal anchor Pith review Pith/arXiv arXiv 2026