Pith. sign in

REVIEW 5 major objections 5 minor 24 references

Exploring generalized Starobinsky Model of Inflation: Observational Constraints

T0 review · 5 major / 5 minor · reviewed 2026-08-09 · deepseek-v4-flash

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

Pith's one-line read A power-law generalization of Starobinsky inflation, with exponent beta = 1.987, is observationally viable under Planck-2018, BICEP/Keck and BAO data.

desk verdict A short proceedings that re-reports the authors' earlier PRD constraints on power-law Starobinsky inflation, with some sloppy equations and an unresolved Npivot contradiction; the central result is plausible but this text alone doesn't pin down the model actually run. read the letter →

arxiv 2502.04401 v1 pith:ZMSGTTFI submitted 2025-02-06 astro-ph.CO gr-qchep-th

Saisandri Saini , Akhilesh Nautiyal This is my paper · ORCID
classification astro-ph.COgr-qchep-th
keywords power-lawStarobinskymodelf(R)gravityinflationarycosmologycosmicmicrowavebackgroundtensor-to-scalarratioMarkovChainMonteCarloreheating
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

The paper argues that a generalized version of Starobinsky inflation, in which the curvature term is raised to a power $\beta$ (a so-called power-law Starobinsky model), is compatible with current cosmological data. Using Planck-2018, BICEP/Keck and BAO observations, the authors run a Markov Chain Monte Carlo analysis over the model parameters M and $\beta$, computing the scalar and tensor power spectra numerically without relying on the slow-roll approximation. The best-fit values are $\beta$ = 1.$987^{{+0.013}}$_{-0.016} and log10 M = -4.$72^{{+0.21}}$_{-0.20}, from which the derived spectral index n_s = 0.$9676^{{+0.0069}}$_{-0.0068} and tensor-to-scalar ratio r = 0.$0074^{{+0.0061}}$_{-0.0044} stay within Planck bounds. A sympathetic reading is that this establishes deviations from the $R^{2}$ Starobinsky model are observationally viable, with implications for supergravity-inspired inflation.

What carries the argument

The central object is the Einstein-frame scalar potential (Eq. 3) obtained from the $f(R)$ action $R+R^\beta$; it is an exponential times a power of $(1-e^{-2\chi/\sqrt{6}})$, whose shape interpolates between the flat plateau of ordinary Starobinsky inflation and steeper potentials as $\beta$ moves away from 2. The argument is carried by numerically integrating the background Friedmann and scalar-field equations together with the Mukhanov-Sasaki equation for scalar perturbations and the tensor mode equation, computing the primordial power spectra $P_R(k)$ and $P_t(k)$ and hence $n_s$ and $r$. These spectra are fed through CAMB and CosmoMC in an MCMC pipeline against Planck-2018, BICEP/Keck and BAO data, with the model parameters $M$ and $\beta$ constrained directly and $N_{pivot}$ treated as the e-folding number at the pivot scale. The machinery's job is to show that a full numerical treatment, not a slow-roll shortcut, produces the reported posterior.

What would settle it

Run the published pipeline on a publicly released version of the modified ModeCode that implements Eq. (3) with the $M_Pl^{2}$ factor restored and with N_pivot set to 50 as stated; if the resulting 95% contours do not contain $\beta$ = 1.987 and log10 M = -4.72, the reported constraints are an artifact of the code rather than of the model.

Watch

Extended reading notes

Core claim

On its own terms, the paper's central claim is that the power-law Starobinsky model, with action $S_J = -\frac{M_{Pl}^2}{2}\int d^4x\sqrt{-g}\left(R + \frac{1}{6M^2}\frac{R^\beta}{M_{Pl}^{2\beta-2}}\right)$, matches the observed CMB and BAO data when $\beta$ is slightly below the Starobinsky value of 2. The analysis treats $M$ and $\beta$ as the primary parameters and $n_s$, $r$ as derived outputs; the posterior peaks at $\beta = 1.987^{+0.013}_{-0.016}$ and $\log_{10} M = -4.72^{+0.21}_{-0.20}$ (95% C.L.), giving $n_s=0.9676^{+0.0069}_{-0.0068}$ and $r=0.0074^{+0.0061}_{-0.0044}$. Because the potential and perturbation spectra are computed numerically rather than through slow-roll approximations, the result is presented as a direct test of the model rather than an analytic estimate. This is what the paper means by confirming that deviations from the $R^2$ Starobinsky model are observationally viable.

Load-bearing premise

The constraints inherit the assumption that the modified ModeCode used for the analysis solves exactly the model defined by the printed equations, with correct mass dimensions and a consistent treatment of N_pivot, an assumption the manuscript does not fully pin down.

Editorial extensions

If this is right

  • If the central claim is right, the original Starobinsky $R^2$ model sits within or very close to the 95% allowed region, since the interval for $\beta$ reaches 2.000.
  • The reported $r=0.0074$ is well below current upper limits, so next-generation CMB experiments targeting $r\sim 10^{-3}$ will be needed to distinguish this model from simpler single-field potentials.
  • The constraints transfer to supergravity-inspired inflation models that produce analogous power-law corrections, narrowing the allowed couplings.
  • Fixing $N_{pivot}$ while scanning reheating scenarios means the model's predictions for $n_s$ and $r$ carry an implicit reheating assumption; testing other reheating histories would shift the contours.

Reading between the lines

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

  • Reading the posterior arithmetically, the 68% interval for $\beta$ tops out at 1.998 while the 95% interval reaches exactly 2.000; I would characterize the tension with pure $R^2$ as suggestive rather than decisive.
  • A natural cross-check that goes beyond the paper is to run the same MCMC with the exact printed potential and with $N_{pivot}$ left free, and to compare the resulting contours against the slow-roll approximation across the prior range; this would show where the full numerical treatment matters.
  • If the method is correct, it offers a template for precision-constraining other $f(R)$ inflation models whose potentials are not slow-roll friendly, including those with running spectral index.
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

5 major / 5 minor

Summary. The paper studies the power-law Starobinsky model, an f(R) modification with an R^β correction, and reports constraints on the two model parameters (β and M) from Planck-2018, BICEP/Keck, and BAO data. The authors state that they numerically solve the background and perturbation equations using an adapted version of ModeCode, and then use CAMB and CosmoMC to perform an MCMC analysis. They report β = 1.987^{+0.013}_{-0.016} and log10M = -4.72^{+0.21}_{-0.20} at 95% CL, with derived values ns = 0.9676^{+0.0069}_{-0.0068} and r = 0.0074^{+0.0061}_{-0.0044}, and conclude that deviations from the original Starobinsky R^2 model are observationally viable. The abstract also announces a reheating analysis with Npivot fixed at 50, but no reheating analysis appears in the body of the paper.

Significance. If the reported constraints are correct, the result is a useful numerical, beyond-slow-roll confirmation of earlier slow-roll analyses of the generalized Starobinsky model, and it provides an explicit example of a fitted inflationary potential whose derived spectral index and tensor-to-scalar ratio are consistent with Planck. The use of the standard ModeCode+CAMB+CosmoMC pipeline is a strength, and the comparison with the authors' earlier work in Ref. [17] helps frame the result. However, the paper as written does not allow the numerical model to be unambiguously reconstructed from its equations: the printed potential is dimensionally incomplete, the second Friedmann equation is missing the kinetic term, and the treatment of Npivot is contradictory between the abstract, §3, and Table 1. The derived ns and r consistency with Planck is not an independent validation because the same data were used in the fit. These issues currently limit the significance of the paper to that of a proceedings contribution that needs substantial clarification before its central claim can be assessed.

major comments (5)
  1. [2.2, Eq. (3)] Equation (3) for the Einstein-frame potential U(χ) is dimensionally incomplete. With M declared dimensionless in Eq. (1) and M_Pl appearing explicitly in Eqs. (1) and (4), the right-hand side of Eq. (3) has no mass dimension, while the potential in Eq. (2) must have mass dimension 4. As printed, this expression cannot be the potential used in the Friedmann equation (4). Please provide the corrected version with the required M_Pl factors, or state explicitly and consistently that M_Pl = 1 units are used throughout, and describe how the adapted ModeCode implements this potential.
  2. [2.3, Eq. (5)] Equation (5) is not a valid second Friedmann equation for a canonical scalar field: the right-hand side should contain at least the kinetic term -χdot²/(2M_Pl²). As written, the equation is not closed and does not determine ˙H from the phase-space variables. Please correct the equation and confirm that the background solver used in the numerical calculations actually integrates the complete equations.
  3. [Abstract; §3; Table 1] The treatment of Npivot is internally contradictory. The abstract and §1 state that Npivot is fixed at 50, but Table 1 reports Npivot as a constrained parameter (48.7^{+2.9}_{-7.8} at 68% CL) and §3 says Npivot is varied. Because ns and r are evaluated at the pivot scale and Fig. 1(a) shows a strong correlation between Npivot and (β, log10M), the reported 95% intervals for β and log10M depend on which convention is used. Please state which analysis produced Table 1, and if both fixed-Npivot and varied-Npivot analyses were run, present the fixed-Npivot results separately.
  4. [Abstract; full text] The abstract announces that 'a general reheating scenario' is analyzed, but no reheating analysis appears anywhere in the text. Fixing Npivot to 50 is not a reheating calculation. Please either present the reheating constraints (e.g., on reheating duration, temperature, or equation-of-state parameter) or remove this claim from the abstract.
  5. [Abstract; §3] The statement that the derived ns and r 'lie within the bounds set by Planck' is not an independent validation, because β, log10M, and (in Table 1) Npivot were fitted to the same Planck-2018 and BICEP/Keck data from which ns and r are then derived via Eqs. (11)-(13). The agreement is a posterior consistency check, not a prediction. Please rephrase this claim accordingly.
minor comments (5)
  1. [§2.2-§2.3] The potential is denoted U(χ) in Eqs. (2) and (3) but V(χ) in Eqs. (4)-(6); please use one notation consistently throughout.
  2. [§2.1-§2.2] The sign of the Einstein-Hilbert term in Eqs. (1) and (2) is opposite to the standard convention implied by Eq. (4). Please clarify the metric signature and the sign convention for R so that the action and the Friedmann equations are mutually consistent.
  3. [§2.3, Eq. (9)] Equation (9) contains notation errors: the first term should be the second derivative d²v_k/dτ², and the term (1/a)da''/dτ² should be written in standard form, e.g., a''/a. Please correct the notation.
  4. [Abstract; Table 1; Ref. [18]] The data set is described inconsistently: the abstract and text say BICEP3, while Table 1 caption says BICEP/Keck (BK15) and Ref. [18] is the BICEP/Keck 2021 result. Please specify exactly which BICEP/Keck likelihood was used.
  5. [§3] Please provide the exact likelihood combination (e.g., Planck high-ℓ TTTEEE, low-ℓ, lensing, BICEP/Keck, and the specific BAO data set) and MCMC convergence diagnostics, since without these details Table 1 cannot be reproduced. It would also help to clarify the incremental contribution of this paper relative to Ref. [17].

Circularity Check

1 steps flagged · score 6.0 of 10

Reported ns and r are consistency statements from the same fit, not independent predictions; parameter constraints are a direct fit with reproducibility caveats

  1. fitted input called prediction [Abstract; Section 3 (Table 1); Eqs. (11)-(13)]
    "However, in this study ns and r are treated as derived quantities, while the primary parameters constrained directly by the CMB data are those of the inflaton potential (3), namely M and β. ... The derived scalar spectral index ns=0.9676+0.0069/−0.0068 and tensor-to-scalar ratio r=0.0074+0.0061/−0.0044 lie within the bounds set by Planck observations."

    By Eqs. (11)-(13), ns and r are logarithmic derivatives and a ratio of the same scalar and tensor power spectra that were computed from M and β and passed to the Planck likelihood. Fitting M and β therefore already determines ns and r; the reported posterior for ns/r is a coordinate transform of the M/β posterior, not an independent quantity. Comparing these derived values with Planck's ns−r bounds uses the same dataset on both sides, so the statement is a consistency check of the fit, not a model prediction.

full rationale

The paper's main quantitative result (β and log10M constraints) comes from a direct MCMC fit of the power-law Starobinsky potential to Planck/BICEP/BAO data; that is parameter estimation, not a circular derivation. The one genuinely circular presentation is the abstract's claim that the derived ns and r 'lie within the bounds set by Planck observations.' Because ns and r are computed from the same scalar and tensor power spectra that define the likelihood, their posterior is a deterministic reparametrization of the fitted M and β posterior; the comparison with Planck's ns−r bounds is therefore a restatement of the fit, not a model prediction. This is a partial circularity (pattern 2), but it is not load-bearing for the central β constraint. The priors are taken from the authors' earlier paper [17], but this is a normal prior choice and does not by itself force the result. Separately, the printed equations are not self-contained: Eq. (3) lacks the M_Pl factors needed for a mass-dimension-4 potential, Eq. (5) omits the kinetic term, and the abstract's claim that Npivot is fixed at 50 conflicts with Table 1 and the text reporting Npivot as a varied, constrained parameter. These are reproducibility/correctness defects, not circularity, but they prevent the reader from verifying that the numerical spectra were generated from the stated model.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The central result is a fit of three parameters, beta, log10 M, and (despite the abstract) Npivot, to external CMB and BAO data. The model rests on standard f(R)-to-Einstein-frame machinery and on standard assumptions about initial conditions and likelihoods. No new entities are introduced.

free parameters (3)
  • beta = 1.987 +0.013/-0.016 (95% CL)
    Power-law exponent in the f(R) action, fitted to Planck+BICEP/Keck+BAO data via MCMC.
  • log10 M = -4.72 +0.21/-0.20 (95% CL)
    Dimensionless coupling of the R^beta term, fitted to the same data.
  • Npivot = 48.7 +2.9/-7.8 (68% CL)
    Number of e-folds at the pivot scale; abstract says fixed at 50 but Table 1 treats it as a fitted parameter.
assumptions (5)
  • domain assumption The generalized f(R) action (Eq. 1) with R^beta is the correct classical gravity model during inflation.
    The entire analysis is built on this action; no derivation from a more fundamental theory is given.
  • standard math The conformal transformation to the Einstein frame and the resulting potential U(chi) (Eq. 3) are valid and completely describe inflation.
    This is a standard f(R) procedure, but Eq. (3) as printed has a dimensional inconsistency (missing M_Pl^2 factor).
  • domain assumption The scalar and tensor perturbations start in the Bunch-Davies vacuum deep in slow roll.
    Stated in Section 2.3; a standard assumption for single-field inflation.
  • domain assumption The Planck-2018, BICEP/Keck, and BAO likelihoods are correctly implemented in CosmoMC.
    The constraints are only as reliable as the external data and likelihood code.
  • domain assumption The prior ranges log10 M in [-6.5,-3.0] and beta in [1.90,2.07] do not artificially exclude viable regions.
    Priors copied from [17]; the narrow beta range restricts the model to near-Starobinsky.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Exploring generalized Starobinsky Model of Inflation: Observational Constraints." pith.science (2026). https://pith.science/paper/ZMSGTTFI

@misc{pith2026250204401,
  author       = {Pith},
  title        = {Pith review of: Exploring generalized Starobinsky Model of Inflation: Observational Constraints},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZMSGTTFI}},
  note         = {Machine review of arXiv:2502.04401}
}
abstract

We examine the power-law Starobinsky model, a generalized version of the Starobinsky inflation model, characterized by a power-law correction to Einstein gravity. Employing the $f(R)$ formalism, the scalar and tensor power spectra were numerically computed as functions of the dimensionless parameters $M$ and $\beta$. A Markov Chain Monte Carlo (MCMC) analysis was conducted using Planck-2018, BICEP3 and BAO observational data, yielding precise constraints on $\beta = 1.987^{+0.013}_{-0.016},\, 95\%\, C.\, L.$. and $ \log_{10}M = -4.72^{+0.21}_{-0.20}$. The derived scalar spectral index $n_s=0.9676^{+0.0069}_{-0.0068}$ and tensor-to-scalar ratio $r=0.0074^{+0.0061}_{-0.0044}$ lie within the bounds set by Planck observations. We analyse a general reheating scenario while keeping the number of e-folds during inflation, $N_{pivot}$, fixed. The analysis confirms that deviations from the Starobinsky $R^2$ model are observationaly viable, with implications for high-energy physics and supergravity-based inflationary models.

Figures

Figures reproduced from arXiv: 2502.04401 by the authors.

Figure 1
Figure 1. Marginalized joint two-dimensional 68% C.L. and 95% C.L. con [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

24 extracted references · 7 canonical work pages

  1. [17]

    Observational constraints on power law Starobinsky inflation

    S. Saini and A. Nautiyal, Phys. Rev. D 108, no.12, 123505 (2023) doi:10.1103/PhysRevD.108.123505 [arXiv:2305.00682 [astro-ph.CO]]

    work page Pith review arXiv 2023

  2. [1]

    A. H. Guth, Phys. Rev. D23, 347 (1981)

    1981

  3. [2]

    A. D. Linde, Phys. Lett. B 108, 389-393 (1982) doi:10.1016/0370-2693(82)91219-9

    doi:10.1016/0370-2693(82)91219-9 1982

  4. [3]

    V. F. Mukhanov and G. V. Chibisov, JETP Lett. 33, 532 (1981)

    1981

  5. [4]

    A. A. Starobinsky, Phys. Lett. B117, 175 (1982)

    1982

  6. [5]

    A. H. Guth and S.-Y. Pi, Phys. Rev. D32, 1899 (1985)

    work page 1985

  7. [6]

    Martin, C

    J. Martin, C. Ringeval and V. Vennin, Phys. Dark Univ. 5-6, 75-235 (2014) doi:10.1016/j.dark.2014.01.003 [arXiv:1303.3787 [astro-ph.CO]]

    arXiv 2014

  8. [7]

    A. A. Starobinsky, Phys. Lett. B 91, 99-102 (1980) doi:10.1016/0370-2693(80)90670-X

    doi:10.1016/0370-2693(80)90670-x 1980

Show all 24 references
  1. [8]

    Akrami et al

    Y. Akrami et al. [Planck], Astron. Astrophys. 641, A10 (2020) doi:10.1051/0004- 6361/201833887 [arXiv:1807.06211 [astro-ph.CO]]

    2020 arXiv

  2. [9]

    Muller, H

    V. Muller, H. J. Schmidt and A. A. Starobinsky, Class. Quant. Gra v. 7, 1163-1168 (1990) doi:10.1088/0264-9381/7/7/012

    1990 doi

  3. [10]

    Gottlober, V

    S. Gottlober, V. Muller, H. J. Schmidt and A. A. Starobinsky, In t. J. Mod. Phys. D 1, 257-279 (1992) doi:10.1142/S0218271892000136

    1992 doi

  4. [11]

    Ellis, D

    J. Ellis, D. V. Nanopoulos and K. A. Olive, JCAP 10, 009 (2013) doi:10.1088/1475- 7516/2013/10/009 [arXiv:1307.3537 [hep-th]]

    2013 arXiv

  5. [12]

    Ellis, D

    J. Ellis, D. V. Nanopoulos, K. A. Olive and S. Verner, JHEP 03, 099 (2019) doi:10.1007/JHEP03(2019)099 [arXiv:1812.02192 [hep-th]]

    2019 arXiv

  6. [13]

    G. K. Chakravarty and S. Mohanty, Phys. Lett. B 746, 242-247 (2015) doi:10.1016/j.physletb.2015.04.056 [arXiv:1405.1321 [hep-ph]]

    2015 arXiv

  7. [14]

    Motohashi, Phys

    H. Motohashi, Phys. Rev. D 91, 064016 (2015) doi:10.1103/PhysRevD.91.064016 [arXiv:1411.2972 [astro-ph.CO]]

    2015 arXiv

  8. [15]

    S. D. Odintsov and V. K. Oikonomou, Int. J. Mod. Phys. D 32, no.01, 2250135 (2023) doi:10.1142/S0218271822501358 [arXiv:2210.11351 [gr-qc]]

    2023 arXiv

  9. [16]

    S. Meza, D. Altamirano, M. Z. Mughal and C. Rojas, Int. J. Mod . Phys. D 30, no.08, 2150062 (2021) doi:10.1142/S0218271821500620 [arXiv:2104.01139 [gr-qc ]]

    2021 arXiv

  10. [18]

    P. A. R. Ade et al. [BICEP and Keck], Phys. Rev. Lett. 127, no.15, 151301 (2021) doi:10.1103/PhysRevLett.127.151301 [arXiv:2110.00483 [astro-ph.C O]]

    2021

  11. [19]

    V. F. Mukhanov, Sov. Phys. JETP 67, 1297-1302 (1988)

    1988

  12. [20]

    Sasaki, Prog

    M. Sasaki, Prog. Theor. Phys. 76, 1036 (1986) doi:10.1143/PTP.76.1036

    1986 doi

  13. [21]

    B. A. Bassett, S. Tsujikawa and D. Wands, Rev. Mod. Phys. 78, 537-589 (2006) doi:10.1103/RevModPhys.78.537 [arXiv:astro-ph/0507632 [astro-p h]]

    2006 arXiv

  14. [22]

    M. J. Mortonson, H. V. Peiris and R. Easther, Phys. Rev. D 83, 043505 (2011) doi:10.1103/PhysRevD.83.043505 [arXiv:1007.4205 [astro-ph.CO]]

    2011 arXiv

  15. [23]

    Lewis, A

    A. Lewis, A. Challinor and A. Lasenby, Astrophys. J. 538, 473-476 (2000) doi:10.1086/309179 [arXiv:astro-ph/9911177 [astro-ph]]

    2000 arXiv

  16. [24]

    Lewis and S

    A. Lewis and S. Bridle, Phys. Rev. D 66, 103511 (2002) doi:10.1103/PhysRevD.66.103511 [arXiv:astro-ph/0205436 [astro-ph]]

    2002 arXiv

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.