pith. sign in

arxiv: 2509.13220 · v2 · submitted 2025-09-16 · 🌌 astro-ph.CO

Dynamic or Systematic? Bayesian model selection between dark energy and supernova biases

A.N. Ormondroyd , W.J. Handley , M.P. Hobson , A.N. Lasenby , D. Yallup This is my paper

Pith reviewed 2026-05-18 16:02 UTC · model grok-4.3

classification 🌌 astro-ph.CO
keywords dark energysupernovaeBayesian model selectionsystematic biasesDES-5YDESI BAOflexknot
0
0 comments X

The pith

A magnitude offset between low- and high-redshift supernovae beats evidence for dynamical dark energy.

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

The paper tests whether the apparent support for evolving dark energy from DES-5Y supernovae plus DESI BAO data comes from new physics or from a data-processing bias. It shows that adding one magnitude offset between the low-redshift legacy supernovae and the high-redshift DES sample produces stronger Bayesian evidence than the flexknot parametrization of dark energy. The same offset also brings the supernova and BAO measurements into closer agreement. The authors reach these conclusions by direct model comparison using Bayes factors, and they test an alternative numerical method for computing those factors.

Core claim

The Bayesian evidence previously reported for flexknot dark energy is exceeded by a model that includes a single magnitude offset between low- and high-redshift supernovae; the offset also substantially reduces the tension between the DES-5Y supernova sample and DESI BAO measurements.

What carries the argument

Bayes factor comparison between dark-energy parametrizations and a systematic-magnitude-offset model applied to supernova distance moduli.

If this is right

  • The current DES-5Y plus DESI preference for dynamical dark energy weakens once the offset is allowed.
  • The apparent tension between the two datasets shrinks under the offset model.
  • Bayesian model selection can be used to decide between physical dark-energy extensions and simple calibration adjustments in future surveys.

Where Pith is reading between the lines

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

  • If the offset traces to selection biases in the legacy low-redshift sample, similar adjustments may be needed for other combined supernova analyses.
  • The result suggests that apparent deviations from a cosmological constant in current data sets should be checked against straightforward calibration differences before invoking new physics.
  • The trial of Nested Bridge Sampling with Sequential Monte Carlo offers a practical route for computing Bayes factors when standard methods become computationally heavy.

Load-bearing premise

That one constant magnitude shift between low- and high-redshift supernovae is enough to capture the main pipeline systematic without needing a full re-modeling of selection effects.

What would settle it

A re-analysis of the DES supernova pipeline that finds no magnitude offset of the size required to remove the dynamical-dark-energy preference, or new supernova data processed without the offset that still favors evolving dark energy.

Figures

Figures reproduced from arXiv: 2509.13220 by A.N. Lasenby, A.N. Ormondroyd, D. Yallup, M.P. Hobson, W.J. Handley.

Figure 1
Figure 1. Figure 1: shows prior samples from the usual CPL prior used in most analyses, and the equivalent flexknot prior, and the correspond￾ing posteriors from DES-5Y combined with DESI as kernel density estimates (KDEs). The CPL prior includes the constraint that the value of 𝑤 at early times, 𝑤0 + 𝑤𝑎, is less than zero, to ensure a period of matter domination. Clearly, the two priors are different, but the posteriors are … view at source ↗
Figure 2
Figure 2. Figure 2: Flexknot reconstruction of the dark energy equation of state pa￾rameter using DES-5Y supernovae only. In red is the standard likelihood, in lilac, the version with the Δ𝑚B offset for the low-redshift supernovae. The overall shape of the reconstructions are very similar, but the functional KL divergence and model evidences tell quite different stories. Firstly, note that the high-𝑎/low-redshift KL divergenc… view at source ↗
Figure 4
Figure 4. Figure 4: Tension values between DES-5Y and DESI BAO, with and without the low-redshift offset. ΛCDM is shown as horizontal dashed lines, for easy comparison with the other points. For ΛCDM the tension has been reduced (more positive) significantly, while for 𝑤CDM and CPL, it has increased slightly. ΛCDM with the offset is now on-par with CPL, though 𝑤CDM remains the model with the best dataset concordance. In contr… view at source ↗
Figure 3
Figure 3. Figure 3: Similar to [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 5
Figure 5. Figure 5: Posterior histograms of Δ𝑚B for ΛCDM and CPL. The left panel uses DES-5Y only, the right also includes DESI BAO. The prior is uniform over the domain of the plot. As predicted by Efstathiou (2024), its value is centred on −0.04. Note that the ΛCDM posteriors (and the right CPL posterior) are well contained within the prior, therefore, the evidence which would have been found had a wider prior been used can… view at source ↗
read the original abstract

DES-5Y supernovae, combined with DESI BAO, appear to favour Chevallier-Polarski-Linder $(w_0, w_a)$ dynamical dark energy over $\Lambda$CDM. arXiv:2408.07175 suggested that this is driven by a systematic in the DES pipeline, which particularly affects the low-redshift supernovae brought in from legacy surveys. It is difficult to investigate these data in isolation, however, as the complicated supernovae pipelines must properly account for selection effects. In this work, we discover that the Bayesian evidence previously found for flexknot dark energy (arXiv:2503.17342) is beaten by a magnitude offset between the low- and high-redshift supernovae. In addition, we find that the possible tension between DES-5Y and DESI is significantly reduced by such an offset. We also take the opportunity to trial Nested Bridge Sampling with Sequential Monte Carlo as an alternative method for calculating Bayes factors.

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.

Referee Report

2 major / 2 minor

Summary. The manuscript claims that a single constant magnitude offset between low- and high-redshift supernovae yields higher Bayesian evidence than the flexknot dynamical dark-energy model when fitting DES-5Y supernovae together with DESI BAO data; the offset also reduces apparent tension between the two datasets. It further reports a trial of Nested Bridge Sampling with Sequential Monte Carlo as an alternative route to Bayes factors.

Significance. If the central result holds after addressing the modeling assumptions, the work would strengthen the case that apparent dynamical-dark-energy signals in current supernova compilations can be reinterpreted as low-redshift systematics rather than new physics, with direct bearing on the interpretation of DES-5Y/DESI tensions. The explicit comparison of evidence ratios and the demonstration of an alternative nested-sampling technique constitute reproducible methodological contributions.

major comments (2)
  1. [§3 and §4] §3 (Model definitions) and §4 (Evidence comparison): the claim that a single magnitude offset fully proxies the pipeline systematic identified in arXiv:2408.07175 is load-bearing for the model-selection conclusion, yet the manuscript provides no test of whether a redshift-dependent bias (e.g., Malmquist or host-galaxy selection) would restore the flexknot evidence once the constant offset is replaced by a more flexible systematic model.
  2. [Abstract and §4.2] Abstract and §4.2 (Bayes-factor results): the reported evidence comparison is stated without quantitative values for the Bayes factor, the prior width adopted for the magnitude offset, or the precise likelihood construction (including how the offset enters the distance-modulus term), preventing verification that the data actually support the ranking as claimed.
minor comments (2)
  1. [Figure 2] Figure 2 caption: the legend does not specify the exact prior ranges used for the offset parameter in the two models being compared.
  2. [§5] §5 (Discussion): the statement that the offset 'significantly reduces' tension would benefit from a quantitative tension metric (e.g., parameter-shift or evidence ratio) rather than a qualitative description.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive and detailed report. We address each major comment below and outline the revisions we will make to strengthen the manuscript.

read point-by-point responses

  1. Referee: [§3 and §4] §3 (Model definitions) and §4 (Evidence comparison): the claim that a single magnitude offset fully proxies the pipeline systematic identified in arXiv:2408.07175 is load-bearing for the model-selection conclusion, yet the manuscript provides no test of whether a redshift-dependent bias (e.g., Malmquist or host-galaxy selection) would restore the flexknot evidence once the constant offset is replaced by a more flexible systematic model.

    Authors: We thank the referee for this observation. Our manuscript does not claim that a constant offset fully captures every detail of the DES pipeline systematic reported in arXiv:2408.07175; rather, we show that even this minimal, physically motivated correction already produces higher Bayesian evidence than the flexknot model and reduces the DES-5Y/DESI tension. We agree that testing a redshift-dependent bias is a valuable robustness check. In the revised manuscript we will add an explicit comparison replacing the constant offset with a linear redshift-dependent magnitude bias and report the resulting evidence ratios. revision: yes

  2. Referee: [Abstract and §4.2] Abstract and §4.2 (Bayes-factor results): the reported evidence comparison is stated without quantitative values for the Bayes factor, the prior width adopted for the magnitude offset, or the precise likelihood construction (including how the offset enters the distance-modulus term), preventing verification that the data actually support the ranking as claimed.

    Authors: We agree that these quantitative details are essential for reproducibility and verification. In the revised version we will (i) report the numerical Bayes factors (both in the abstract and §4.2), (ii) specify the exact prior width and functional form adopted for the magnitude-offset parameter, and (iii) provide a clear equation showing how the offset is added to the distance modulus inside the likelihood for the low-redshift supernovae. revision: yes

Circularity Check

0 steps flagged

Bayesian evidence comparison is self-contained with no circular reduction

full rationale

The paper computes Bayesian evidences for a flexknot dark-energy model versus a model that adds a single magnitude offset nuisance parameter between low- and high-redshift supernovae, using DES-5Y and DESI BAO data. This is a standard model-selection procedure whose evidence ratio is determined by explicit likelihoods and priors rather than by construction from the inputs. The offset is introduced to test the systematic hypothesis referenced from arXiv:2408.07175; the comparison itself does not reduce to a fit renamed as a prediction or to a self-citation chain that forces the result. The additional trial of Nested Bridge Sampling with Sequential Monte Carlo is an independent numerical check. No load-bearing step equates the claimed outcome to its own definition or fitted values.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The central claim rests on the validity of Bayesian evidence for model comparison and the assumption that a single magnitude offset adequately represents the systematic bias.

free parameters (1)
  • magnitude offset
    A free parameter introduced to model the difference between low- and high-redshift supernova magnitudes.
axioms (1)
  • domain assumption Bayesian evidence ranks models correctly when priors and likelihoods are properly specified
    Invoked when claiming the offset model beats flexknot dark energy.

pith-pipeline@v0.9.0 · 5725 in / 1188 out tokens · 45685 ms · 2026-05-18T16:02:00.819942+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Information-Geometric Perspective on the Hubble Tension: Eigenmode Rotation and Curvature Suppression in wCDM

    astro-ph.CO 2026-04 unverdicted novelty 7.0

    Extending to wCDM mainly suppresses the leading Planck Fisher eigenvalue to 2.7% of its LambdaCDM value with only modest eigenmode rotation, while late-time data adds curvature that limits tension relief.

  2. Nested Sampling for ARIMA Model Selection in Astronomical Time-Series Analysis

    astro-ph.IM 2025-12 unverdicted novelty 6.0

    Nested sampling applied to ARIMA models enables Bayesian order selection and parameter inference that recovers ground truth in simulations and fits stochastic variability in sunspot, Kepler, and TESS light curves.

Reference graph

Works this paper leans on

62 extracted references · 62 canonical work pages · cited by 2 Pith papers · 7 internal anchors

  1. [1]

    C., Abazajian K

    Aslanyan G., Price L. C., Abazajian K. N., Easther R., 2014, @doi [ ] 10.1088/1475-7516/2014/08/052 , https://ui.adsabs.harvard.edu/abs/2014JCAP...08..052A 2014, 052

    work page doi:10.1088/1475-7516/2014/08/052 2014

  2. [2]

    Bayes T., 1763, Philosophical Transactions of the Royal Society of London , 53, 370

    work page

  3. [3]

    Bradbury J., et al., 2018, JAX : composable transformations of P ython+ N um P y programs, http://github.com/jax-ml/jax

    work page 2018

  4. [4]

    Brout D., et al., 2022, @doi [ ] 10.3847/1538-4357/ac8e04 , https://ui.adsabs.harvard.edu/abs/2022ApJ...938..110B 938, 110

    work page doi:10.3847/1538-4357/ac8e04 2022

  5. [5]

    Cabezas A., Corenflos A., Lao J., Louf R., 2024, BlackJAX: Composable B ayesian inference in JAX ( @eprint arXiv 2402.10797 )

    work page arXiv 2024

  6. [6]

    G., 2000, Monte Carlo Methods in Bayesian Computation, 1 edn

    Chen M.-H., Shao Q.-M., Ibrahim J. G., 2000, Monte Carlo Methods in Bayesian Computation, 1 edn. Springer Series in Statistics, Springer New York, NY, @doi 10.1007/978-1-4612-1276-8

    work page doi:10.1007/978-1-4612-1276-8 2000

  7. [7]

    Chevallier M., Polarski D., 2001, @doi [Int. J. Mod. Phys. D] 10.1142/S0218271801000822 , 10, 213

    work page doi:10.1142/s0218271801000822 2001

  8. [8]

    DES Collaboration et al., 2024, @doi [ ] 10.3847/2041-8213/ad6f9f , https://ui.adsabs.harvard.edu/abs/2024ApJ...973L..14D 973, L14

    work page doi:10.3847/2041-8213/ad6f9f 2024

  9. [9]

    DESI 2024 VI: Cosmological Constraints from the Measurements of Baryon Acoustic Oscillations

    DESI Collaboration et al., 2024, @doi [arXiv e-prints] 10.48550/arXiv.2404.03002 , https://ui.adsabs.harvard.edu/abs/2024arXiv240403002D p. arXiv:2404.03002

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2404.03002 2024

  10. [10]

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

    DESI Collaboration et al., 2025a, arXiv e-prints, https://ui.adsabs.harvard.edu/abs/2025arXiv250314738D p. arXiv:2503.14738

    work page internal anchor Pith review Pith/arXiv arXiv

  11. [11]

    DESI DR2 Results I: Baryon Acoustic Oscillations from the Lyman Alpha Forest

    DESI Collaboration et al., 2025b, arXiv e-prints, https://ui.adsabs.harvard.edu/abs/2025arXiv250314739D p. arXiv:2503.14739

    work page internal anchor Pith review Pith/arXiv arXiv

  12. [12]

    Extended Dark Energy analysis using DESI DR2 BAO measurements

    DESI Collaboration et al., 2025c, arXiv e-prints, https://ui.adsabs.harvard.edu/abs/2025arXiv250314743D p. arXiv:2503.14743

    work page internal anchor Pith review Pith/arXiv arXiv

  13. [13]

    Efstathiou,Evolving Dark Energy or Supernovae Systematics?, arXiv:2408.07175 (2025)

    Efstathiou G., 2024, @doi [arXiv e-prints] 10.48550/arXiv.2408.07175 , https://ui.adsabs.harvard.edu/abs/2024arXiv240807175E p. arXiv:2408.07175

    work page doi:10.48550/arxiv.2408.07175 2024

  14. [14]

    Einstein A., 1917, Sitzungsberichte der Königlich Preussischen Akademie der Wissenschaften, https://ui.adsabs.harvard.edu/abs/1917SPAW.......142E pp 142--152

    work page 1917

  15. [15]

    Finelli F., et al., 2018, @doi [ ] 10.1088/1475-7516/2018/04/016 , https://ui.adsabs.harvard.edu/abs/2018JCAP...04..016F 2018, 016

    work page doi:10.1088/1475-7516/2018/04/016 2018

  16. [16]

    J., et al., 2017, @doi [Monthly Notices of the Royal Astronomical Society] 10.1093/mnras/stx3136 , 475, 193

    Foley R. J., et al., 2017, @doi [Monthly Notices of the Royal Astronomical Society] 10.1093/mnras/stx3136 , 475, 193

    work page doi:10.1093/mnras/stx3136 2017

  17. [17]

    A Tutorial on Bridge Sampling

    Gronau Q. F., et al., 2017, @doi [arXiv e-prints] 10.48550/arXiv.1703.05984 , https://ui.adsabs.harvard.edu/abs/2017arXiv170305984G p. arXiv:1703.05984

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1703.05984 2017

  18. [18]

    Guy J., et al., 2007, @doi [ ] 10.1051/0004-6361:20066930 , https://ui.adsabs.harvard.edu/abs/2007A&A...466...11G 466, 11

    work page doi:10.1051/0004-6361:20066930 2007

  19. [19]

    Handley W., 2019, @doi [The Journal of Open Source Software] 10.21105/joss.01414 , 4, 1414

    work page doi:10.21105/joss.01414 2019

  20. [20]

    Handley W., Lemos P., 2019, @doi [ ] 10.1103/PhysRevD.100.043504 , https://ui.adsabs.harvard.edu/abs/2019PhRvD.100d3504H 100, 043504

    work page doi:10.1103/physrevd.100.043504 2019

  21. [21]

    J., Hobson M

    Handley W. J., Hobson M. P., Lasenby A. N., 2015a, @doi [ ] 10.1093/mnrasl/slv047 , https://ui.adsabs.harvard.edu/abs/2015MNRAS.450L..61H 450, L61

    work page doi:10.1093/mnrasl/slv047

  22. [22]

    J., Hobson M

    Handley W. J., Hobson M. P., Lasenby A. N., 2015b, @doi [ ] 10.1093/mnras/stv1911 , https://ui.adsabs.harvard.edu/abs/2015MNRAS.453.4384H 453, 4384

    work page doi:10.1093/mnras/stv1911

  23. [23]

    J., Lasenby A

    Handley W. J., Lasenby A. N., Peiris H. V., Hobson M. P., 2019, @doi [ ] 10.1103/PhysRevD.100.103511 , https://ui.adsabs.harvard.edu/abs/2019PhRvD.100j3511H 100, 103511

    work page doi:10.1103/physrevd.100.103511 2019

  24. [24]

    Harris, K

    Harris C. R., et al., 2020, @doi [Nature] 10.1038/s41586-020-2649-2 , 585, 357

    work page doi:10.1038/s41586-020-2649-2 2020

  25. [25]

    J., Hobson M

    Hee S., Handley W. J., Hobson M. P., Lasenby A. N., 2016, @doi [ ] 10.1093/mnras/stv2217 , https://ui.adsabs.harvard.edu/abs/2016MNRAS.455.2461H 455, 2461

    work page doi:10.1093/mnras/stv2217 2016

  26. [26]

    S., Fialkov A., Lorimer D

    Heimersheim S., Sartorio N. S., Fialkov A., Lorimer D. R., 2022, @doi [ ] 10.3847/1538-4357/ac70c9 , https://ui.adsabs.harvard.edu/abs/2022ApJ...933...57H 933, 57

    work page doi:10.3847/1538-4357/ac70c9 2022

  27. [27]

    Heimersheim S., R nneberg L., Linton H., Pagani F., Fialkov A., 2024, @doi [ ] 10.1093/mnras/stad3936 , https://ui.adsabs.harvard.edu/abs/2024MNRAS.52711404H 527, 11404

    work page doi:10.1093/mnras/stad3936 2024

  28. [28]

    T., Handley W

    Hergt L. T., Handley W. J., Hobson M. P., Lasenby A. N., 2021, @doi [ ] 10.1103/PhysRevD.103.123511 , https://ui.adsabs.harvard.edu/abs/2021PhRvD.103l3511H 103, 123511

    work page doi:10.1103/physrevd.103.123511 2021

  29. [29]

    Hicken M., et al., 2009, @doi [The Astrophysical Journal] 10.1088/0004-637X/700/1/331 , 700, 331

    work page doi:10.1088/0004-637x/700/1/331 2009

  30. [30]

    Hicken M., et al., 2012, @doi [The Astrophysical Journal Supplement Series] 10.1088/0067-0049/200/2/12 , 200, 12

    work page doi:10.1088/0067-0049/200/2/12 2012

  31. [31]

    D., 2007, @doi [Computing in Science and Engineering] 10.1109/MCSE.2007.55 , 9, 90

    Hunter J. D., 2007, @doi [Computing in Science & Engineering] 10.1109/MCSE.2007.55 , 9, 90

    work page doi:10.1109/mcse.2007.55 2007

  32. [32]

    Clarendon Press, Oxford

    Jeffreys H., 1939, The Theory of Probability, second edn. Clarendon Press, Oxford

    work page 1939

  33. [33]

    D., et al., 2021, @doi [The Astrophysical Journal] 10.3847/1538-4357/ac30d8 , 923, 265

    Kenworthy W. D., et al., 2021, @doi [The Astrophysical Journal] 10.3847/1538-4357/ac30d8 , 923, 265

    work page doi:10.3847/1538-4357/ac30d8 2021

  34. [34]

    Krisciunas K., et al., 2017, @doi [The Astronomical Journal] 10.3847/1538-3881/aa9a3d , 154, 278

    work page doi:10.3847/1538-3881/aa9a3d 2017

  35. [35]

    Krisciunas K., et al., 2020, @doi [The Astronomical Journal] 10.3847/1538-3881/abc431 , 160, 289

    work page doi:10.3847/1538-3881/abc431 2020

  36. [36]

    Li J., 2023, AstroJacobLi/smplotlib: v0.0.9, @doi 10.5281/zenodo.8126529 , https://doi.org/10.5281/zenodo.8126529

    work page doi:10.5281/zenodo.8126529 2023

  37. [37]

    Exploring the expansion history of the universe,

    Linder E. V., 2003, @doi [Phys. Rev. Lett.] 10.1103/PhysRevLett.90.091301 , 90, 091301

    work page doi:10.1103/physrevlett.90.091301 2003

  38. [38]

    2010, in Proceedings of the 9th Python in Science Conference, ed

    M c K inney W., 2010, in S t\'efan van der W alt J arrod M illman eds, P roceedings of the 9th P ython in S cience C onference. pp 56 -- 61, @doi 10.25080/Majora-92bf1922-00a

    work page doi:10.25080/majora-92bf1922-00a 2010

  39. [39]

    Millea M., Bouchet F., 2018, @doi [ ] 10.1051/0004-6361/201833288 , https://ui.adsabs.harvard.edu/abs/2018A&A...617A..96M 617, A96

    work page doi:10.1051/0004-6361/201833288 2018

  40. [40]

    and Lindsten, Fredrik and Sch

    Naesseth C. A., Lindsten F., Sch \"o n T. B., 2019, @doi [arXiv e-prints] 10.48550/arXiv.1903.04797 , https://ui.adsabs.harvard.edu/abs/2019arXiv190304797N p. arXiv:1903.04797

    work page doi:10.48550/arxiv.1903.04797 2019

  41. [41]

    Notari A., Redi M., Tesi A., 2025, @doi [ ] 10.1088/1475-7516/2025/04/048 , https://ui.adsabs.harvard.edu/abs/2025JCAP...04..048N 2025, 048

    work page doi:10.1088/1475-7516/2025/04/048 2025

  42. [42]

    O'Raifeartaigh C., O'Keeffe M., Nahm W., Mitton S., 2017, @doi [European Physical Journal H] 10.1140/epjh/e2017-80002-5 , https://ui.adsabs.harvard.edu/abs/2017EPJH...42..431O 42

    work page doi:10.1140/epjh/e2017-80002-5 2017

  43. [43]

    P., Feroz F., Grainge K

    Olamaie M., Hobson M. P., Feroz F., Grainge K. J. B., Lasenby A., Perrott Y. C., Rumsey C., Saunders R. D. E., 2018, @doi [ ] 10.1093/mnras/sty2495 , https://ui.adsabs.harvard.edu/abs/2018MNRAS.481.3853O 481, 3853

    work page doi:10.1093/mnras/sty2495 2018

  44. [44]

    Ormondroyd A., 2025, Nonparametric reconstructions of dynamical dark energy using flexknots , @doi 10.5281/zenodo.15025604 , https://doi.org/10.5281/zenodo.15025604

    work page doi:10.5281/zenodo.15025604 2025

  45. [45]

    N., Handley W

    Ormondroyd A. N., Handley W. J., Hobson M. P., Lasenby A. N., 2023, @doi [arXiv e-prints] 10.48550/arXiv.2310.08490 , https://ui.adsabs.harvard.edu/abs/2023arXiv231008490O p. arXiv:2310.08490

    work page doi:10.48550/arxiv.2310.08490 2023

  46. [46]

    N., Handley W

    Ormondroyd A. N., Handley W. J., Hobson M. P., Lasenby A. N., 2025a, @doi [arXiv e-prints] 10.48550/arXiv.2503.17342 , https://ui.adsabs.harvard.edu/abs/2025arXiv250317342O p. arXiv:2503.17342

    work page doi:10.48550/arxiv.2503.17342

  47. [47]

    N., Handley W

    Ormondroyd A. N., Handley W. J., Hobson M. P., Lasenby A. N., 2025b, @doi [ ] 10.1093/mnras/staf1144 , https://ui.adsabs.harvard.edu/abs/2025MNRAS.541.3388O 541, 3388

    work page doi:10.1093/mnras/staf1144

  48. [48]

    Perlmutter S., et al., 1999, @doi [ ] 10.1086/307221 , https://ui.adsabs.harvard.edu/abs/1999ApJ...517..565P 517, 565

    work page internal anchor Pith review doi:10.1086/307221 1999

  49. [49]

    Planck Collaboration et al., 2014, @doi [ ] 10.1051/0004-6361/201321569 , https://ui.adsabs.harvard.edu/abs/2014A&A...571A..22P 571, A22

    work page doi:10.1051/0004-6361/201321569 2014

  50. [50]

    Planck Collaboration et al., 2016a, @doi [ ] 10.1051/0004-6361/201525830 , https://ui.adsabs.harvard.edu/abs/2016A&A...594A..13P 594, A13

    work page doi:10.1051/0004-6361/201525830

  51. [51]

    Planck Collaboration et al., 2016b, @doi [ ] 10.1051/0004-6361/201525898 , https://ui.adsabs.harvard.edu/abs/2016A&A...594A..20P 594, A20

    work page doi:10.1051/0004-6361/201525898

  52. [52]

    Planck Collaboration et al., 2021, @doi [ ] 10.1051/0004-6361/201833910e , https://ui.adsabs.harvard.edu/abs/2021A&A...652C...4P 652, C4

    work page doi:10.1051/0004-6361/201833910e 2021

  53. [53]

    Prathaban, D

    Prathaban M., Yallup D., Alvey J., Yang M., Templeton W., Handley W., 2025, @doi [arXiv e-prints] 10.48550/arXiv.2509.04336 , https://ui.adsabs.harvard.edu/abs/2025arXiv250904336P p. arXiv:2509.04336

    work page doi:10.48550/arxiv.2509.04336 2025

  54. [54]

    G., et al., 1998, @doi [AJ] 10.1086/300499 , http://adsabs.harvard.edu/abs/1998AJ....116.1009R 116, 1009

    Riess A. G., et al., 1998, @doi [ ] 10.1086/300499 , https://ui.adsabs.harvard.edu/abs/1998AJ....116.1009R 116, 1009

    work page internal anchor Pith review doi:10.1086/300499 1998

  55. [55]

    Shen E., Anstey D., de Lera Acedo E., Fialkov A., 2024, @doi [ ] 10.1093/mnras/stae614 , https://ui.adsabs.harvard.edu/abs/2024MNRAS.529.1642S 529, 1642

    work page doi:10.1093/mnras/stae614 2024

  56. [56]

    Taylor G., et al., 2023, @doi [Monthly Notices of the Royal Astronomical Society] 10.1093/mnras/stad320 , 520, 5209

    work page doi:10.1093/mnras/stad320 2023

  57. [57]

    The pandas development team 2023, pandas-dev/pandas: Pandas, @doi 10.5281/zenodo.8092754 , https://doi.org/10.5281/zenodo.8092754

    work page doi:10.5281/zenodo.8092754 2023

  58. [58]

    A., Bridges M., Hobson M

    V \'a zquez J. A., Bridges M., Hobson M. P., Lasenby A. N., 2012a, @doi [ ] 10.1088/1475-7516/2012/06/006 , https://ui.adsabs.harvard.edu/abs/2012JCAP...06..006V 2012, 006

    work page doi:10.1088/1475-7516/2012/06/006 2012

  59. [59]

    A., Bridges M., Hobson M

    V \'a zquez J. A., Bridges M., Hobson M. P., Lasenby A. N., 2012b, @doi [ ] 10.1088/1475-7516/2012/09/020 , https://ui.adsabs.harvard.edu/abs/2012JCAP...09..020V 2012, 020

    work page doi:10.1088/1475-7516/2012/09/020 2012

  60. [60]

    Vincenzi M., et al., 2025, @doi [ ] 10.1093/mnras/staf943 , https://ui.adsabs.harvard.edu/abs/2025MNRAS.541.2585V 541, 2585

    work page doi:10.1093/mnras/staf943 2025

  61. [61]

    Virtanen P., et al., 2020, @doi [Nature Methods] 10.1038/s41592-019-0686-2 , https://rdcu.be/b08Wh 17, 261

    work page doi:10.1038/s41592-019-0686-2 2020

  62. [62]

    https://openreview.net/forum?id=ekbkMSuPo4

    Yallup D., Kroupa N., Handley W., 2025, in Frontiers in Probabilistic Inference: Learning meets Sampling. https://openreview.net/forum?id=ekbkMSuPo4

    work page 2025