REVIEW
ECH0EZ: Estimating the Hubble constant with ZTF SN Ia DR2
Cosmology and Nongalactic Astrophysics · Submitted 26 Oct 2026 · 46 references
Pith's verdict: First H0 from ZTF DR2 SNe: a careful, honestly caveated analysis whose main risk — the Pan-STARRS cross-calibration — is exactly where the authors put it.
Why Pith says this
Highlight any text to comment on it.
Referee report
The paper
The paper claims
A homogeneous supernova sample from ZTF measures H0 = 71.6 ± 1.5 km/s/Mpc, consistent with local distance-ladder values but in 2.7σ tension with Planck.
The reading
What carries the argument
The key machinery is the Simple-BayeSN Bayesian hierarchical model (Mandel et al. 2017), which simultaneously fits the SALT light-curve parameters, intrinsic scatter, dust extinction and dust-law parameters, population distributions of stretch and colour, and cosmological distances. The critical data are the 772 ZTF DR2 type Ia supernovae in the volume-limited subset (z < 0.06), which is designed to be free of systematic selection biases, and the 42 calibrators from SH0ES. The photometric calibration is corrected using average offsets from Lacroix et al. (2025) and Rigault et al. (2025), and peculiar velocities are corrected using the Carrick et al. (2015) map.
Extended reading notes
Core claim
The central discovery is a new measurement of the Hubble constant from a homogeneous, untargeted sample of type Ia supernovae, nearly free of systematic selection biases. Using the volume-limited subset (z < 0.06) of ZTF DR2, together with 42 SH0ES calibrators and a Bayesian hierarchical model, the authors obtain H0 = 71.6 ± 1.5 km/s/Mpc. The uncertainty is dominated by the number of calibrators. The measurement is consistent with the SH0ES value and in tension with Planck, reinforcing the conclusion that the Hubble tension cannot be attributed to inhomogeneous supernova samples or simplified modeling.
Load-bearing premise
The cross-calibration between ZTF and Pan-STARRS photometry is assumed to be correctly accounted for by the average corrections from Lacroix et al. (2025) and Rigault et al. (2025); any residual systematic offset in the photometric calibration between Hubble-flow and calibrator samples propagates directly into H0.
Editorial extensions
If this is right
- If correct, the measurement implies that the Hubble tension is not resolved by supernova data alone, pushing the source of the discrepancy to the Cepheid calibration, geometric anchors, or new physics beyond ΛCDM.
- The homogeneous ZTF sample validates the potential of single-telescope, untargeted surveys for precision cosmology, reducing selection biases that have plagued previous compilations.
- The result sets a benchmark for future ZTF data releases with improved calibration, which could reduce systematic uncertainties and sharpen the tension with Planck.
- The dominance of calibrator statistics in the error budget suggests that a dedicated ZTF calibrator program could substantially reduce the H0 uncertainty.
Reading between the lines
- An unaccounted cross-calibration offset between ZTF and Pan-STARRS as large as ~0.13 mag could shift the inferred H0 toward the Planck value, as the authors acknowledge; this provides a concrete systematic path to check with improved photometry.
- The similarity between the ZTF+H0 result and SH0ES, despite the very different Hubble-flow sample, strengthens the case that the Hubble tension is not due to supernova selection effects, focusing attention on the calibrator rung of the distance ladder.
- Future work could test this result by replacing the SH0ES calibrators with ZTF-observed calibrators, which would eliminate the cross-calibration systematic between the two samples.
- The weak constraint on q0 (deceleration) from ZTF alone highlights the need for higher-redshift homogeneous samples to break the H0–q0 degeneracy without external priors.
Original abstract
Type Ia supernovae (SNe Ia) are instrumental for a precise measurement of the Hubble constant ($H_0$) from the local distance ladder, yet the data used hitherto for this purpose are inhomogeneous and subject to severe selection effects. We present the first, preliminary, cosmological analysis of 772 SNe Ia from the Zwicky Transient Facility's second data release: an untargeted single-telescope sample whose volume-limited subset (z<0.06) is largely free of Malmquist bias. We correct the public data for known instrumental and calibration biases and the redshifts for peculiar motions. Using the 42 SH0ES calibrators (their host distances computed from Cepheids and geometric anchors) and a Bayesian hierarchical model of SN Ia standardisation, dust extinction, redshift and calibration uncertainties, we infer $H_0 = (71.6 \pm 1.5)$ km/s/Mpc after marginalising over deceleration with a flat prior (or $(72.3 \pm 1.1)$ km/s/Mpc if a prior on matter density from Pantheon+ is imposed), with the uncertainty dominated by the number of calibrators. We demonstrate robustness to the choice of light-curve and Milky Way dust models and to extensions of the model to include peculiar-velocity correlations, a host-mass step, and two stretch populations. These results are in agreement with previous distance-ladder measurements and $2.7\sigma$ ($4.1\sigma$) away from the value obtained from Planck data, suggesting that the resolution of the Hubble tension lies outside of SNe Ia---or the distance ladder entirely.
The compiled record
The paper's own sentences, arranged, kept beside the review and separate from it. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.
The abstract, sentence by sentence
Type Ia supernovae (SNe Ia) are instrumental for a precise measurement of the Hubble constant ($H_0$) from the local distance ladder, yet the data used hitherto for this purpose are inhomogeneous and subject to severe selection effects. We present the first, preliminary, cosmological analysis of 772 SNe Ia from the Zwicky Transient Facility's second data release: an untargeted single-telescope sample whose volume-limited subset (z<0.06) is largely free of Malmquist bias. We correct the public data for known instrumental and calibration biases and the redshifts for peculiar motions. Using the 42 SH0ES calibrators (their host distances computed from Cepheids and geometric anchors) and a Bayesian hierarchical model of SN Ia standardisation, dust extinction, redshift and calibration uncertainties, we infer $H_0 = (71.6 \pm 1.5)$ km/s/Mpc after marginalising over deceleration with a flat prior (or $(72.3 \pm 1.1)$ km/s/Mpc if a prior on matter density from Pantheon+ is imposed), with the uncertainty dominated by the number of calibrators. We demonstrate robustness to the choice of light-curve and Milky Way dust models and to extensions of the model to include peculiar-velocity correlations, a host-mass step, and two stretch populations. These results are in agreement with previous distance-ladder measurements and $2.7\sigma$ ($4.1\sigma$) away from the value obtained from Planck data, suggesting that the resolution of the Hubble tension lies outside of SNe Ia---or the distance ladder entirely.
background contribution method result interpretation main claim
| # | sentence | role | the authors' own work | states the main conclusion | strength |
|---|---|---|---|---|---|
| s0 | Type Ia supernovae (SNe Ia) are instrumental for a precise measurement of the Hubble constant ($H_0$) from the local distance ladder, yet the data used hitherto for this purpose are inhomogeneous and subject to severe selection effects. | background 0.84 | 0.06 | 0.03 | asserted 1.00 |
| s1 | We present the first, preliminary, cosmological analysis of 772 SNe Ia from the Zwicky Transient Facility's second data release: an untargeted single-telescope sample whose volume-limited subset (z<0.06) is largely free of Malmquist bias. | contribution 1.00 | 0.96 | 0.09 | asserted 1.00 |
| s2 | We correct the public data for known instrumental and calibration biases and the redshifts for peculiar motions. | method 1.00 | 0.92 | 0.02 | asserted 0.99 |
| s3 | Using the 42 SH0ES calibrators (their host distances computed from Cepheids and geometric anchors) and a Bayesian hierarchical model of SN Ia standardisation, dust extinction, redshift and calibration uncertainties, we infer $H_0 = (71.6 \pm 1.5)$ km/s/Mpc after marginalising over deceleration with a flat prior (or $(72.3 \pm 1.1)$ km/s/Mpc if a prior on matter density from Pantheon+ is imposed), with the uncertainty dominated by the number of calibrators. | result 1.00 | 0.97 | 0.84 | asserted 1.00 |
| s4 | We demonstrate robustness to the choice of light-curve and Milky Way dust models and to extensions of the model to include peculiar-velocity correlations, a host-mass step, and two stretch populations. | result 0.77 | 0.96 | 0.04 | asserted 1.00 |
| s5 | These results are in agreement with previous distance-ladder measurements and $2.7\sigma$ ($4.1\sigma$) away from the value obtained from Planck data, suggesting that the resolution of the Hubble tension lies outside of SNe Ia---or the distance ladder entirely. | interpretation 1.00 | 0.75 | 0.60 | hedged possibility 0.95 |
What was examined, and what was not
- arXiv listing abstract v1
- full text
- figures and tables
- references and citing sentences
- relations to other papers
- P abstract
- S sections
- T body sentences
- C citations
- R cross-paper relations
Figures
Figures reproduced from arXiv:2610.36051 by the authors.
References
Discussion
Disagree with the review, or know something it missed? Leave a comment, highlight any passage above to comment on it, or challenge this review formally.