{"id":"f1d554d6-d619-4326-998f-483758e88a41","arxiv_id":"2412.14262","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"High-velocity Type Ia supernovae occur at comparable rates in low- and high-mass host galaxies in the ZTF DR2 volume-limited sample, contrary to earlier work.","lead":"This paper studies 1,523 Type Ia supernovae from the Zwicky Transient Facility and compares their properties with the masses of their host galaxies. It reports that high-velocity supernovae are not confined to massive galaxies and provides new rate measurements for low-mass hosts.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 'similar HV rates' conclusion is not robust to the SNID redshift–velocity degeneracy in low-mass hosts, where 69% of the peak sample lacks spectroscopic redshifts; the paper acknowledges the degeneracy for SN 2020lmd but does not propagate it through the Table 3 quadrant analysis.","rationale":"Good-faith reading: the paper is honest, includes many sensitivity tests, and the 'exc' rows show the conclusion is not driven solely by the 45 template-matched SNe. However, the central claim is a null result—'similar rates' of HV SNe Ia in low- and high-mass hosts. For such a claim, the dominant threat is systematic bias, not Poisson noise. Table 3 shows the agreement is only ~1σ in every configuration, so even a modest bias could erase it. The most plausible bias pathway is the SNID redshift–velocity degeneracy in faint low-mass hosts, where spectroscopic redshifts are absent for 69% of the peak sample. The authors demonstrate the degeneracy exists (SN 2020lmd) but do not propagate it through the quadrant analysis. The 'exc' rows reduce the sample but do not establish that the remaining spectroscopic subset is representative of the missing 69%; if the missing events are preferentially HV, the observed low-mass HV fraction is underestimated. This is a concrete, testable mechanism rather than a disagreement with prior results. The reader's weakest_assumption pointed to the same general area (template matching and incompleteness), so agreement is partial; we sharpen the mechanism to the velocity–redshift degeneracy and the inadequacy of the 'exc' test. The verdict remains CONDITIONAL: the paper should either bound this systematic (via the proposed re-fit test) or soften the conclusion to 'no significant difference detected at current sensitivity.'","tokens_in":23220,"tokens_out":6149,"duration_ms":60263,"concrete_test":"Re-fit all 45 SNID-based SNe in the [−3,3] d sample with two restricted template sets: one containing only NV templates (v < 12,000 km/s) and one containing only HV templates (v ≥ 12,000 km/s). For each SN, record the 68% redshift posterior under each prior. Flag any SN where the HV-prior redshift shifts by more than Δz = 0.005 (enough to change the host galaxy association at z ~ 0.03) or where the host mass inferred from the HV-prior crosses log(M*/M_sun) = 10. Recompute Table 3 after excluding flagged SNe and after reassigning them to the alternative host mass. If the observed-vs-expected low-mass HV count changes by more than 1σ in either test, the 'similar rates' conclusion is not robust.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 3.4 concludes that HV SNe Ia occur at similar rates in low- and high-mass hosts, based on Table 3, where the observed low-mass HV counts agree with the null expectation within 1σ in all configurations. This null result is only valid if the low-mass sample is not systematically missing HV events or misassigning their host masses. In the peak-light sample, 69% of low-mass SNe Ia lack spectroscopic redshifts and rely on SNID template matching (Section 3.3). For these objects, redshift and Si II velocity are not independent: SNID cross-correlates the SN spectrum, so a template with a different intrinsic velocity can trade off against redshift. SN 2020lmd (Section 2.3, Table 2) is a documented case where an HV feature produces acceptable SNID matches over Δz ≈ 0.007, and the authors adopt the upper-end redshift. If similar degeneracies affect the 45 template-matched SNe used in Table 3, some true low-mass HV events could be assigned to higher-mass hosts or lost entirely, artificially suppressing the observed low-mass HV count and creating the apparent 'similar rates'. The paper's 'exc' rows remove template-matched SNe, but they do not test whether the remaining spectroscopic subset is unbiased, nor do they bound the systematic redshift error. Because the claim is a null result, this unquantified selection systematic is the load-bearing risk.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper analyzes host-galaxy stellar masses for 1523 SNe Ia from the volume-limited ZTF DR2 sample, focusing on low-mass and intermediate-mass hosts. Using Si II λ6355 velocities from peak-phase spectra, SALT2 light-curve parameters, and host masses, it examines the relation between velocity and host mass, subtype-host preferences, the nature of six hostless SNe Ia, and the specific SN Ia rate as a function of host mass. The central claims are that HV SNe Ia occur at similar rates in low-mass and high-mass hosts, in contrast to the earlier PTF-based result of Pan et al. (2020), and that the low-mass specific SN Ia rate is flatter than previous estimates and consistent with little or no metallicity dependence in the delay-time distribution.","tokens_in":23435,"tokens_out":7060,"duration_ms":62395,"significance":"If the HV-rate null result holds, it is a valuable counter to the previous claim that high-mass (metal-rich) hosts preferentially produce HV SNe Ia, and it supports the idea that HV SNe Ia are not a distinct progenitor population. The paper provides the largest spectroscopically confirmed sample of SNe Ia in low-mass hosts to date, including a careful treatment of six hostless events, and it explicitly tests redshift-source selection effects by comparing template-matched and spectroscopic redshifts. The analysis is not circular: the HV/NV classification uses Si II velocity measurements that are independent of host mass, and the rate calculations use external stellar mass functions rather than fitting the target result. The manuscript also makes a clear falsifiable prediction: future larger samples should confirm a flat HV fraction across host mass if the conclusion is correct.","major_comments":[{"comment":"The central null result—that HV SNe Ia occur at similar rates in low- and high-mass hosts—does not bound the dominant systematic, the SNID redshift–velocity degeneracy. The paper itself documents in Section 2.3 (Table 2) that SN 2020lmd has a prominent HV feature that produces reasonable SNID matches over z ≈ 0.011–0.018, and the authors adopt the upper-end redshift. The same degeneracy can affect the 45 template-matched SNe in the Section 3.4 sample and the 69% of low-mass SNe without spectroscopic redshifts in the Section 3.3 peak sample, because redshift and v_Si are not independent in SNID matching: a template with a different intrinsic velocity can trade off against redshift. If some true low-mass HV events are assigned to higher-mass hosts or to NV, the observed counts in the low-mass HV quadrant of Table 3 (34 or 26 for the [−5,5] rows) could be underestimated, and the agreement with the expectation (40±6 or 23±5) could be an artifact of the systematic. The 'exc' rows remove the template-matched SNe but do not establish that the remaining spectroscopic subset is an unbiased tracer of the low-mass HV population. I request a quantitative estimate of SNID redshift errors (e.g., from fits to the spectroscopic subsample or from simulations) and a propagation of those errors through the host-mass and velocity assignments, or an explicit upper bound on the fraction of low-mass HV events that could be misassigned.","section":"Section 3.4, Table 3"},{"comment":"The claim that HV SNe Ia are found in low-mass hosts rests on exactly one event below log(M*/M_sun) = 8 (Fig. 5 and Section 3.3). The abstract and Section 3.4 do not distinguish this bin from the intermediate-mass range 8 < log(M*/M_sun) < 10, where the Table 3 mass split at 10^10 places substantial statistical power. With a single HV event in the truly low-mass bin, the data are consistent with a wide range of HV fractions there, including values far from the 'similar rates' conclusion. Please rephrase the claim to distinguish the mass range where the sample has statistical power (roughly 8 < log(M*/M_sun) < 10 vs > 10) from the log(M*/M_sun) <= 8 bin, and state explicitly that the low-mass bin is unconstrained (0 or 1 HV events). This caveat should appear in the abstract or conclusions.","section":"Section 3.3, Fig. 5"},{"comment":"The specific-rate analysis assumes that the host stellar-mass distribution of the 34 missed 91bg-like events between z = 0.04 and 0.06 matches that of the observed 91bg-like events in those redshift bins. This assumption is load-bearing for the rates at log(M*/M_sun) >= 10, where 91bg-like SNe are most common. If the missed events preferentially occur in fainter hosts, or if the ZTF magnitude limit removes a mass-dependent fraction of 91bg events, the corrected rates and the comparison to ASAS-SN and DES could shift. Please test the sensitivity of the rates to this assumption, for example by using the complete z <= 0.04 91bg sample to define the mass distribution, or by assigning the missed events according to a range of plausible mass distributions and showing how the corrected rates change.","section":"Section 3.5"}],"minor_comments":[{"comment":"The column headers of Table 2 appear garbled (e.g., 'M g,S N 3 Mi,host upper'); the table should use clearly formatted headings that separate the SN absolute magnitude, host upper-limit magnitude, and upper-limit stellar mass.","section":"Table 2"},{"comment":"Several typographical issues appear throughout the text, including 'di fferent' for 'different', 'we also can apply' for 'we can also apply', and 'wich' for 'which' in Section 3.5. A careful proofread is needed.","section":"Section 2.1"},{"comment":"The sentence stating that the [−3,3] day sample yields observed numbers 'higher than or very close to the predicted values' is not supported by the first two rows of Table 3 (observed 20 vs expected 31±6 and observed 16 vs expected 19±4); please rephrase to 'consistent within the uncertainties' or similar.","section":"Section 3.4"},{"comment":"The claim that 91T-like and 99aa-like SNe Ia have a higher percentage in low and intermediate mass bins relative to the total sample would benefit from a quantitative statement or an explicit reference to the percentages shown in Fig. 4, rather than only a qualitative description.","section":"Section 3.2"},{"comment":"The abstract states that HV SNe Ia come from both older and young populations, but the age inference is based on host g−z color rather than direct stellar-age measurements; please clarify this in the abstract or conclusions so the claim is not over-stated.","section":"Abstract and Section 4"}],"recommendation":"major_revision","confidential_remarks":"The paper is part of the ZTF DR2 series and will be read largely by the SN Ia community. The statistical foundation is sound in the sense that the HV-rate difference is not significant, but the wording 'similar rates' is stronger than the data support once the SNID redshift–velocity degeneracy is considered. The authors should be encouraged to add a quantitative systematic analysis rather than only removing template-matched sources. The paper is publishable after that work is done, but I would not accept it in its current form."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Useful paper, but the headline result is less secure than the abstract implies. The real contribution is the sample: the largest volume-limited collection of SNe Ia in low-mass hosts, with six hostless events handled carefully and included as upper limits. The paper also does the right thing by repeating the quadrant analysis with and without SNID template-matched redshifts and by showing the specific rate under two stellar mass functions. That is solid, honest work, and the DR2 context makes the data valuable even where the interpretation is soft.\n\nThe main soft spot is the claim in Section 3.4 that HV SNe Ia show similar rates in high- and low-mass hosts. The numbers do not really support that as a positive statement. For the primary split, you have 34 observed low-mass HV events versus 40 +/- 6 expected, which is about a one-sigma deficit. The other configurations are consistent too, but they are also consistent with a modest preference for high-mass hosts. The paper should give a confidence interval on the HV/NV rate ratio rather than a point null. The abstract and conclusion overstate what a 1-sigma consistency check can establish.\n\nThe second issue is that the Table 3 'low-mass' bin uses a 10^10 Msun split, so it is dominated by intermediate-mass galaxies. The truly low-mass regime (log M* <= 8), which the title and abstract emphasize, contains exactly one HV event in the peak sample. That is not enough to support the claim that HV SNe Ia occur at similar rates in low-mass hosts. The authors should either frame the result as applying to hosts below 10^10 Msun, or explicitly say the low-mass bin is unconstrained.\n\nOn the stress-test concern about SNID redshift-velocity degeneracy: it is a real physical worry, but the paper already tests it by rerunning the analysis without template-matched redshifts. The 'exc' rows show the result does not go away, so the concern is partly addressed. What is missing is a bound on the systematic redshift error in the template-matched subset and a check that the spectroscopic-redshift subset is not biased by host brightness. That would be a useful addition rather than a fatal flaw.\n\nThe rate analysis is fine and appropriately shows that the SMF choice dominates the low-mass behavior. The 91bg correction comes from a companion paper and is not fitted to the target result, so I do not see a circularity problem. The heavy reliance on unpublished companion-paper products is normal for a DR2 paper, but the host catalog and masses should be public by the time this is accepted.\n\nVerdict: send it to peer review. The sample and rate measurements deserve referee time, but the authors should temper the abstract, add a rate-ratio confidence interval, and be clearer about where the evidence actually stops.","headline":"A useful volume-limited sample paper, but the 'similar HV rates' conclusion is a 1-sigma null that gets overstated in the abstract, and the truly low-mass claim rests on a single event.","tokens_in":24181,"tokens_out":3480,"would_cite":true,"duration_ms":37736,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"High-velocity Type Ia supernovae occur at the same rate in low-mass and high-mass host galaxies, according to a volume-limited sample of 1,523 events.","keywords":["Type Ia supernovae","Host galaxies","High-velocity supernovae","Stellar mass","SN Ia rates","Si II velocity","Volume-limited sample","Delay-time distribution"],"falsifier":"Obtain spectroscopic redshifts and peak-phase spectra for the low-mass-host SNe Ia (log(M*/M☉) ≤ 8, plus the six hostless events) and recount the HV fraction; if it is significantly lower than the ~40% seen above 10¹⁰ M☉, the equal-rate conclusion collapses. The current sample contains only one HV event below that mass, so a single additional low-mass HV event or a confirmed absence in a larger sample would directly test the claim.","tokens_in":22967,"feed_emoji":"💥","tokens_out":7542,"duration_ms":62646,"temperature":0.7,"pith_summary":"The paper uses the volume-limited ZTF DR2 sample of 1,523 Type Ia supernovae to ask whether high-velocity (HV) events are really produced preferentially by massive galaxies, as an earlier study claimed. It finds that HV and normal-velocity (NV) SNe Ia occur in similar proportions in low- and high-mass host galaxies, and that the apparent earlier preference likely came from small samples and a bias toward observing bright, massive hosts. The same sample shows that faint, fast-declining subtypes such as 91bg-like events concentrate in massive, old hosts, while bright 91T-like and 99aa-like events prefer lower-mass, younger hosts. It also derives specific SN Ia rates per stellar mass and finds them roughly flat at low masses, consistent with little or no metallicity dependence in the delay-time distribution.","feed_headline":"High-velocity supernovae are as common in small galaxies as large","feed_subtitle":"A volume-limited sample finds fast Type Ia supernovae in low-mass hosts, challenging the earlier mass-metallicity link.","key_machinery":"The comparison engine is the volume-limited ZTF DR2 SN Ia sample (1,523 events at z ≤ 0.06, including 52 low-mass hosts and six 'hostless' events assigned upper-limit masses). Velocities come from the Si II λ6355 absorption feature measured near maximum light, with the standard 12,000 km s⁻¹ split into high-velocity and normal-velocity classes. Host stellar masses are estimated from photometry with PÉGASE.2 spectral energy distribution fitting, and the SN-to-host offset scaled by galaxy size (dDLR) is used to check that host-light contamination does not mimic velocity differences. The rate calculation divides each mass bin's SN count by the integral of a galaxy stellar mass function, normalized to 10¹⁰ solar masses, and applies a completeness correction for faint 91bg-like events.","core_discovery":"The central claim is that high-velocity SNe Ia are not a preferentially high-mass-galaxy phenomenon: in the volume-limited ZTF DR2 sample, the number of HV SNe Ia observed in low-mass hosts is consistent with the number expected if the HV fraction is independent of host stellar mass (34 observed versus 40 ± 6 expected under the equal-rate assumption in the [−5, 5] day sample; 26 versus 23 ± 5 when template-matched redshifts are excluded). The paper concludes that earlier reports of a mass preference were driven by sample size and selection bias rather than by a physical link between host mass or metallicity and high ejecta velocity. It also argues that HV SNe Ia appear in both old and young stellar populations, so they may be a continuous extension of the normal population rather than a separate explosion channel.","pith_inferences":["A natural extension is that the fixed 12,000 km s⁻¹ HV/NV split may not mark two distinct populations; if the HV fraction varies continuously with mass and age, other spectral diagnostics may better separate explosion mechanisms.","With only one HV event below log(M*/M☉) = 8 and many template-matched redshifts, a modest spectroscopic follow-up campaign could either confirm the equal-rate claim or reveal a downturn at the very lowest masses; the current data cannot distinguish these.","If the specific rate is genuinely flat at low masses, using host mass as a proxy for metallicity in cosmological corrections may be less effective than direct metallicities, and the standard 'mass step' may need a different physical origin.","The 91bg-like completeness correction assumes missing faint events follow the host-mass distribution of observed ones; if faint 91bg events are even more concentrated in massive hosts, the high-mass specific rate would rise and affect the metallicity comparison."],"forward_implications":["If HV rates are truly mass-independent, models that tie high ejecta velocity to high host metallicity or to a single progenitor channel need revision.","The apparent HV–mass relation in earlier samples should be reexamined as a selection effect rather than a physical correlation.","The low and roughly flat specific SN Ia rate at low masses, matched by delay-time-distribution models with little or no metallicity dependence, weakens the case for a strong metallicity scaling in the SN Ia rate.","The inclusion of hostless events with upper-limit masses keeps the lowest-mass bin populated, so future deeper imaging can turn these upper limits into detections and sharpen the rate measurement.","The similar dDLR and host-color distributions of HV and NV events imply both classes suffer comparable host-galaxy contamination, so the velocity diversity is intrinsic."],"supporting_citations":[{"why":"Provides the earlier claim that high-velocity SNe Ia preferentially occur in more massive, higher-metallicity hosts; this is the result the paper directly tests and contradicts.","marker":"P20"},{"why":"Established the trend between Si II velocity and host stellar mass that the paper re-examines with a larger, volume-limited sample.","marker":"Pan et al. (2015)"},{"why":"Supplies the Si II λ6355 velocity measurements and spectral subtype classifications used in the analysis.","marker":"Burgaz et al. (2024)"},{"why":"Defines the ZTF DR2 SN Ia sample and the light-curve quality cuts that underlie the volume-limited selection.","marker":"Rigault et al. (2024a)"},{"why":"Simulations demonstrating completeness of the DR2 sample for non-peculiar SNe Ia at z ≤ 0.06, the key assumption behind the rate and HV-fraction comparisons.","marker":"Amenouche et al. (2024)"},{"why":"Provides subtype classifications for non-peak spectra and the completeness limits for faint 91bg-like events used to correct the rates.","marker":"Dimitriadis et al. (2024)"},{"why":"Supplies the stellar-mass-function-based method for specific SN Ia rates and the delay-time-distribution models with metallicity dependence that the paper compares to ZTF data.","marker":"Gandhi et al. (2022)"},{"why":"An earlier untargeted-survey claim that HV SNe Ia are found almost exclusively in high-mass hosts; the paper shows the significance depends on the mass split and that ZTF data match equal rates.","marker":"Dettman et al. (2021)"}],"fun_headline_variants":["Fast supernovae not tied to massive galaxies","HV SNe Ia show no host mass preference in ZTF DR2","Fast supernovae equally likely in small and large galaxies","High-velocity supernovae not exclusive to massive galaxies","No host-mass bias for fast Type Ia supernovae"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The conclusion assumes the volume-limited sample is complete for high-velocity SNe Ia down to the faintest low-mass hosts, even though 69% of low-mass peak-light events lack spectroscopic redshifts (they are template-matched) and only one HV event sits below log(M*/M☉) = 8.","fun_headline_variants_meta":{"raw":{"variants":["Fast supernovae not tied to massive galaxies","HV SNe Ia show no host mass preference in ZTF DR2","Fast supernovae equally likely in small and large galaxies","High-velocity supernovae not exclusive to massive galaxies","No host-mass bias for fast Type Ia supernovae"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000829,"raw_usage":{"total_tokens":3651,"prompt_tokens":1007,"completion_tokens":2644,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":623,"completion_tokens_details":{"reasoning_tokens":2572}},"tokens_in":623,"tokens_out":2644,"duration_ms":16987,"temperature":1.0,"reasoning_tokens":2572,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T12:22:51.902399+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Obtain spectroscopic redshifts and peak-phase spectra for the low-mass-host SNe Ia (log(M*/M☉) ≤ 8, plus the six hostless events) and recount the HV fraction; if it is significantly lower than the ~40% seen above 10¹⁰ M☉, the equal-rate conclusion collapses. The current sample contains only one HV event below that mass, so a single additional low-mass HV event or a confirmed absence in a larger sample would directly test the claim.","supporting_citations":[{"cited_title":"ZTF SN Ia DR2: Simulations and volume limited sample","cited_arxiv_id":"2409.04650","evidence_quote":"Simulations demonstrating completeness of the DR2 sample for non-peculiar SNe Ia at z ≤ 0.06, the key assumption behind the rate and HV-fraction comparisons."}],"review_version":1}