{"id":"15f63694-9d0b-4255-9046-0fd0914a418c","arxiv_id":"2507.18946","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"A 95-FRB analysis gives H0 = 71.28 +1.90/-2.08 km/s/Mpc under fixed modeling assumptions, while alternative choices shift the result by up to several sigma.","lead":"The authors used 95 localized fast radio bursts to estimate the Hubble constant at about 71 km/s/Mpc, with a statistical error below 3%. The result depends strongly on modeling choices, and the paper maps how those choices move the answer by several km/s.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Central value H0=71.28 is conditional on point-estimate external parameters; paper's own Tables 4–5 show F, host-DM, fIGM, and halo alternatives shift H0 by 2.5–3.9σ, so the quoted 2.8% precision is not the total uncertainty.","rationale":"The paper is a careful and transparent sensitivity analysis, and it merits credit for enumerating the systematic choices. However, the central claim—that a 95-FRB sample yields H0=71.28 with a ~2.8% statistical uncertainty and thereby improves previous FRB measurements—is only meaningful if the fixed external parameters are correct. The paper's own Tables 4 and 5 demonstrate that the central value moves by 2.5–3.9σ under alternative, published choices for F, host-DM parameters, fIGM, and the Galactic halo. Because these parameters enter the likelihood (Eq. 7) at the same level as H0 and are not marginalized over, the quoted error bars are conditional statistical errors, not the total uncertainty of the measurement. This is precisely the reader's weakest assumption. I agree with the CONDITIONAL verdict: the paper should be accepted as a sensitivity/systematics study, but the headline measurement should be presented with a model-averaged or combined systematic uncertainty. I would not change the verdict; the concern is real but the paper is honest and the analysis is a useful contribution. No internal inconsistency was found in the likelihood construction; the main numerical choices are clearly reported and tested. The most useful single test is a full marginalization over F and host parameters.","tokens_in":20223,"tokens_out":15556,"duration_ms":143489,"concrete_test":"Re-run the baseline 95-FRB MCMC and the 85 one-off subsample with F and (µ_host, σ_host) marginalized rather than fixed: use a Gaussian prior on F centered at the Baptista et al. (2024) best fit with its quoted uncertainty (or a uniform prior over the physically motivated range [0.09,0.32]), and a two-component mixture prior for the host parameters corresponding to Zhang et al. (2020) and Fortunato et al. (2023). Also, in a simpler check, propagate the stated fIGM scaling: compute H0×0.93/0.865−H0; if this exceeds the quoted 1σ error, the statistical-only error bar is incomplete. If the marginalized H0 interval broadens beyond ~±5 km/s or shifts by more than the quoted error, the claim of a 2.8%-precision refined measurement is unsupported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central measurement is produced by the likelihood in Eq. (7), which treats H0 as the only free parameter and fixes all astrophysical inputs to point estimates. The most consequential fixed inputs are the IGM fluctuation amplitude F and the host-galaxy DM distribution (µ_host, σ_host). The paper's own Table 5 (85 one-off FRBs) shows that F=0.32 (the value used in Macquart et al. 2020 and many subsequent studies) yields H0=80.68±2.6, and the Fortunato et al. (2023) host parameters yield H0=61.22±2.0—shifts of roughly 2.9σ and >3σ from the fiducial 70.89±~2.0. Section 4.4 also states that replacing fIGM=0.93 with fIGM=0.865 (Zhang et al. 2025) scales H0 by 0.93/0.865 ≈ 1.075, i.e., to ~76.6, a ~2.8σ shift. Table 4 shows a fixed Galactic halo of 50 pc cm^-3 shifts H0 down to 63.75, >2.7σ. None of these uncertainties is propagated into the quoted error bar. Therefore the headline value 71.28+1.90/−2.08 is a conditional estimate whose central value is largely set by the assumed external parameters, not by the FRB data alone.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper constrains the Hubble constant from the dispersion measure-redshift relation of 95 localized fast radio bursts selected from an initial sample of 117 events. Using a fixed Planck18 background cosmology, NE2001 for the Galactic ISM, the YT2020 model for the Galactic halo, fIGM=0.93 from Connor et al. (2024), F=10^-0.75 from Baptista et al. (2024), and host-galaxy DM parameters from Zhang et al. (2020), the authors report H0 = 71.28^{+1.90}_{-2.08} km/s/Mpc and claim a statistical uncertainty below 2.8%, the lowest among relevant FRB studies. The paper then systematically examines the sensitivity of this result to outliers, Galactic electron density models, the Galactic halo contribution, fIGM, F, and the host-galaxy DM distribution, and shows that several alternative choices shift H0 by more than 2-3 sigma. The paper concludes that these systematic effects need careful treatment before FRBs can provide reliable cosmological constraints.","tokens_in":20613,"tokens_out":4937,"duration_ms":47380,"significance":"If the quoted measurement were robust, it would be a competitive FRB-only H0 determination at about 2.8% statistical precision, and the paper's systematic study is a genuine advance over previous work, which often adopts one set of nuisance parameters without testing alternatives. The explicit tabulation of how H0 shifts under different modeling choices in Tables 3-5, the use of a uniform latitude cut to remove unreliable low-latitude events, and the identification of the degeneracy between H0 and the IGM fluctuation amplitude F are useful contributions. However, the central claim as presented is conditional on point estimates of several external astrophysical parameters, and the paper does not fold the large shifts it documents into a final systematic error budget. This makes the current headline number a conditional estimate rather than a refined measurement with the stated uncertainty.","major_comments":[{"comment":"The headline central value is conditional on point estimates for F and the host-galaxy DM distribution. The paper's own Table 5 shows that replacing F=10^-0.75 with F=0.32 shifts H0 from 70.89 to 80.68, a change of about 2.9 sigma, and that replacing the Zhang et al. (2020) host parameters with those of Fortunato et al. (2023) shifts H0 to 61.22, a change of more than 3 sigma. These shifts are larger than the quoted statistical uncertainty, so the 'statistical uncertainty below 2.8%' reported in the abstract is not the uncertainty of the measurement unless these nuisance parameters are marginalized over or an explicit systematic error is added to the final budget. The paper currently does neither.","section":"Section 4.4, Table 5"},{"comment":"The choice of the Galactic halo model also moves the result by more than the statistical error: fixing DMMW,halo=50 pc cm^-3 gives H0=63.75^{+1.83}_{-1.90}, a downward shift of more than 2.7 sigma relative to the fiducial YT2020-based value. Because the paper itself states that a constant halo model 'has not been conclusively ruled out,' the YT2020 choice cannot be treated as a known input. Some estimate of model uncertainty, such as a prior over halo models or a quoted systematic error, is needed before the central value can stand as a refined constraint.","section":"Section 4.3, Table 4"},{"comment":"The |b| cut is an analysis choice that affects the result: the NE2001-based H0 varies between 65.60^{+1.83}_{-1.80} and 71.28^{+1.90}_{-2.08} across the cuts tabulated, a spread of roughly 2.1 sigma, with the no-cut case giving 65.84^{+1.85}_{-1.70}. The paper states that results converge beyond |b|>15 degrees, but the table shows 69.87 at |b|>20 degrees and 69.95 at |b|>30 degrees, so the residual dependence on the cut should be quantified or included in the error budget rather than asserted to have converged.","section":"Section 4.2, Table 3"},{"comment":"The claim that replacing fIGM=0.93 with fIGM=0.865 scales the result by a factor of 0.93/0.865 assumes a strict proportionality that holds only to first order in the likelihood. Since the paper elsewhere emphasizes that the IGM PDF is skewed and its mean and peak do not coincide, this scaling should be stated as an approximation and, ideally, checked by re-running the MCMC with the alternative fIGM value.","section":"Section 4.4, fIGM scaling"}],"minor_comments":[{"comment":"Equation (5) defines a probability density for DMhost, but Equation (7) writes phost(DMhost,i | mu_host, sigma_host) without specifying that DMhost,i must be marginalized over the log-normal distribution in the convolution with pIGM. The notation should be defined explicitly to avoid ambiguity.","section":"Equation (7)"},{"comment":"Table 6 lists the full 117-FRB sample, but the final 95-FRB subset and the 85 one-off FRB subset used in Table 5 are not identified in the table. Please add a column or marker indicating which bursts enter each sample for reproducibility.","section":"Appendix A, Table 6"},{"comment":"The MCMC settings are not reported: chain length, burn-in, number of walkers, and convergence checks are absent. Table 1 is labeled as a summary of settings, but it lists only astrophysical parameters and results, so the sampling details should be added for reproducibility.","section":"Section 3, Table 1"},{"comment":"The caption states that DMMW,halo and DMhost are 'both assumed with 50 pc cm^-3', which is inconsistent with the fiducial analysis that adopts YT2020 and a log-normal host-galaxy distribution. The caption appears to describe only panel (a) with Macquart et al. (2020) settings, and this should be clarified.","section":"Figure 1 caption"},{"comment":"The acknowledgments contain a duplicated word: 'supported by by the Leading Innovation and Entrepreneurship Team' should read 'supported by the Leading Innovation and Entrepreneurship Team'.","section":"Acknowledgments"}],"recommendation":"major_revision","confidential_remarks":"This is a useful and timely paper, and the systematic study is a genuine contribution. The central issue is that the abstract and introduction frame H0 = 71.28^{+1.90}_{-2.08} as a refined measurement, while the paper's own tables show that the result shifts by 2.7-3.9 sigma under alternative, currently viable modeling choices. This is fixable within the paper's scope by adding a systematic error budget or by explicitly reframing the result as conditional on a specific set of nuisance parameters. I do not see grounds for rejection, provided the authors address the conditional nature of the headline number and the convergence claim for the latitude cut. The citation practice appears appropriate and no scope concerns arise for this journal."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"If you read only the abstract, you'd think the field just got a firm 2.8% FRB-only H0. The headline is 71.28+1.90/−2.08 from 95 localized FRBs, and the paper calls it the lowest statistical uncertainty to date. That part is fine as a statistical statement, but it is not the total uncertainty. The paper's own Tables 4 and 5 show that swapping F to 0.32 moves H0 to 80.7, using Fortunato et al.'s host-DM parameters moves it to 61.2, fixing the Galactic halo at 50 pc cm−3 drops it to 63.8, and replacing fIGM=0.93 with 0.865 scales it to roughly 76.6. Each of those shifts is 2.7–3.9σ against the quoted error bars. The stress-test note is right: the central value is conditional on point estimates of external parameters, and the quoted precision is not the whole story.\n\nWhat the paper does well is the systematic assessment itself. The sample of 117 localized FRBs with 95 retained is new, and the authors test latitude cuts, NE2001 versus YMW16, the YT2020 halo model, the F parameter, host-DM distributions, and fIGM scaling in a transparent way. The tables are clear, the likelihood is standard Macquart-relation MCMC, and the dataset is in the appendix. They are also unusually honest: they do not hide the large shifts, and they explicitly discuss the degeneracy between H0 and F. That is real and citable work.\n\nThe main soft spot is the framing. The abstract and summary present 71.28 as the constrained value, but the analysis never propagates the external-parameter uncertainty into a combined systematic error nor quotes a model-averaged result. The |b|>15° cut is defensible, yet the no-cut NE2001 result is 65.8, a ~2σ shift, so the post hoc nature of the selection matters. None of these are fatal flaws; they are reasons the headline should carry a caveat.\n\nWho is this for? Anyone using or reviewing FRB cosmology, or studying systematics in the Hubble tension. It is a sensitivity map more than a final measurement, and as such it deserves a serious referee rather than a desk reject. I would ask the authors to add an explicit systematic error budget or model-averaged value before publication, but I would not block the paper on that. I'd probably cite it in a methods footnote, and I might bring it to a reading group focused on Hubble-tension systematics.","headline":"A genuinely useful systematic sensitivity map for FRB-based H0, but the headline 2.8% measurement is statistical only and shifts by 2–4σ under the paper's own alternative model choices.","tokens_in":21109,"tokens_out":2531,"would_cite":true,"duration_ms":28287,"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":"Ninety-five localized fast radio bursts pin the Hubble constant at 71.28 km/s/Mpc with 2.8 percent statistical uncertainty.","keywords":["fast radio bursts","Hubble constant","Hubble tension","dispersion measure","Macquart relation","cosmological parameters","baryonic feedback","Galactic halo"],"falsifier":"Measure the amplitude of intergalactic density fluctuations directly—for example, by comparing dispersion-measure scatter among many FRBs at the same redshift using an independent redshift indicator that does not assume $H_0$. If the true $F$ is near 0.32 rather than $10^{-0.75}$, the paper's Table 5 implies the central value would move from about 71 to about 81 km s$^{-1}$ Mpc$^{-1}$.","tokens_in":20026,"feed_emoji":"📡","tokens_out":9651,"duration_ms":91252,"temperature":0.7,"pith_summary":"Within the standard ΛCDM model, the paper seeks to establish that a carefully curated sample of 95 localized fast radio bursts, interpreted through the dispersion measure-redshift (Macquart) relation, gives $H_0 = 71.28^{+1.90}_{-2.08}$ km s$^{-1}$ Mpc$^{-1}$, a statistical uncertainty below 2.8 percent that the authors report as the smallest achieved so far with FRBs. The measurement matters because FRBs are an independent probe of cosmic expansion, and the value sits between the CMB estimate (67.66) and the local distance-ladder estimate (73.04) that define the Hubble tension. The paper's second, diagnostic aim is to explain why earlier FRB analyses disagreed: it traces the discrepancies to concrete modeling choices—the Milky Way electron density model, the Galactic halo contribution, the intergalactic baryon fraction, the amplitude of intergalactic density fluctuations, and the host-galaxy DM distribution—and quantifies how far each choice moves $H_0$.","feed_headline":"95 fast radio bursts pin Hubble constant at 71.28","feed_subtitle":"Fast radio bursts now rival other cosmic probes; systematics still shift the value by several sigma.","key_machinery":"The central object is the Macquart relation, the statistical correlation between an FRB's observed dispersion measure and its redshift. After subtracting the Milky Way disk and halo contributions, the remaining extragalactic dispersion measure is the sum of a host-galaxy term, modeled as log-normal, and an intergalactic term whose mean scales roughly as $\\Omega_b H_0 f_{\\mathrm{IGM}} z$ at low redshift and whose scatter is $\\sigma_{\\mathrm{IGM}} = F z^{-0.5}$. The likelihood multiplies the skewed probability density for the intergalactic fluctuation by the log-normal probability density for the host-galaxy term, and MCMC sampling turns this product into a posterior for $H_0$ alone, with all other cosmological parameters fixed. This machinery converts a growing catalog of localized bursts into a single cosmological number, but it does so by depending on external inputs ($F$, $f_{\\mathrm{IGM}}$, and the host DM parameters) that directly set the scale of the inferred $H_0$.","core_discovery":"Within the standard ΛCDM framework, and using the Macquart dispersion measure-redshift relation, the paper reports $H_0 = 71.28^{+1.90}_{-2.08}$ km s$^{-1}$ Mpc$^{-1}$ from 95 of 117 localized FRBs. The retained sample excludes bursts with ambiguous host associations, bursts hosted by elliptical galaxies, and bursts at Galactic latitude $|b| \\le 15^\\circ$. The analysis fixes the background cosmology to Planck 2018 values, adopts the NE2001 Galactic electron density model, the direction-dependent YT2020 Galactic halo model, an intergalactic baryon fraction of 0.93, an IGM fluctuation amplitude $F = 10^{-0.75}$, and host-galaxy DM parameters fitted to simulations by burst type. The paper claims this is the lowest statistical uncertainty yet for an FRB-only measurement, about 2.8 percent, and demonstrates with controlled comparisons that the choice of $F$ and of the host DM distribution shifts $H_0$ by several $\\sigma$, which it offers as the main explanation for the spread among earlier FRB studies.","pith_inferences":["If $F$ and the host DM distribution are as uncertain as the paper's own comparisons suggest, the true uncertainty on this $H_0$ is larger than the quoted statistical error; a joint analysis treating $F$, $f_{\\mathrm{IGM}}$, and host parameters as free, with priors from independent probes, would likely return error bars closer to 8–10 km s$^{-1}$ Mpc$^{-1}$.","The strong coupling between $F$ and $H_0$ implies FRB-only Hubble constraints cannot be treated as feedback-free cosmological rulers; conversely, fixing $H_0$ externally could turn the same Macquart-relation likelihood into a competitive measurement of the baryon fluctuation amplitude and the diffuse baryon fraction.","A testable prediction of the YT2020 halo model is that future FRB sightlines toward the Galactic center should show systematically higher Milky Way halo dispersion measures; if pulsar-based halo measurements instead find no longitude dependence, the YT2020-based $H_0$ would need revision."],"forward_implications":["An FRB-only route to $H_0$ at roughly 2.8 percent statistical precision is now competitive with other single-probe estimates, so each new localized burst with a host redshift tightens the measurement rather than merely adding a data point.","The measured central value sits between the Planck and SH0ES values, with quoted errors reaching both regimes, so this data set does not break the Hubble tension; it adds an independent point that is compatible with either side at the quoted precision.","The analysis identifies the IGM fluctuation amplitude $F$, tied to baryonic feedback, and the host-galaxy DM distribution as the dominant systematic levers: reverting to $F = 0.32$ raises the inferred $H_0$ to about 80.7, while adopting a fixed log-normal host DM distribution with median 100 pc cm$^{-3}$ lowers it to about 61.2 on the one-off subset.","Adopting a Galactic latitude cut at $|b| > 15^\\circ$ is recommended for FRB cosmology: it removes events with unphysical negative extragalactic dispersion measures and stabilizes the $H_0$ estimate against Galactic electron-density model errors."],"supporting_citations":[{"why":"Establishes the dispersion measure-redshift framework, the skewed IGM fluctuation distribution, and the five-FRB baseline that this analysis extends.","marker":"Macquart et al. (2020)"},{"why":"Supplies the observationally constrained IGM fluctuation amplitude $F = 10^{-0.75}$ adopted in the likelihood.","marker":"Baptista et al. (2024)"},{"why":"Provides the simulation-fitted host-galaxy DM log-normal parameters used for each FRB type.","marker":"Zhang et al. (2020)"},{"why":"Provides the adopted intergalactic baryon fraction $f_{\\mathrm{IGM}} = 0.93$ from FRB observations.","marker":"Connor et al. (2024)"},{"why":"Provides the direction-dependent YT2020 Galactic halo DM model used to subtract the Milky Way halo contribution.","marker":"Yamasaki & Totani (2020)"},{"why":"Provides the NE2001 Galactic electron density model used for the default Milky Way disk subtraction.","marker":"Cordes & Lazio (2002)"},{"why":"Fixes the background cosmological parameters and supplies the CMB comparison value that frames the Hubble tension.","marker":"Planck Collaboration et al. (2020)"},{"why":"Simulation evidence that earlier IllustrisTNG-based variance tables may underestimate baryonic feedback, motivating the choice of $F$ over those tables.","marker":"Zhang et al. (2025)"}],"fun_headline_variants":["FRBs measure Hubble constant to 2.8% precision","95 fast radio bursts yield Hubble value of 71.28","FRB cosmology: 71.28 Hubble constant with systematics check","Fast radio bursts tighten Hubble constant measurement"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The quoted $H_0$ rests on the external inputs being right: the host-galaxy DM distribution and the IGM fluctuation amplitude $F = 10^{-0.75}$; the paper's own table shows that replacing either with plausible alternatives shifts $H_0$ by roughly 3 $\\sigma$ or more.","fun_headline_variants_meta":{"raw":{"variants":["FRBs measure Hubble constant to 2.8% precision","95 fast radio bursts yield Hubble value of 71.28","FRB cosmology: 71.28 Hubble constant with systematics check","Fast radio bursts tighten Hubble constant measurement"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000658,"raw_usage":{"total_tokens":3038,"prompt_tokens":1000,"completion_tokens":2038,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":616,"completion_tokens_details":{"reasoning_tokens":1970}},"tokens_in":616,"tokens_out":2038,"duration_ms":13794,"temperature":1.0,"reasoning_tokens":1970,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T18:04:37.087757+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the amplitude of intergalactic density fluctuations directly—for example, by comparing dispersion-measure scatter among many FRBs at the same redshift using an independent redshift indicator that does not assume $H_0$. If the true $F$ is near 0.32 rather than $10^{-0.75}$, the paper's Table 5 implies the central value would move from about 71 to about 81 km s$^{-1}$ Mpc$^{-1}$.","supporting_citations":[],"review_version":1}