{"id":"ebdb67e2-33fc-403f-b097-7fc3619d2315","arxiv_id":"2506.01013","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":3.0,"correctness_risk":"high","formal_verification":"none","parameter_count":14,"one_line_summary":"Short GRB rates from three detectors show a low-redshift excess over star formation histories, but the claimed power-law residual is an artifact of the fitted model.","lead":"This paper compares the redshift, luminosity, and event rates of short gamma-ray bursts detected by Swift, Fermi, and Konus-Wind, and finds that the rates exceed star formation histories at low redshift. The authors interpret the leftover rate after subtracting star formation as a power-law decline from old stellar populations, but this residual is largely built into the two-component fit they use.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The steep power-law residual after SFR subtraction is imposed by the assumed two-component fit in Eq. (16) and an unspecified normalization of the SFR curves, rather than being an independently measured excess.","rationale":"The reader identifies the same weakest assumption: the decomposition into a Gaussian SFR-like component plus an independent power law (Eq. 16), combined with an unspecified normalization for the SFR subtraction. This is precisely the load-bearing point for the central claim in the Abstract and Section 4.5. My independent reading confirms the concern is concrete and not merely a stylistic objection. The paper's other results, such as the multi-detector comparison of redshift/luminosity distributions and the known low-redshift excess, have value, but the headline new finding is not independently established. The abstract overstates the case by calling the residual 'surprising' when it is a direct consequence of the assumed fit. The local-rate comparison across detectors is also weakened by differing minimum luminosities (Table 5), and several redshifts/energies in Table 1 are estimated using empirical relations with unpropagated errors; these issues reinforce the correctness-risk assessment but are secondary to the central decomposition problem. On the positive side, the paper provides a clear parametric derivation and uses standard SFR models, and the qualitative low-redshift excess is consistent with past work; however, the quantitative power-law residual claim needs the Monte Carlo test proposed above to be credible. Given the high correctness risk and the fact that the main new result is an artifact of the analysis, the REJECT verdict is appropriate.","tokens_in":35593,"tokens_out":4470,"duration_ms":47808,"concrete_test":"Run a Monte Carlo simulation under the null hypothesis that the true SGRB rate follows a single delayed SFR model (e.g., power-law delay on the Yuksel 2008 SFR), with the same detector selection, sample size, and luminosity function used by the authors. Apply the exact pipeline of Eq. (11) and Eq. (16) to synthetic samples. If the recovered 'power-law residual' appears with similar significance in the simulated pure-SFR data, the claimed old-population component in the real data is an artifact of the fitting procedure and arbitrary normalization. Additionally, refit the observed rates with a single SFR component with a free normalization and compare using BIC; if the two-component model is not strongly preferred, the residual is not statistically required.","verdict_should_be":"REJECT","load_bearing_attack":"The paper's headline new result (Section 4.5, Abstract) is that after subtracting various delayed/undelayed SFR models from the observed SGRB rates, the remaining rates decline steeply as a power law, implying an old-population/compact-binary component. This conclusion rests on two untested assumptions. First, Eq. (16) forces the observed rate to be the sum of a Gaussian (interpreted as the SFR-like component) and a power law A2*Z^{-D}. Fitting this form guarantees that a power-law component appears in the decomposition; the 'residual' is not discovered but manufactured by the choice of fitting function. Second, the SFR curves used in Figures 5 and 6 are not derived from the same normalization as the observed rates. Figure 2 states the redshift distributions are normalized to unity at z=0, but the comparison in Figures 4-6 requires an arbitrary vertical scaling of each theoretical SFR to the observed rate because the theoretical models give only the redshift shape, not the absolute rate. The paper never specifies how this scaling is chosen. If the SFR component is scaled to match at high redshift, the low-z excess that remains is imposed; if scaled differently, the residual can change shape, disappear, or even rise with redshift. Therefore the claimed power-law-like decline is not a robust, independent measurement; it is an artifact of the assumed functional form and the arbitrary SFR normalization. The qualitative low-redshift excess of SGRB rates over SFR models is well known and may be real, but the specific quantitative claim of a power-law residual from old populations is not supported by the analysis as presented.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper compiles 103 short gamma-ray bursts (SGRBs) with measured redshifts from Swift/BAT, Fermi/GBM, and Konus-Wind, fits their luminosity distributions with a smoothly broken power law, derives redshift distribution functions for delayed and undelayed star-formation-rate (SFR) models, and computes event rates. It reports that Swift SGRBs have lower luminosities and redshifts than Fermi or Konus-Wind SGRBs, that the local rate of Swift SGRBs is about two orders of magnitude larger than the other two samples, and that the observed SGRB rates can be fitted with a Gaussian plus a power-law function. The central claim is that after subtracting the delayed/undelayed SFR components, the remaining SGRB rate declines steeply with redshift in a power-law-like form, which the authors interpret as evidence for an old-population or compact-binary merger component.","tokens_in":36046,"tokens_out":10389,"duration_ms":109688,"significance":"The paper provides a useful compilation of SGRB redshifts and luminosities across three detectors and a systematic comparison of two SFR models with three delay-time distributions, which is a valuable reference exercise. If the claimed power-law residual were established by a genuinely independent test, it would be an interesting indication of an SGRB population not tracking recent star formation. However, the central new claim is currently not independently supported: the residual is the power-law term imposed by the assumed fitting function in Eq. (16), and the normalization of the subtracted SFR curves is not specified. The cross-detector local-rate comparison is also weakened by the different Lmin values used in Table 5. These are load-bearing issues that can potentially be repaired with a revised analysis, but as presented the paper overstates what is demonstrated.","major_comments":[{"comment":"The headline claim is circular. Equation (16) fits the observed SGRB rate as a Gaussian plus a term A2 Z^{-D}; the 'remaining rate' shown as purple squares in Figures 5 and 6 is exactly this fitted power-law term, not an independently measured residual after SFR subtraction. To support the claim of a distinct old-population component, the authors need to demonstrate with a model-comparison statistic (for example, Δχ² or BIC) that the data require a power-law component in addition to the best-fitting delayed or undelayed SFR model, and they must report the uncertainties on A2 and D.","section":"Section 4.5, Eq. (16); Abstract"},{"comment":"The manuscript never states how the dimensionless SFR curves of Figure 2, which are normalized to unity at z=0, are scaled to the absolute rates of Eq. (11) before being compared or subtracted. If the vertical scaling is chosen arbitrarily for each panel, then the shape and even the sign of the residual after subtraction are arbitrary. If the scaling instead uses the fitted local rates ρ0 of Table 5, that procedure must be described explicitly and its uncertainties propagated. Without this information, the 'deduction' of SFR components in Figures 5 and 6 is not a well-defined operation.","section":"Figures 4-6 and Section 4.5"},{"comment":"The comparison leading to the claim that the Swift local rate is about two orders of magnitude larger than the Fermi or Konus-Wind rates is not meaningful as presented because the three rows of Table 5 use different minimum luminosities: 2.57×10^48 erg/s for Swift, 1.97×10^50 erg/s for Fermi, and 2.91×10^50 erg/s for Konus-Wind. Since Eq. (10) integrates the luminosity function from Lmin, the Swift sample's much lower Lmin will produce a larger local rate by construction. The authors should compare rates above a common luminosity threshold, using the cumulative ρ0,>L fits in Table 4, or present differential rates at a fixed luminosity.","section":"Table 5 and Section 4.4"},{"comment":"The fitted parameters A1, A2, B, C, and D of Eq. (16) are not reported anywhere in the text or tables; only reduced chi-square values appear in Figure 4. Consequently, the statement that the residual 'steeply declines with redshift in a power-law-like form' cannot be quantitatively evaluated: the reader does not know the slope D, its uncertainty, or the relative amplitude of the power-law term. These parameters should be tabulated for each detector and each SFR model considered.","section":"Section 4.5, Eq. (16)"}],"minor_comments":[{"comment":"There are numerous typographical errors, including 'matche', 'redshit', 'Lognrmal', 'bianary', 'detetors', 'supporse', 'impirical', and 'Univeristy' in the references; a careful proofreading pass is needed.","section":"Throughout"},{"comment":"The sentence 'It needs to point out that we have only taken into account the luminosity errors and Poisson errors of the local event rate density, so the actual errors will be larger and the actual chi-squares will be smaller than the current ones' is confusing, because adding more sources of uncertainty should not decrease a chi-square statistic; this should be rewritten.","section":"Section 4.3"},{"comment":"The sentence 'the local event rates of Swift/BAT, Fermi/GBM and Konus-wind SGRBs are around two orders of magnitude larger than that of either Fermi or Konus-wind SGRBs' is grammatically ambiguous; the intended meaning is that the Swift rate is two orders of magnitude larger than the Fermi or Konus-Wind rates, and this should be stated clearly.","section":"Section 4.4"},{"comment":"The claim that the Fermi and Konus-Wind redshift and luminosity distributions are 'identical' is based on visual inspection of Figure 1; a two-sample statistical test such as Kolmogorov-Smirnov or Anderson-Darling should be reported to support this statement.","section":"Section 4.1"},{"comment":"The number of SGRBs whose redshifts are estimated from the Ep-luminosity relation is not stated, and the systematic uncertainty from this calibration is not propagated into the redshift distributions or event-rate estimates; the authors should quantify this effect.","section":"Table 1, note b"},{"comment":"Several in-text citations lack corresponding entries in the reference list, including Zhang et al. (2025), Pan et al. (2025), and Rong et al. (2025), and the speculative statement connecting SGRBs and FRBs should be clearly labeled as a speculation rather than a result of this analysis.","section":"Section 5"}],"recommendation":"major_revision","confidential_remarks":"The manuscript reads like an early draft with many typographical errors and missing references. The central logical issue—the power-law residual being an imposed component of the assumed fitting function rather than an independent measurement—is serious, but I believe it is fixable by reanalyzing the data with explicit model comparison and by specifying the SFR normalization. The cross-detector local-rate comparison can also be repaired by using a common Lmin. I therefore recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a solid multi-detector compilation of Swift, Fermi, and Konus-Wind SGRB samples, and the per-detector luminosity function fits and local rate tables will be handy. But the headline result — a steep power-law decline in the rate after 'subtracting' SFR components — is not an independent measurement. It is effectively the power-law term of the authors' own two-component fit in Eq. (16), and the subtraction in Figures 5 and 6 depends on an arbitrary vertical normalization of the SFR curves that is never specified. The qualitative low-redshift excess over SFR is real and already in the literature (Dainotti et al. 2021; Zhang & Wang 2018); the quantitative claim about a separate old-population component is not supported by the analysis as presented.\n\nWhat is good: the multi-detector comparison is genuinely useful. The paper treats each detector's selection function separately, fits smoothly broken power laws to the luminosity functions, and tabulates local rates for five SFR models and three delay distributions. The Fermi and Konus-Wind redshift/luminosity distributions looking alike while Swift differs is a clear, testable statement, even without a formal KS test.\n\nSoft spots: the circularity in the residual claim is the big one. The local rate comparison across detectors is also misleading because the Swift sample goes down to Lmin = 2.6e48 erg/s while Fermi and Konus-Wind have Lmin around 2–3e50 erg/s; the reported 'two orders of magnitude larger' Swift rate is mostly a threshold difference, not a physical excess. The redshift estimates using the Ep–L relation and the peak energy estimates add unpropagated systematics, and the abstract's 'identical' claim deserves a statistical test. These are fixable with reframing, but they undercut the paper's central quantitative conclusions.\n\nNet: this is borderline. The data work is worth having, but the central new claim overreaches. Send it to a referee, but expect the residual claim to be reframed or dropped. I would not cite the power-law residual, but I might cite the per-detector rate tabulations if I needed them.","headline":"Useful multi-detector SGRB compilation, but the headline 'power-law residual' is an artifact of the fitting function and an unspecified SFR normalization, not an independent discovery.","tokens_in":36648,"tokens_out":5137,"would_cite":false,"duration_ms":50203,"reading_group":"maybe","serious_thinker":"no","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["98.70.Rz"],"model":"deepseek-v4-flash","headline":"After subtracting star-formation components from observed short gamma-ray burst rates, this paper finds a leftover rate that declines steeply with redshift as a power law, and reads that residual as bursts from old stellar populations or…","keywords":["short gamma-ray bursts","star formation rate","merger time delay","compact binary mergers","luminosity function","event rate density","redshift distribution","detector selection effects"],"falsifier":"Take a redshift-complete sample of SGRBs at $z<1$ and measure the stellar ages of their host galaxies: if those hosts are predominantly young and star-forming, the old-stellar-population explanation of the residual is falsified. A second check is to repeat the subtraction with the star-formation template normalized by an independent efficiency estimate rather than by the fit, and see whether the power-law residual persists or disappears.","tokens_in":35372,"feed_emoji":"💥","tokens_out":16463,"duration_ms":126883,"temperature":0.7,"pith_summary":"This paper asks whether short gamma-ray bursts (SGRBs) track the cosmic star formation rate, with or without a merger delay, using 103 bursts with measured redshifts detected by Swift/BAT, Fermi/GBM, and Konus-Wind. The three detector samples are not interchangeable: Swift sees dimmer, closer bursts with a local rate about two orders of magnitude higher than the other two instruments, whose redshift and luminosity distributions look statistically identical. A Gaussian-plus-power-law fit in $Z=1+z$ reproduces the observed SGRB rate of every detector. The paper's central result is the next step: after subtracting the star-formation templates from the observed rates, a steep, power-law-like decline with redshift remains, which the authors attribute to bursts from old stellar populations or compact binary mergers rather than recent star formation. If right, the residual explains why SGRB rates exceed every star-formation history at $z<1$ and implies that a second, delayed channel produces short bursts.","feed_headline":"Short GRB rates exceed star formation — and the excess is a power law","feed_subtitle":"Stripping out star-formation templates leaves a steep decline with redshift, pointing to compact binary mergers.","key_machinery":"The load-bearing object is the two-component rate model of Eq. (16),\n$$R_{\\rm SGRB}(Z)=\\frac{A_1}{C\\sqrt{2\\pi}}\\exp\\left[-\\frac{(Z-B)^2}{$2C^{2}$}\\right]+A_2 $Z^{{-D}}$,$$\nwith $Z=1+z$: the Gaussian part is the component that tracks the star-formation templates, and the power-law part is what remains after those templates are subtracted from the observed rate. The rate estimates themselves come from the parametric method of Eq. (11), which converts each detector's redshift-luminosity distribution into an event rate using the instrument's field of view, operation time, and sensitivity. The comparison curves are built by convolving two star-formation histories (Yüksel et al. 2008 and Madau & Dickinson 2014) with three merger delay-time distributions: Gaussian with $\\tau_0=2$ Gyr, lognormal with $\\tau_0=2.9$ Gyr, and power-law with $\\alpha_\\tau=0.81$.","core_discovery":"The paper's central claim is that the SGRB rate is two-component: one part follows the (possibly time-delayed) star formation rate, and a second part declines steeply with redshift in a power-law-like form (the rate falling roughly as $(1+z)^{-D}$). The two-component model of Eq. (16), a Gaussian added to a power law in $Z=1+z$, fits the observed rates of the Swift, Fermi, and Konus-Wind samples, and when the star-formation curves are subtracted from the data the leftover rate steeply declines toward higher redshift, as shown in Figures 5 and 6. The paper interprets that residual as evidence that a substantial fraction of low-redshift SGRBs come from old stellar populations or compact binary mergers rather than from young massive stars. Along the way it establishes that the detector samples differ systematically: Swift/BAT SGRBs have a median luminosity roughly an order of magnitude below that of Fermi/GBM or Konus-Wind SGRBs and a local rate roughly two orders of magnitude higher, while the Fermi and Konus-Wind samples have statistically identical redshift and luminosity distributions. The detector dependence is attributed to the instruments' energy bands and sensitivity limits, and the jet-corrected local rates, between roughly 0.30 and 66.47 Gpc$^{-3}$ yr$^{-1}$, overlap the inferred BH–NS merger rate of about 35 Gpc$^{-3}$ yr$^{-1}$.","pith_inferences":["If the residual is real, the natural test is host-galaxy stellar ages at $z<1$: a two-population model predicts bimodal host ages, with the residual channel living in passive, old galaxies and the SFR-tracking channel in star-forming ones — a measurement the paper does not carry out.","The subtraction is sensitive to how each star-formation template is normalized to the data; fixing that normalization with an independently estimated burst efficiency would settle whether the power-law tail survives or dissolves, and the paper leaves that normalization unspecified.","The same Gaussian-plus-power-law decomposition could be applied to other transients with reported low-redshift excesses (long GRBs, FRBs, black holes, AGN) to test whether one delayed channel underlies the pattern across very different source sizes.","Comparing the fitted power-law index $D$ across the three detectors and against gravitational-wave merger-rate evolution would tie the residual to a merger delay-time distribution — or expose it as detector-dependent."],"forward_implications":["A physical residual would mean the low-redshift excess of SGRBs is a genuine population rather than a selection effect: bursts whose progenitors formed long before the burst, consistent with compact binary mergers.","Rate models for compact-object mergers would need two channels, one tracking star formation and one rising toward the present, rather than a single time-delayed convolution of the SFR.","Because the three detector samples imply local rates differing by roughly two orders of magnitude, single-instrument rate estimates carry strong energy-band biases and should be compared only with like-instrument samples.","The analogy the paper draws with non-repeating fast radio bursts gains force: both transient classes show low-redshift excesses that a delayed, old-population channel could explain uniformly.","The jet-corrected local rate interval of 0.30–66.47 Gpc$^{-3}$ yr$^{-1}$ brackets the inferred BH–NS merger rate of about 35 Gpc$^{-3}$ yr$^{-1}$, linking the residual SGRB population to gravitational-wave sources."],"supporting_citations":[{"why":"Supplies the primary undelayed star-formation history (Eq. 1) on which the delayed rate curves are built.","marker":"Yüksel et al. 2008"},{"why":"Supplies the second undelayed star-formation template used to build delayed variants and to test the SGRB-rate comparison.","marker":"Madau & Dickinson 2014"},{"why":"Supplies the Gaussian merger delay parameters ($\\tau_0=2$ Gyr, $\\sigma=0.3$ Gyr) used in the convolution.","marker":"Virgili et al. 2011"},{"why":"Supplies the lognormal ($\\tau_0=2.9$ Gyr, $\\sigma=0.2$ Gyr) and power-law ($\\alpha_\\tau=0.81$) delay parameters.","marker":"Wanderman & Piran 2015"},{"why":"Provides the three merger delay-time distributions the paper follows in convolving star formation into delayed SGRB rates.","marker":"Zhang et al. 2021"},{"why":"Provides the empirical redshift-distribution functions and the local-rate formula (Eq. 11) used to convert each detector sample into a rate.","marker":"Sun et al. 2015"},{"why":"Provides the updated empirical $f(z)$ fitting scheme and the preference for Gaussian and lognormal delays that the paper tests against all three detectors.","marker":"Zhu et al. 2021"},{"why":"Sets the minimum merger delay ($\\tau_{\\min}=10$ Myr) and supports the power-law delay model used in the comparison.","marker":"Paul 2018"},{"why":"Establishes the earlier finding that SGRB rates exceed the power-law delayed star-formation rate at low redshift, which this paper's two-component analysis extends.","marker":"Dainotti et al. 2021"}],"fun_headline_variants":["Short GRB rates hide a power-law excess from old mergers","Excess short GRBs decline as power law after SFR stripped","SGRB rates: SFR plus a steep power-law residual","Short gamma-ray bursts outpace star formation at low z","Residual SGRB rate falls as power law, hinting at mergers"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The analysis assumes the observed SGRB rate is exactly the sum of a Gaussian component tied to star formation and an independent power-law component (Eq. 16), and that the star-formation curves can be subtracted after a normalization the fit itself supplies; if the true rate is not this sum, the leftover power-law decline is an artifact of the decomposition rather than a new population.","fun_headline_variants_meta":{"raw":{"variants":["Short GRB rates hide a power-law excess from old mergers","Excess short GRBs decline as power law after SFR stripped","SGRB rates: SFR plus a steep power-law residual","Short gamma-ray bursts outpace star formation at low z","Residual SGRB rate falls as power law, hinting at mergers"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000966,"raw_usage":{"total_tokens":4211,"prompt_tokens":1147,"completion_tokens":3064,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":763,"completion_tokens_details":{"reasoning_tokens":2974}},"tokens_in":763,"tokens_out":3064,"duration_ms":21256,"temperature":1.0,"reasoning_tokens":2974,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T11:53:39.356058+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a redshift-complete sample of SGRBs at $z<1$ and measure the stellar ages of their host galaxies: if those hosts are predominantly young and star-forming, the old-stellar-population explanation of the residual is falsified. A second check is to repeat the subtraction with the star-formation template normalized by an independent efficiency estimate rather than by the fit, and see whether the power-law residual persists or disappears.","supporting_citations":[],"review_version":1}