{"id":"ed34648b-8474-4b3c-a0bc-cf11d0fea53c","arxiv_id":"1908.07249","paper_version":1,"verdict":"UNVERDICTED","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"IceCube describes the analysis method for an 11-year neutrino search for Galactic core collapse supernovae, with projected sensitivities, but the data remain blinded and no result is reported.","lead":"This IceCube paper lays out the tools for an 11-year neutrino search for galactic supernovae that are hidden by dust or fail to explode, but it does not present a final result yet. The payoff would be a dust-independent neutrino census of stellar collapses in the Milky Way, complementing optical sky surveys.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Projected 99% coverage and the ξ>9.3σ cut rest entirely on an unshown ASTERIA/GEANT-4 equivalence; a few-percent simulation bias would directly change the claimed upper-limit coverage.","rationale":"The reader's weakest assumption identifies the asserted ASTERIA/GEANT-4 equivalence as the load-bearing point, and I agree. The strongest claim is explicitly a projection: it applies only after unblinding and only if no signal is observed, so the paper contains no completed measurement to accept or reject. The data remain blinded, and the upper limit 'will be provided' in the future; therefore UNVERDICTED is the appropriate verdict, and my concern does not move it. The paper does provide real supporting material: detector livetimes, quality cuts, and a data-driven σΔμ that accounts for non-Poissonian background. These support the method, but the numerical cut that defines the projected coverage is calibrated entirely inside an unshown simulation validation. The only cited verification is a reference to an unpublished tool, and no comparison material is included. My concern is not about internal consistency; it is that the projected 99% coverage cannot be checked from the presented evidence. If the collaboration later releases the ASTERIA/GEANT-4 comparison or a public repository showing negligible shifts in the ξ distributions, the concern would be resolved and the projected sensitivity could be reassessed. Until then, the UNVERDICTED verdict stands.","tokens_in":7351,"tokens_out":7663,"duration_ms":78490,"concrete_test":"Recompute Fig. 6 (right panel) using the full GEANT-4 simulation instead of ASTERIA for at least 1,000 injected 8.8 M_sun Hüdepohl supernovae over the same Galactic distance distribution and the same uniformly sampled SNDAQ background, then compare the 99th-percentile ξ cutoff and the retained fraction at ξ>9.3σ with the ASTERIA-based values. If the cutoff shifts by more than ~0.5σ or the retained fraction changes by more than ~2%, the projected 99% coverage is not robust to the asserted ASTERIA/GEANT-4 equivalence.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central forward claim is the projected upper limit valid for 99% of Galactic core-collapse supernovae with fluxes at or above the 8.8 M_sun Hüdepohl model, operationalized as a cut at ξ>9.3σ with systematics (Section 3). This cut is calibrated by 18,000 simulated supernovae whose ASTERIA hit rates are added to recorded SNDAQ DOM rates (Figs. 5 and 6). The only support for ASTERIA is the sentence 'The parameterized simulation [10], verified to produce the same results as the GEANT-4 Monte Carlo'; no comparison plots, residuals, code, or repository are shown. Because ASTERIA is explicitly the faster, less sophisticated scheme and uses GEANT-4 effective volumes, the verification should cover the full chain that generates ξ, including the 1.5 s maximization, the 10 s selection, and the 250 μs deadtime. The injection step is also sensitive: adding simulated hit rates to real DOM rates is only valid if ASTERIA's output is already corrected for deadtime as a function of the total signal+background rate, which the text does not state. A few-percent bias in effective volume or deadtime handling at ~10 MeV would shift the 99% retention threshold and change the projected coverage. This is not an internal inconsistency, but it is an unverified load-bearing calibration step for a claim that is itself prospective.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript, an IceCube Collaboration proceedings paper submitted to ICRC 2019, describes the methodology for an eleven-year search for Galactic core-collapse supernovae using 3911 days of IceCube data from April 2008 to December 2018. The paper introduces the detector response simulation chain (GEANT-4, ASTERIA, SNOwGLoBES), defines the SNDAQ test statistic ξ, and presents a projected sensitivity study using five progenitor models with distances drawn from Galactic radial distributions. The central forward claim is that, if no signal is observed upon unblinding, the analysis will provide an upper limit valid for 99% of all Galactic core-collapse supernovae with neutrino fluxes at or above the conservative 8.8 solar-mass Hüdepohl model, corresponding to a significance cut at ξ > 9.3σ with systematics included (Section 3, Figures 5 and 6).","tokens_in":7542,"tokens_out":4037,"duration_ms":43754,"significance":"If the projected sensitivity is realized, the analysis would deliver the most stringent neutrino-based constraint on the rate of obscured and failed core-collapse supernovae in the Milky Way, improving on the Baksan and LVD limits. The paper has strengths: it uses a large live-time dataset, a clear definition of the test statistic on real background data, multiple progenitor models spanning an order of magnitude in predicted flux, and an explicit, falsifiable projection of 99% signal retention. The use of recorded DOM rates as background and the inclusion of detector systematic uncertainties in the projected cut are positive features. However, the projected 99% coverage and the ξ > 9.3σ cut rest entirely on an asserted equivalence between the fast parameterized simulation ASTERIA and the full GEANT-4 Monte Carlo; this equivalence is not demonstrated in the manuscript, and the paper itself notes order-of-magnitude uncertainties in supernova flux models. The central claim is therefore plausible but not yet fully substantiated.","major_comments":[{"comment":"The projected 99% retention and the ξ > 9.3σ cut depend on the sentence 'The parameterized simulation [10], verified to produce the same results as the GEANT-4 Monte Carlo.' This verification is load-bearing for every sensitivity number in the paper: the ξ distributions for all five progenitor models are generated with ASTERIA, and the 99% coverage threshold is read directly from these distributions. Yet the manuscript provides no comparison plots, residuals, validation statistics, or public code/repository to support the equivalence. ASTERIA is described as a faster and less sophisticated parameterized simulation that uses GEANT-4 effective volumes, so a small bias in effective volume or energy response at ~10 MeV could shift the 99% retention threshold and change the projected upper limit. Please provide a quantitative ASTERIA-vs-GEANT-4 comparison for the full analysis chain that produces ξ, including the 1.5 s maximization, the 10 s selection, and the 250 μs deadtime handling, and state the energy range over which the verification holds. Without this, the central claim is an unsupported calibration step.","section":"Section 3, Figures 5-6"},{"comment":"The paper states that 'The simulated hit rates were added to uniformly sampled DOM rates recorded by SNDAQ between 2008 and 2018.' This injection step is not self-evidently valid with respect to the artificial deadtime. The deadtime fraction in IceCube depends on the total PMT rate, so adding simulated signal hits to recorded background rates changes the total rate and hence the deadtime loss. The text does not state whether the ASTERIA hit rates are corrected for deadtime as a function of the combined signal-plus-background rate, or whether the recorded DOM rates already include the deadtime suppression in a way that makes simple addition correct. If the deadtime coupling is not modeled, the effective signal size is biased and the 99% retention cut derived from Figure 6 will be incorrect. Please clarify the deadtime treatment in the injection procedure and justify the additive model quantitatively, or modify the simulation to account for the rate-dependent deadtime.","section":"Section 3, simulated signal injection"}],"minor_comments":[{"comment":"The text says a 30 M⊙ progenitor model 'yields 1.97×10^53 M⊙'; the unit should presumably be erg, not solar masses.","section":"Section 3, benchmark models"},{"comment":"There is a typo 'artificial deadime' where 'deadtime' is intended.","section":"Section 2"},{"comment":"The black hole model is described as lying entirely beyond the right edge of the plot, but the caption does not explicitly state this; please add a note in the caption so that readers do not infer the model is absent from the analysis.","section":"Figure 6, left panel"},{"comment":"Reference [10] is given only as 'ASTERIA: A Supernova TEst Routine for IceCube Analysis, 2019' with no author list, publication venue, or DOI. Since the manuscript relies on ASTERIA for the central projection, a full citation or a publicly accessible code repository is needed.","section":"References"},{"comment":"The formula for the deadtime suppression factor '0.87/(1+Rdark(t)/NDOM·τ)' is stated without derivation or a citation; please clarify the origin of the 0.87 factor and define all quantities precisely, as this factor directly affects the simulated rates.","section":"Section 1, deadtime formula"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is a proceedings-style methods paper whose central claim is a projected sensitivity. The unshown ASTERIA/GEANT-4 equivalence is a serious gap because the projected limit inherits any simulation bias. In my view, this is fixable with a supplementary verification section or public code release, so major revision rather than rejection is appropriate. The paper would also benefit from an explicit statement that the 99% coverage statement is conditional on the chosen benchmark flux models and does not incorporate the order-of-magnitude flux-model uncertainties acknowledged in Section 2."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a conference writeup, not a measurement. No unblinded data, no candidate, no limit; the promised 11-year upper limit is conditional on seeing nothing. What is actually new is the 99% retention projection for a 3670-day IceCube search, calibrated on 18,000 simulated supernovae from five benchmark models including a failed/black-hole case, plus the Bayesian Blocks trigger and a careful comparison of IceCube's effective volume to Super-K, JUNO, and Hyper-K. The paper is honest about what is inherited: SNDAQ, ASTERIA, and SNOwGLoBES are all earlier IceCube tools. It also flags the big physics systematics—model fluxes vary by an order of magnitude, oscillation assumptions matter—and it does not oversell. The soft spot is exactly where the stress-test note lands. The whole 99% coverage and the xi > 9.3 sigma cut come from ASTERIA output added to recorded DOM rates, and the equivalence to GEANT-4 is asserted in one sentence without a validation plot, residual distribution, or statement about how deadtime is handled when simulated hits are added to the real rate stream. That matters. A few percent bias in effective volume at ~10 MeV, or a wrong deadtime correction as a function of total rate, would shift the retention curve and change the 'valid for 99%' claim. This is not an internal contradiction, but it is an unverified load-bearing step. I would not publish the projected number without seeing the verification; the collaboration presumably has it, since ASTERIA is an internal tool, so this is a fixable presentation gap. The background side is second-order. The paper acknowledges muon bundles broaden the xi distribution and that muon-corrected distributions were not used. That could affect false-alarm rates, but not the stated retention cut directly. Worth a footnote, not a fatal objection. Bottom line: this is a methods-readiness paper for a search that is still blinded. The useful content is the explicit benchmark-model coverage projection and the timing precision from Bayesian Blocks. It deserves a serious referee if submitted to a journal; the referee should ask for the ASTERIA/GEANT validation and a precise statement of the injection deadtime. It is not a result paper and should not be cited as one. I would bring it to a reading group only for people thinking about Galactic supernova search strategies; otherwise it can wait for the unblinded limit.","headline":"A methods-readiness conference paper for a still-blinded 11-year IceCube supernova search; the projected 99% coverage is the one substantive claim, and it hinges entirely on an ASTERIA/GEANT-4 equivalence that is asserted, not shown.","tokens_in":768,"tokens_out":874,"would_cite":false,"duration_ms":38209,"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":"An eleven-year IceCube search will set a neutrino-based limit that covers 99 percent of Galactic core-collapse supernovae at or above a conservative low-mass flux.","keywords":["core-collapse supernovae","supernova neutrinos","IceCube","Galactic supernova rate","failed supernovae","dust-obscured supernovae","neutrino burst detection","SNDAQ"],"falsifier":"Re-run the fast simulation and the full detector simulation on identical 8.8-solar-mass supernovae at 10 kiloparsecs and compare the per-module and total hit rates; if the disagreement is larger than the 14-percent systematic band used in the paper, the $\\xi>9.3\\sigma$ cut would not in fact retain 99 percent of supernovae. The same check can be done on the unblinded data: the measured background $\\xi$ distribution should match the simulated background, with no excess tail above 9 $\\sigma$ once muon-correlated hits are accounted for.","tokens_in":7061,"feed_emoji":"🌌","tokens_out":9586,"duration_ms":93469,"temperature":0.7,"pith_summary":"IceCube was built for high-energy neutrinos, but its unusually quiet optical sensors also let it see the tens-of-MeV neutrino burst of a Galactic core-collapse supernova as a collective rise in the count rate of all its photomultipliers. This paper presents the method and simulation tools for an eleven-year search over 3,911 days of data (3,670 days after quality cuts) aimed at supernovae that are hidden by dust or that fail to explode, since neutrinos escape where light does not. Its central claim is that once the data are unblinded, and if no burst is seen, IceCube will set an upper limit valid for 99 percent of all Galactic core-collapse supernovae whose neutrino flux is at least as high as the conservative 8.8-solar-mass benchmark model. A cut on the search statistic at 9.3 sigma, with detector systematics included, is what achieves that 99-percent retention. If correct, this would be the strongest neutrino-based constraint yet on the rate of obscured and failed stellar collapses in the Milky Way.","feed_headline":"IceCube neutrino search will cover 99% of Milky Way supernovae","feed_subtitle":"Unblinding 11 years of detector data could yield the tightest neutrino limit yet on dust-hidden collapses.","key_machinery":"The argument runs on the moving-average test statistic $\\xi=\\Delta\\mu/\\sigma_{\\Delta\\mu}$, computed in overlapping 1.5-second windows stepped by 500 milliseconds, with the largest value in each 10-second interval kept. Here $\\Delta\\mu$ is the most likely collective deviation of all optical-module hit rates from their running average, and $\\sigma_{\\Delta\\mu}$ is estimated from the data themselves, so non-Poissonian dark-noise fluctuations are folded into the significance. The expected signal shapes come from ASTERIA, a fast parameterized simulation of the detector response that the paper states, without showing verification, reproduces the full simulation; simulated bursts are added to recorded rates and then passed through the same trigger logic that runs online. The retention claim follows from the separation between the simulated $\\xi$ distributions for the five benchmark models and the background-only distribution in the data.","core_discovery":"The central result is a projected sensitivity statement rather than a detection. For the five benchmark core-collapse models considered — a low-mass 8.8-solar-mass electron-capture collapse, an 11.2-solar-mass star, a 27-solar-mass star, a forced explosion of a 30-solar-mass star, and a failed collapse that forms a black hole — the expected values of the test statistic $\\xi$ separate cleanly from the background distribution measured in real IceCube data. Only the lowest-mass model shows any overlap. The black-hole model, which produces a high average neutrino energy and a sharp cutoff after about one second, yields $\\xi$ values that lie entirely beyond the background. Because the simulated supernova positions are drawn from published radial distributions of Galactic structure and the signals are added to recorded detector rates, the projected $\\xi>9.3\\sigma$ cut retains 99 percent of the simulated population under the conservative flux assumption. The search deliberately requires no optical counterpart, so it is sensitive to dust-obscured and failed supernovae that optical surveys would miss.","pith_inferences":["An implicit consequence the paper does not spell out: a null result would bound the rate of failed collapses specifically, because the black-hole model produces the strongest signal of all five benchmarks; separating that rate from the total would require a model-dependent fit.","The quoted 9.3-sigma cut depends on the assumptions of a normal neutrino mass hierarchy and a particular Galactic progenitor distribution; if either changes, the 99-percent coverage cut shifts, so comparisons with other experiments should be made model by model rather than on a single number.","If the equivalence between the fast simulation and the full simulation were publicly verified, the same tool could be extended to estimate IceCube's sensitivity to other short MeV-scale neutrino transients, such as pre-supernova neutrinos, where the collective-rate method would also apply.","A reader could test the pipeline on the Magellanic Clouds by applying the same simulation with muon-subtracted rates, which the paper identifies as future work; this would tell whether the 99-percent coverage extends beyond the Milky Way."],"forward_implications":["If no burst is found after unblinding, the result will set an upper limit on the rate of Galactic core-collapse supernovae that is valid for 99 percent of all collapses with neutrino fluxes at or above the conservative low-mass model, independent of any optical detection.","The same test statistic already runs online in real time, so a genuine Galactic supernova would be flagged immediately and the buffered full waveforms would allow follow-up timing measurements with other neutrino detectors, potentially triangulating the source.","The clean separation of the black-hole model's $\\xi$ distribution from background means that a failed collapse in the Milky Way should be detected with high significance if it happens during the live period; a null result therefore constrains the rate of such collapses.","The search is independent of external information, so it covers dust-obscured supernovae and complements optical and gravitational-wave based rate estimates, including the 1.7-to-2.5 per century expectation from stellar evolution."],"supporting_citations":[{"why":"Supplies the conservative 8.8-solar-mass electron-capture model that defines the flux threshold for the 99-percent coverage claim.","marker":"[13]"},{"why":"Supplies ASTERIA, the fast parameterized simulation used to generate all expected detector rates for the five benchmark models.","marker":"[10]"},{"why":"Defines the moving-average test statistic and the treatment of non-Poissonian dark noise on which the search is built.","marker":"[4]"},{"why":"Supplies the 30-solar-mass forced-explosion and failed-collapse models used as benchmarks.","marker":"[12]"},{"why":"Supplies the black-hole formation model whose high average energy and sharp cutoff give the strongest signal.","marker":"[15]"},{"why":"Supplies a radial distribution of Galactic progenitors used to place simulated supernovae and to estimate systematics.","marker":"[16]"},{"why":"Supplies the alternate Galactic radial distribution used in the systematic band for the retention curve.","marker":"[25]"},{"why":"Supplies the estimated Galactic collapse rate of 1.7 to 2.5 per century, the quantity the search aims to constrain.","marker":"[19]"},{"why":"Provides a previous 90-percent-confidence upper limit on the Galactic core-collapse rate that this analysis aims to supersede.","marker":"[20]"},{"why":"Provides another previous 90-percent-confidence upper limit, setting the comparison point for the projected IceCube sensitivity.","marker":"[21]"}],"fun_headline_variants":["IceCube's 11-year neutrino hunt covers 99% of Milky Way","11-year IceCube data to probe 99% of galactic supernovae","Failed supernovae in sight: IceCube's 11-year neutrino search","No light needed: IceCube neutrino search eyes hidden supernovae"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The 99-percent coverage and the $\\xi>9.3\\sigma$ cut assume the fast parameterized simulation produces the same detector response as the full simulation, and that adding simulated bursts to recorded rates faithfully mimics real backgrounds; the paper asserts both without showing verification.","fun_headline_variants_meta":{"raw":{"variants":["IceCube's 11-year neutrino hunt covers 99% of Milky Way","11-year IceCube data to probe 99% of galactic supernovae","Failed supernovae in sight: IceCube's 11-year neutrino search","No light needed: IceCube neutrino search eyes hidden supernovae"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00021,"raw_usage":{"total_tokens":1438,"prompt_tokens":997,"completion_tokens":441,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":613,"completion_tokens_details":{"reasoning_tokens":360}},"tokens_in":613,"tokens_out":441,"duration_ms":4707,"temperature":1.0,"reasoning_tokens":360,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:22:01.953526+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Re-run the fast simulation and the full detector simulation on identical 8.8-solar-mass supernovae at 10 kiloparsecs and compare the per-module and total hit rates; if the disagreement is larger than the 14-percent systematic band used in the paper, the $\\xi>9.3\\sigma$ cut would not in fact retain 99 percent of supernovae. The same check can be done on the unblinded data: the measured background $\\xi$ distribution should match the simulated background, with no excess tail above 9 $\\sigma$ once muon-correlated hits are accounted for.","supporting_citations":[{"cited_title":"Hüdepohl et al., Phys","cited_arxiv_id":null,"evidence_quote":"Supplies the conservative 8.8-solar-mass electron-capture model that defines the flux threshold for the 99-percent coverage claim."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies ASTERIA, the fast parameterized simulation used to generate all expected detector rates for the five benchmark models."},{"cited_title":"Abbasi et al., Astron","cited_arxiv_id":null,"evidence_quote":"Defines the moving-average test statistic and the treatment of non-Poissonian dark noise on which the search is built."},{"cited_title":"Nakazato et al., Astrophys","cited_arxiv_id":null,"evidence_quote":"Supplies the 30-solar-mass forced-explosion and failed-collapse models used as benchmarks."},{"cited_title":"Sumiyoshi, S","cited_arxiv_id":null,"evidence_quote":"Supplies the black-hole formation model whose high average energy and sharp cutoff give the strongest signal."},{"cited_title":"Ahlers, P","cited_arxiv_id":null,"evidence_quote":"Supplies a radial distribution of Galactic progenitors used to place simulated supernovae and to estimate systematics."},{"cited_title":"Mirizzi, G","cited_arxiv_id":null,"evidence_quote":"Supplies the alternate Galactic radial distribution used in the systematic band for the retention curve."},{"cited_title":"Giunti and C","cited_arxiv_id":null,"evidence_quote":"Supplies the estimated Galactic collapse rate of 1.7 to 2.5 per century, the quantity the search aims to constrain."},{"cited_title":"Petkov et al.,PoS(ICRC2017)960 (2018)","cited_arxiv_id":null,"evidence_quote":"Provides a previous 90-percent-confidence upper limit on the Galactic core-collapse rate that this analysis aims to supersede."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides another previous 90-percent-confidence upper limit, setting the comparison point for the projected IceCube sensitivity."}],"review_version":1}