{"id":"cf104ced-da3c-4062-a644-44fbbe8d584a","arxiv_id":"2505.12785","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"SHiP with a far detector at 120 m is projected to reach |U_alpha4|^2 ~ 10^-2 near Delta m^2_41 ~ 10^3 eV^2, exceeding current muon and tau mixing limits.","lead":"This paper estimates how well the proposed SHiP experiment could detect a fourth, 'sterile' neutrino that mixes with the three known neutrinos. It finds that adding a second, farther detector would make SHiP two to ten times more sensitive to sterile mixing than its single-detector design.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"FSND gains rely on fully correlated NSND–FSND systematics; realistic detector differences could erase the advertised factor-2–10 improvement and sigma_norm independence.","rationale":"The paper is a sensitivity projection for a concrete design question: whether adding a far detector at 120 m to SHiP's near detector would meaningfully extend sterile-neutrino reach. The mathematics is standard and the inputs are transparent (event numbers from the SHiP report, efficiencies from OPERA and a thesis, tau background ratio chosen by hand). The novel and central claim is that the dual-baseline combination becomes independent of the normalization uncertainty and reaches |U|^2 ~ 10^-2 for electron and muon flavors. That claim is a direct consequence of the Section 3.2 modeling choice that NSND and FSND share identical nuisance parameters. This is not a small technical detail: it is the mechanism that cancels the leading systematic. The paper's own caveat admits that real detector properties will reduce the correlation, but no quantitative estimate is given. We therefore regard the headline improvement as an idealized upper bound; the paper is appropriately cautious in its conclusion, but the projection's practical value for the SHiP design depends on how much correlation survives. Other possible concerns—lack of a coverage study for the profiled Feldman–Cousins method, energy-independent efficiencies, and the ad hoc R_s/b = 2 for the tau background—are real but secondary: they affect the absolute calibration of the contours, not the qualitative dual-baseline mechanism. The proposed test, introducing a partially uncorrelated normalization uncertainty and re-evaluating the contours, would directly establish whether the advertised factor-2–10 gain survives realistic detector differences. On the substance, the reader's weakest-assumption identification matches ours; the verdict should remain CONDITIONAL, as the paper states its assumption and caveat but does not quantify the degradation.","tokens_in":15313,"tokens_out":12698,"duration_ms":132050,"concrete_test":"Repeat the sensitivity analysis of Section 4 with the FSND signal and background multiplied by an additional factor (1 + delta_rel), where delta_rel is a Gaussian nuisance parameter with standard deviation sigma_rel (0%, 2%, 5%, 10%) independent of the NSND nuisance parameters, and recompute the 90% CL contours in Fig. 3 for R_F/N = 10% and 100%. If the factor-2–10 improvement and the sigma_norm independence vanish for sigma_rel of a few percent, the dual-baseline advantage is not robust to realistic detector differences; the paper should then quote sensitivities as a function of sigma_rel or of the NSND–FSND correlation coefficient.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central projection—that adding FSND improves sensitivity by factors of 2–10 and removes the sigma_norm dependence—rests on the assumption in Section 3.2 that the FSND signal and background are governed by the same nuisance parameters as NSND, scaled only by a fixed R_F/N. This is equivalent to assuming perfectly correlated systematic uncertainties between the two detectors. Under this assumption, common-mode flux and shape uncertainties cancel in the NSND/FSND ratio, which is exactly what produces the sigma_norm-independent contours and the |U|^2 ~ 10^-2 reach. In practice, FSND is a distinct detector at a different location: its efficiency, energy response, and background contamination will differ from NSND, and the relative normalization R_F/N itself carries an uncertainty. Any uncorrelated component of the systematics weakens the cancellation; the paper acknowledges this in Section 3.2 ('uncertainties due to the properties of detectors reduce the correlation between NSND and FSND...') but does not assess the quantitative impact on the headline reach. If the uncorrelated normalization uncertainty is even a few percent, the projected improvement may degrade substantially and the sigma_norm independence may be lost. Thus the advertised sensitivity is an idealized upper bound rather than a robust projection for the proposed dual-baseline configuration.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper estimates the sensitivity of the SHiP experiment to the 3+1 sterile neutrino model using charged-current deep inelastic scattering event spectra. The authors adopt a Feldman-Cousins construction with a parametric bootstrap to handle nuisance parameters, and they compare two configurations: the proposed Near SND (NSND) at 27 m and a hypothetical Far SND (FSND) at 120 m. The main results are that the NSND-only configuration can probe |U_alpha4|^2 ~ 0.1 near Delta m^2_41 ~ 10^3 eV^2, and that adding FSND improves the sensitivity by factors of 2-10 depending on flavor and systematic uncertainty, while making the sensitivity independent of the normalization uncertainty in non-averaged oscillation regions. The paper also discusses two-flavor mixing scenarios, where a cancellation between appearance and disappearance creates kinks in the NSND-only sensitivity curves, which disappear in the dual-baseline approach.","tokens_in":15413,"tokens_out":3858,"duration_ms":42314,"significance":"If the central projection holds, the paper would be a valuable extension of the SHiP physics case: it provides a concrete statistical framework, uses a standard profile-likelihood Feldman-Cousins method with nuisance parameters, and identifies a specific dual-baseline configuration that could substantially improve sterile-neutrino sensitivity near Delta m^2_41 ~ 10^3 eV^2. The comparison with existing constraints from MINOS, NOvA, IceCube-DeepCore, tritium beta decay, and other experiments is useful. However, the headline factor-2-to-10 improvements and the claimed independence from the normalization uncertainty rest on an idealized assumption about systematic uncertainties between the two detectors. The paper itself acknowledges this limitation but does not quantify how much of the advertised sensitivity would survive under more realistic, partially correlated systematics. The analysis is otherwise clearly presented, and the statistical construction is standard and reproducible in principle.","major_comments":[{"comment":"The FSND signal and background are computed with the same nuisance parameters phi_beta and phi_beta,i as the NSND, scaled only by R_F/N. This is equivalent to assuming perfectly correlated systematic uncertainties between the two detectors. The claimed results in Section 4.1 that the dual-baseline sensitivity is 'independent of sigma_norm if neutrino oscillation is not averaged' and that FSND improves the sensitivity by factors of 2-10 are direct consequences of this full correlation, because common-mode normalization uncertainties cancel in the ratio of the two detectors. The paper acknowledges in Section 3.2 that 'uncertainties due to the properties of detectors reduce the correlation between NSND and FSND, which may contribute to a reduction in the sensitivity compared to our estimation,' but it does not assess the magnitude of this reduction. I ask the authors to quantify the impact by introducing a correlation coefficient or an uncorrelated normalization uncertainty for FSND, and to show how the contours in Fig. 3 degrade as the correlation is reduced. Without such a study, the advertised reach is an idealized upper bound rather than a robust projection.","section":"Section 3.2, Eq. (3.7) and subsequent text"},{"comment":"The tau-neutrino channel, which is central to the claimed factor-of-7 improvement over existing constraints near Delta m^2_41 ~ 10^3 eV^2, relies on three imported inputs: the detection efficiency epsilon_tau = 10%, the energy-independent signal-to-background ratio R_s/b = 2, and the assumption that the tau background is dominated by muon-neutrino CC events with charm production. The efficiency and R_s/b values are chosen on the basis of OPERA results and a PhD thesis analysis, but no dedicated SHiP simulation is used, and the sensitivity of the results to these choices is not explored. Since the tau channel is where the paper claims the most striking improvement over past experiments, the authors should provide a brief scan over plausible values of epsilon_tau and R_s/b (e.g., 5-20% and 1-4, respectively) to show that the factor-of-7 claim is not an artifact of optimistic assumptions. This would also clarify whether the energy dependence of the efficiency and the background can affect the conclusions.","section":"Section 3.2, Eqs. (3.9)-(3.10)"}],"minor_comments":[{"comment":"The sentence 'the sensitivity is only dependent on sigma_norm' should read 'the sensitivity depends only on sigma_norm.'","section":"Section 4.1"},{"comment":"In the concluding paragraph, 'kinks and disappearance' should be 'kinks and discontinuities,' and 'prober' should be 'probe.'","section":"Section 5"},{"comment":"The text uses both 'gray' and 'grey' inconsistently; please standardize.","section":"Section 3.2 and Figure 2"},{"comment":"The statement that 'Wilks' theorem with two degrees of freedom is satisfied on the non-averaged regions of electron and tau flavor mixing cases, while for muon flavor mixing cases, the number of effective degrees of freedom is approximately 4' is unexplained and does not appear to be used in the analysis; either justify it with a reference or remove it.","section":"Section 4.1"},{"comment":"Eq. (3.7) does not explicitly state whether anti-neutrino contributions are included in N_beta. The SHiP beam contains both neutrinos and antineutrinos, and the imported event numbers from Ref. [63] presumably include both; please clarify this in the text.","section":"Section 3.2"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within the scope of JHEP as a phenomenological sensitivity study. The main issue is the idealized treatment of NSND-FSND systematic correlations; a quantitative study of partial correlations is needed before the central claim can be taken at face value. The paper does not contain machine-checkable code or data, but the statistical method is standard and the presentation is clear. I would be willing to review a revised version."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nShort version: this is a solid, transparent sensitivity study of the 3+1 sterile neutrino model at SHiP, extending the authors' earlier tau-focused work to all three flavors and both single- and two-flavor mixing. The headline numbers—NSND alone probing |Uα4|^2 ~ 0.1 near Δm^2 ~ 10^3 eV^2, and a dual-baseline FSND gaining factors of 2–10—are plausible as projections. The paper should be read as an idealized estimate, not a final reach statement.\n\nWhat's genuinely new: the full flavor coverage, the Feldman-Cousins treatment with parametric bootstrap, and the observation that appearance-disappearance cancellation creates kinks in two-flavor sensitivity curves that a second baseline removes. That last point is a real phenomenological insight and the strongest part of the paper. The method is standard, the formulas are clearly stated, and the input choices (event numbers from the SHiP proposal, OPERA efficiencies, hand-picked tau background ratio) are all stated openly rather than buried.\n\nSoft spots, in proportion. The load-bearing assumption is that FSND and NSND share the same nuisance parameters, i.e., fully correlated systematics. That's what produces the sigma_norm independence and the factor-of-2–10 improvement. The authors note in Sec. 3.2 that real detector differences would reduce the correlation and weaken the sensitivity, but they don't quantify how much. If uncorrelated systematics are even a few percent, the advertised FSND advantage shrinks, possibly a lot. So treat the FSND contours as an upper bound. Second, the statistical construction isn't validated with coverage studies; for a sensitivity projection that's common practice, but it means the 90% CL statements are approximate. Third, no code or data artifacts are shipped, which limits reproducibility. None of these are disqualifying for a projection paper; they just set expectations.\n\nComparison with existing constraints is done carefully, including the tricky part that different experiments constrain different parameter-space slices. The authors avoid overclaiming: they note NSND alone mostly doesn't beat current limits, and electron flavor gains little.\n\nWho is this for? SHiP collaboration members deciding whether a second far SND is worth building, and sterile neutrino phenomenologists who want a ready-made sensitivity map. It deserves a serious referee; the main request should be a sensitivity scan with partially uncorrelated systematics, not a rewrite.\n\nRecommendation: send to peer review, with the correlation caveat as the central point to address.","headline":"A careful SHiP sensitivity projection for 3+1 sterile neutrinos; the dual-baseline gain is real but depends on fully correlated systematics, which the authors acknowledge but do not quantify.","tokens_in":16177,"tokens_out":1995,"would_cite":true,"duration_ms":21673,"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":"SHiP with a second, far detector can probe sterile neutrino mixing down to $|U_{\\alpha 4}|^2 \\sim 10^{-2}$ and beat present muon- and tau-flavor limits by factors of 7-10.","keywords":["3+1 sterile neutrino model","SHiP experiment","charged-current deep inelastic scattering","Feldman-Cousins method","parametric bootstrap","dual-baseline sensitivity","tau neutrino appearance","short-baseline neutrino oscillations"],"falsifier":"A concrete way to settle the central claim is to repeat the sensitivity calculation with separate, only partially correlated nuisance parameters for NSND and FSND; if the $R_{F/N}=100\\%$ contour at $\\Delta m_{41}^2 = 10^3\\ \\mathrm{eV}^2$ no longer reaches $|U_{\\alpha 4}|^2\\sim 10^{-2}$ or becomes dependent on $\\sigma_{\\rm norm}$, the cancellation mechanism underlying the improvement is refuted.","tokens_in":14927,"feed_emoji":"⚛️","tokens_out":12876,"duration_ms":114518,"temperature":0.7,"pith_summary":"This paper quantifies what the SHiP fixed-target neutrino experiment could learn about the 3+1 sterile neutrino model, the minimal extension of the Standard Model that accommodates short-baseline anomalies. Using the charged-current deep inelastic scattering event spectrum, the paper finds that the planned near detector alone can probe mixing parameters $|U_{\\alpha 4}|^2 \\gtrsim 0.1$ near $\\Delta m_{41}^2 \\sim 10^3\\ \\mathrm{eV}^2$. The main claim is that adding a second, identical detector at 120 m, so that the two baselines see different oscillation phases while sharing the same systematic uncertainties, improves sensitivity by a factor of 2 to 10 depending on flavor and systematics, and makes the sensitivity nearly independent of the flux normalization uncertainty in non-averaged regions. If the estimate is right, SHiP with a far detector reaches $|U_{\\alpha 4}|^2 \\sim 10^{-2}$ for electron and muon flavors and exceeds current muon- and tau-flavor limits by about a factor of 10 and 7, respectively, near $\\Delta m_{41}^2 \\sim 10^3\\ \\mathrm{eV}^2$. This matters because the $\\Delta m^2 \\sim 10^3\\ \\mathrm{eV}^2$ region, and especially tau-flavor mixing, is only weakly constrained by existing experiments.","feed_headline":"Two SHiP baselines could beat sterile-neutrino limits tenfold","feed_subtitle":"Pairing near and far detectors cancels flux systematics, reaching |U|^2 ~ 10^-2 near the 10^3 eV^2 splitting.","key_machinery":"The load-bearing mechanism is the dual-baseline comparison between two copies of the same scattering detector: NSND at 27 m and FSND at 120 m, whose signal and background are computed with the same set of flux nuisance parameters $\\phi_\\beta$ and $\\phi_{\\beta,i}$, with the far detector's events scaled by the single ratio $R_{F/N}$. Oscillation probabilities depend on $L/E_\\nu$, while systematic shifts rescale the spectrum as a whole, so comparing the two baselines separates the oscillation pattern from normalization uncertainty; this is what makes the sensitivity $\\sigma_{\\rm norm}$-independent where oscillations are not averaged. The statistical engine is the Feldman-Cousins method with a parametric bootstrap: a profile-likelihood-ratio test statistic is calibrated by repeated sampling of Poisson-distributed event counts and auxiliary-measurement point estimates of the nuisance parameters, yielding 90% CL confidence regions on the $(\\Delta m_{41}^2, |U_{\\alpha 4}|^2)$ and $(|U_{\\alpha 4}|^2, |U_{\\beta 4}|^2)$ planes.","core_discovery":"On its own terms, the paper claims that SHiP can test the 3+1 sterile-neutrino model through charged-current deep inelastic scattering spectra, and that a second detector at 120 m turns an otherwise uncompetitive search into one that exceeds present limits. With only the near detector at 27 m, the expected 90% CL sensitivity reaches $|U_{\\alpha 4}|^2 \\gtrsim 0.1$ near $\\Delta m_{41}^2 \\sim 10^3\\ \\mathrm{eV}^2$, improving by roughly a factor of two when the normalized flux uncertainty drops from 20% to 10%. Adding the far detector with signal scaled to 10% of the near detector's improves sensitivity by about a factor of 10 for electron and muon flavors and by 2-3 for tau flavor near that mass splitting, and scaling to 100% adds another factor of about 2.3. In regions where oscillations are not averaged, the dual-baseline sensitivity becomes independent of $\\sigma_{\\rm norm}$, and in two-flavor mixing scenarios the kinks caused by appearance-disappearance cancellation disappear. If the observed events match the three-flavor expectation, the estimated contours extend to $|U_{\\alpha 4}|^2 \\sim 10^{-2}$ for electron and muon flavors, exceeding the current muon-flavor constraint by about a factor of 10 and the tau-flavor constraint by about a factor of 7 near $\\Delta m_{41}^2 \\sim 10^3\\ \\mathrm{eV}^2$.","pith_inferences":["A testable extension beyond this paper: recompute the dual-baseline sensitivities with partially correlated nuisance parameters between NSND and FSND, for example one common flux pull plus independent detector-efficiency pulls, to map how quickly the improvement degrades as the correlation falls below 100%.","An implication the authors leave implicit is that the same near/far cancellation logic could be applied to other short-baseline neutrino programs with two detector positions, where normalization uncertainty rather than statistics dominates the sensitivity.","The disappearance of the kinks suggests that in two-flavor 3+1 scenarios, appearance and disappearance can conspire to mimic the standard three-flavor spectrum at a single baseline; the dual baseline is what exposes that degeneracy.","The absolute reach quoted here assumes energy-independent detection efficiencies and a Gaussian energy resolution of $\\sigma = 0.2\\,E_\\nu$; refining the detector response would shift the numerical factors, though the direction of the dual-baseline benefit would likely survive."],"forward_implications":["With only the near detector, SHiP can probe $|U_{\\alpha 4}|^2 \\gtrsim 0.1$ near $\\Delta m_{41}^2 \\sim 10^3\\ \\mathrm{eV}^2$, and cutting the normalized flux uncertainty from 20% to 10% roughly doubles that reach.","Adding the far detector improves sensitivity by about a factor of 10 for electron and muon flavors and by 2-3 for tau flavor near $\\Delta m_{41}^2 \\sim 10^3\\ \\mathrm{eV}^2$, while raising $R_{F/N}$ from 10% to 100% adds another factor of about 2.3.","In non-averaged oscillation regions, the dual-baseline sensitivity is independent of $\\sigma_{\\rm norm}$, so the contours are unchanged whether the normalization uncertainty is 10% or 20%.","In two-flavor mixing scenarios, the appearance-disappearance cancellation that produces kinks and discontinuities in single-baseline sensitivity curves disappears when both baselines are combined.","At $R_{F/N}=100\\%$, the expected 90% CL sensitivity exceeds the current muon-flavor limit by roughly a factor of 10 and the tau-flavor limit by roughly a factor of 7 near $\\Delta m_{41}^2 \\sim 10^3\\ \\mathrm{eV}^2$."],"supporting_citations":[{"why":"This previous study introduced the dual-baseline NSND/FSND geometry and the flavor flux estimates on which the present sensitivity calculation is built.","marker":"[64]"},{"why":"The SHiP/BDF proposal supplies the expected charged-current event numbers and detector configuration used to compute the signal spectra.","marker":"[63]"},{"why":"OPERA results provide the detection efficiencies and background purities adopted for the electron and muon neutrino channels.","marker":"[32]"},{"why":"The Feldman-Cousins method is the confidence-interval construction procedure that the paper adapts, via a parametric bootstrap, to handle nuisance parameters.","marker":"[66]"},{"why":"The parametric bootstrap defines how repeated sampling of nuisance-parameter point estimates enters the profiled Feldman-Cousins p-value.","marker":"[67]"},{"why":"The analysis supplies the reconstructed-energy resolution $\\sigma = 0.2 E_\\nu$ used to bin the CC DIS event spectra.","marker":"[74]"},{"why":"It provides the SHiP tau-neutrino flux and the neutrino cross-section ratios used to compute appearance and disappearance contributions.","marker":"[75]"},{"why":"The tau-channel reconstruction criterion and efficiency used for the $\\nu_\\tau$ signal come from this analysis.","marker":"[77]"},{"why":"The IceCube-DeepCore limit is the existing tau- and muon-flavor constraint against which the paper's sensitivity is compared.","marker":"[33]"},{"why":"The NOvA dual-baseline search provides the current tau/muon sterile-mixing limit used to quantify the factor-of-7-to-10 improvement.","marker":"[79]"}],"fun_headline_variants":["Dual-baseline SHiP search reaches |U|^2 ~ 10^-2","SHiP's near+far detectors multiply sterile neutrino sensitivity","Near-far detector pair could probe |U|^2 ~ 10^-2","Canceling flux errors, SHiP's dual baseline hits |U|^2 10^-2","Dual-baseline SHiP could outdo muon constraints tenfold"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the near and far detectors share one common set of systematic nuisance parameters, so that all relevant uncertainties, most importantly the neutrino flux normalization, are fully correlated between the two baselines and cancel in the comparison; if the detectors' systematics are actually independent or only partially correlated, the advertised factor-of-2-to-10 improvement and the $\\sigma_{\\rm norm}$ independence weaken.","fun_headline_variants_meta":{"raw":{"variants":["Dual-baseline SHiP search reaches |U|^2 ~ 10^-2","SHiP's near+far detectors multiply sterile neutrino sensitivity","Near-far detector pair could probe |U|^2 ~ 10^-2","Canceling flux errors, SHiP's dual baseline hits |U|^2 10^-2","Dual-baseline SHiP could outdo muon constraints tenfold"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000792,"raw_usage":{"total_tokens":3570,"prompt_tokens":1104,"completion_tokens":2466,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":720,"completion_tokens_details":{"reasoning_tokens":2354}},"tokens_in":720,"tokens_out":2466,"duration_ms":16307,"temperature":1.0,"reasoning_tokens":2354,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T20:26:48.486826+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A concrete way to settle the central claim is to repeat the sensitivity calculation with separate, only partially correlated nuisance parameters for NSND and FSND; if the $R_{F/N}=100\\%$ contour at $\\Delta m_{41}^2 = 10^3\\ \\mathrm{eV}^2$ no longer reaches $|U_{\\alpha 4}|^2\\sim 10^{-2}$ or becomes dependent on $\\sigma_{\\rm norm}$, the cancellation mechanism underlying the improvement is refuted.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The parametric bootstrap defines how repeated sampling of nuisance-parameter point estimates enters the profiled Feldman-Cousins p-value."},{"cited_title":"Buonaura,Study ofν τ Properties with the SHiP Experiment, Ph.D","cited_arxiv_id":null,"evidence_quote":"The analysis supplies the reconstructed-energy resolution $\\sigma = 0.2 E_\\nu$ used to bin the CC DIS event spectra."},{"cited_title":"Prompt neutrinos and intrinsic charm at SHiP","cited_arxiv_id":"1807.02746","evidence_quote":"It provides the SHiP tau-neutrino flux and the neutrino cross-section ratios used to compute appearance and disappearance contributions."},{"cited_title":"Iuliano,Event reconstruction and data analysis techniques for the SHiP experiment, Ph.D","cited_arxiv_id":null,"evidence_quote":"The tau-channel reconstruction criterion and efficiency used for the $\\nu_\\tau$ signal come from this analysis."}],"review_version":1}