{"id":"6214e4e4-df37-47b8-a22c-809d93b10996","arxiv_id":"2607.19183","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"N3LL' resummation is required to reconcile the 13 TeV ATLAS Z-pT spectrum with global PDF fits, but lowering the pT cut below 30 GeV is not yet supported.","lead":"This paper tests whether ultra-precise LHC measurements of how Z bosons move sideways can be added to global fits of the proton's internal structure. It finds that adding special QCD corrections makes the most precise new data agree, but that using very low-momentum data is still not safe.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Central claim depends on the K-factor approximation of eq. (2.5): resummation is applied as a fixed, PDF-dependent bin-by-bin rescaling, so the observed chi2 improvement and gluon stabilization could be artifacts of this approximation.","rationale":"The reader's weakest assumption is exactly the K-factor approximation, and the paper's own limitation statements (Sect. 3.3.1: 'This approach is an approximation...') confirm this is the least secure link. No independent evidence rules out the possibility that the approximation drives the headline result. The omitted resummation-scale variation (Sect. 2.2.3) is a secondary issue: the paper asserts it is marginal but does not display a test, so this is a missing support flag, but it is less directly load-bearing than the K-factor mechanism. The overstatement that the 13 TeV ATLAS data become 'compatible' (chi2/Ndat ~ 4 after resummation) is a semantic concern but does not undermine the direction of the effect. A concrete test—exact RadISH-PineAPPL grids—would settle whether the K-factor approximation is valid. The reader's CONDITIONAL verdict remains appropriate, so no change is recommended.","tokens_in":37552,"tokens_out":5375,"duration_ms":61866,"concrete_test":"Implement exact resummed predictions by interfacing RadISH to PineAPPL, producing grids that include the resummed calculation (eqs. 2.1-2.4) rather than the K-factor of eq. (2.5). Re-run the 13RES fit with these exact grids and compare the resulting ATLAS 13 TeV chi2 per datum and the gluon PDF with the published 13RES results. If the chi2 remains near 4 and the gluon PDF stays compatible with the baseline, the K-factor approximation is validated; if the chi2 shifts by more than ~0.2 or the gluon central value moves by more than the PDF uncertainty, the approximation is inadequate and the central claim is not established.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that small-pT resummation is 'essential' for compatibility of ATLAS 13 TeV Z pT data with the global fit. This is tested only through the multiplicative K-factor K_RES = (dσRES+FO)/(dσFO) in eq. (2.5), computed once with NNPDF4.0 PDFs. The paper itself states (Sect. 3.3.1) that K-factors do not reweigh individual partonic channels and are PDF-dependent, and that an exact RadISH-PineAPPL implementation is left to future work. The concern is load-bearing because the ATLAS 13 TeV data have sub-percent uncertainties; even few-percent bin-by-bin corrections matter. Resummation corrections involve convolutions with PDFs at low scales (eq. 2.3), precisely where the fit is claimed to modify the gluon. If the PDF dependence or partonic-channel dependence of K_RES is significant, the chi2 drop from 7.83 to 3.94 (Table 3.4) and the stabilization of PDFs could be an artifact of using a fixed K-factor rather than genuine evidence that resummation is essential. The paper's own argument for the approximation relies on the NNLO K-factor experience [121], but resummation K-factors are not validated similarly. This is a specific, testable gap in the mechanism underlying the central claim.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper investigates the impact of small-transverse-momentum resummation for neutral-current Drell-Yan production on collinear PDF determination. Using the NNPDF methodology, the authors include ATLAS and CMS Z-boson pT measurements at 8 and 13 TeV. Fixed-order NNLO predictions are supplemented with N3LL' RadISH resummation implemented as bin-by-bin K-factors (Eq. 2.5). A staged series of PDF fits is performed: first revisiting the 8 TeV data with exact NNLOjet predictions and removing the previous 1% numerical uncertainty; then adding the 13 TeV ATLAS/CMS data; then applying resummation with fixed and progressively lowered pT cuts. The main findings are that resummation markedly improves the description of the 13 TeV ATLAS data (chi2 per point from 7.83 to 3.94), stabilizes the extracted PDFs, and that lowering the pT cut below 30 GeV leads to deteriorating fit quality, partly attributed to the treatment of theory-uncertainty correlations.","tokens_in":37942,"tokens_out":3625,"duration_ms":43784,"significance":"If the central claim holds, the paper provides important guidance for future global PDF fits: small-pT resummation should be included even with a conservative pT>30 GeV cut, and apparent PDF distortions from the 13 TeV ATLAS Z-pT data at fixed order are largely artifacts of missing logarithmic corrections. The paper is also valuable as a methodological demonstration: it uses exact NNLO interpolation grids, full-colour dijet predictions, a theory covariance matrix, and a careful staged comparison of data sets and cuts. The availability of the PDF sets from the authors is a useful resource. However, the central mechanism is tested only through an approximate K-factor that is PDF-dependent and does not reweight partonic channels, and the residual chi2 remains significantly above one even after resummation; these issues need to be addressed before the 'essential' conclusion can be considered fully established.","major_comments":[{"comment":"The central claim that resummation is 'essential' for compatibility of the 13 TeV ATLAS data is tested only through the multiplicative K-factor K_RES = (dσ_RES+FO)/(dσ_FO), evaluated once with the previous NNPDF4.0 PDF set. As the paper itself states in Sect. 3.3.1, this approach does not reweight individual partonic channels and is PDF-dependent. Since K_RES is a few percent in the fitted pT>30 GeV region while the ATLAS 13 TeV data have sub-percent uncertainties, the improvement from 7.83 to 3.94 and the gluon stabilization could in principle be an artifact of using a fixed, PDF-dependent rescaling. The analogy with NNLO K-factors [121] is not a substitute for a validation of the resummation K-factor. I request a quantitative robustness test, e.g., recomputing K_RES with an alternative input PDF set or with PDFs from the fitted ensemble, or a comparison with an exact RadISH-PineAPPL im","section":"Eq. (2.5), Sect. 3.3.1"},{"comment":"The theory covariance matrix is built only from renormalisation- and factorisation-scale variations; resummation-scale variations are explicitly excluded. The paper states that 'its effect is marginal' and refers to Sect. 3.3, but I could not find any numerical evidence supporting this claim. Given that the analysis is precisely about the low-pT region where resummation effects are largest, the omission is load-bearing. The authors should either include the resummation-scale variation in the covariance matrix or provide a dedicated test (e.g., varying Q between m_ll/2, m_ll, and 2m_ll) showing the impact on the ATLAS 13 TeV chi2 and on the PDFs. Without such a test, the quoted improvement cannot be distinguished from a particular choice of resummation scale.","section":"Sect. 2.2.3"},{"comment":"The paper describes the improvement of the ATLAS 13 TeV description from chi2/Ndat=7.83 to 3.94 as making the data 'compatible' with the baseline data set. However, a chi2 per point of 3.94 still represents a very poor description for 19 data points. Appendix A confirms that even heavily weighting the data set only reduces this to 3.57 while severely degrading the global fit, leading the authors to conclude that the measurement may have underestimated uncertainties. This should be reflected more carefully in the abstract and conclusions: resummation substantially improves but does not fully restore consistency. The current wording overstates the strength of the conclusion.","section":"Table 3.4, Sect. 3.3.1"}],"minor_comments":[{"comment":"Please clarify the resummation scale choice: Eq. (2.2) defines Q~m_ll, and the text says 'we choose Q=m_ll/2', but Fig. 2.1 appears to use a single scale. A brief sentence on the sensitivity to this choice would help.","section":"Sect. 2.2.2"},{"comment":"The total number of data points Ndat is reported as 4551 for the baseline and all subsequent fits, but the 13 TeV ATLAS and CMS data are added in Tables 3.3 and 3.4. The text states that 'the total number of data points does not include the data points in the 13 TeV ATLAS and CMS data sets', but this convention should also be stated in the main text where the total chi2 is discussed.","section":"Table 3.2 and Table 3.3"},{"comment":"The footnote about the contradiction with Ref. [175] is useful but somewhat terse. Please specify precisely which 'mismatch in the accounting of experimental uncertainties' was identified, to allow the reader to assess the claim.","section":"Footnote 2"},{"comment":"The 'diagonal' and 'envelope' prescriptions used as diagnostics produce anomalously small chi2 values. It would be helpful to state explicitly whether these prescriptions are used only as diagnostics and are not recommended for production fits, to avoid misinterpretation.","section":"Sect. 3.3.2, Table 3.6"},{"comment":"There are occasional formatting issues in the equations and tables, such as the rendering of 'p_ll_T' and the use of 'N dat' in the text. These should be corrected in the final version.","section":"General"}],"recommendation":"major_revision","confidential_remarks":"The paper is technically solid and well within the scope of the journal. The main concern is that the central phenomenological claim — that resummation is essential for the consistency of the ATLAS 13 TeV Z-pT data — rests on an approximate K-factor whose PDF and partonic-channel dependence is not quantified. The authors themselves acknowledge this limitation and defer the exact implementation to future work. I believe this can be addressed with additional validation and by tempering the language, so major revision rather than rejection is appropriate."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my read. The paper does something genuinely new: it puts N3LL' resummed Z-pT predictions into a global NNPDF-style fit, replaces the old N-jettiness K-factors with exact NNLOjet grids, and tests lowering the pT cut. The staged fits are well done, and the finding that resummation drops the ATLAS 13 TeV chi2/N from 7.8 to 3.9 is real. The PDF stabilization effect — the ATLAS data no longer distort the gluon and strange PDFs once resummation is included — is the most useful result for the community.\n\nNow the soft spots. The central test goes through the bin-by-bin K-factor of eq. (2.5), computed once with NNPDF4.0. That is an approximation, as the paper says, and the partonic-channel mixing is not reweighted. The stress-test note is right that the PDF dependence of the K-factor could in principle soak up part of the effect. I don't think it kills the paper, though: both numerator and denominator of the K-factor use the same PDFs, so the main correction is a shape effect, and the fit returns PDFs close to the baseline, which makes the fixed-K approximation roughly self-consistent. Still, the 'essential' claim is one notch stronger than the evidence. An exact RadISH-PineAPPL grid would settle it, and the authors say they leave that for future work.\n\nThe second soft spot is the residual chi2. After resummation, ATLAS 13 TeV still sits at 3.94, which is a poor fit in absolute terms. The appendix shows weighting the dataset barely moves it. That suggests the problem is partly experimental uncertainties, as they conclude. So the phrase 'essential for ensuring consistency' is doing a lot of work. A more precise statement would be: resummation is necessary but not sufficient; the dataset remains hard to accommodate.\n\nThe third issue, resummation-scale variation excluded from the theory covariance matrix, is minor because they show the effect is marginal, but it would be nice to see the numbers.\n\nOverall, this is a careful, honest study. The K-factor caveat is flagged by the authors themselves. The main new quantitative result — that resummation stabilizes PDFs and that lowering the cut below 30 GeV is not warranted — is directionally solid. It deserves a serious referee; the exact-implementation follow-up is the obvious next step. I'd cite it and bring it to the reading group.","headline":"Resummation helps a lot, but the 'essential' claim leans on an approximate K-factor and a dataset that still sits at chi2/N=3.9.","tokens_in":38452,"tokens_out":2919,"would_cite":true,"duration_ms":33715,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Small-transverse-momentum resummation is essential for making the precise ATLAS 13 TeV Z-boson measurement compatible with global PDF fits, and the apparent PDF shifts at fixed order are largely artifacts of missing logarithmic corrections.","keywords":["parton distribution functions","Z-boson transverse momentum","small-pT resummation","N3LL' accuracy","Drell-Yan production","LHC precision measurements","theory covariance matrix","NNPDF methodology"],"falsifier":"Compute the Z-pT spectrum with an exact resummed interpolation grid (e.g., interfacing RadISH with PineAPPL) for the ATLAS 13 TeV kinematics and compare the resulting fit chi2 and PDFs with the K-factor-based fit; if the exact fit shows chi2 per point above 6 or restores the fixed-order PDF distortions, the paper's central claim fails.","tokens_in":37442,"feed_emoji":"⚛️","tokens_out":6287,"duration_ms":64883,"temperature":0.7,"pith_summary":"This paper establishes that the high-precision ATLAS 13 TeV Z-boson transverse-momentum measurement can only be brought into agreement with the rest of the global data used for PDF determination once all-order small-pT resummation corrections are included. At fixed order the data are inconsistent (chi2 per point around 8) and induce shifts in the strange and gluon PDFs, including a distortion that would affect Higgs production predictions. With N3LL' resummation applied as a multiplicative K-factor, the ATLAS chi2 per point drops to roughly 4, the PDFs return to values compatible with the baseline fit, and the gluon PDF is stabilized. The paper also finds that lowering the pT cut below 30 GeV is not yet supported, because the treatment of correlations among theoretical uncertainties becomes the limiting factor.","feed_headline":"Resummation halves 13 TeV Z-data tension in PDF fits","feed_subtitle":"The apparent PDF distortions from the precise ATLAS measurement vanish once logarithmic corrections are included.","key_machinery":"The load-bearing mechanism is a bin-by-bin resummation K-factor, K_RES = (d(sigma_RES+FO)/dpT)/(d(sigma_FO)/dpT), which rescales NNLO fixed-order predictions for the Z-pT spectrum by the ratio of RadISH N3LL'-resummed to fixed-order cross sections. The K-factor is computed with a fixed input PDF set (the previous NNPDF4.0 nominal set) and then applied multiplicatively to the data in the PDF fit; missing higher-order uncertainties are included through a theory covariance matrix built from 7-point renormalisation- and factorisation-scale variations. This K-factor is what converts an inconsistent data set into a compatible one, and the paper explicitly notes that testing its faithfulness requir","core_discovery":"The central claim, stated on the paper's own terms, is that small-pT resummation corrections are not an optional refinement but fundamental for the compatibility of the very precise 13 TeV ATLAS Z-boson pT measurements with the baseline data set, even with a conservative pT >= 30 GeV cut. The apparent PDF distortions induced by these data at fixed order — a 5% depletion of the strange sea at low x and a 2.5% suppression of the gluon at x less than about 10^-3 with an enhancement near x around 0.03 — are largely mitigated once resummation is included. The impact on PDFs is moderate and manifests mainly as a stabilization of the gluon PDF in the x-region relevant for LHC phenomenology.","pith_inferences":["If the K-factor approximation is as accurate as the paper suggests, resummation should be standard practice for any high-precision DY observable entering PDF fits, including phi* and W-pT measurements; the mechanism is generic, not specific to the ATLAS 13 TeV data.","The residual chi2 of about 4 for the ATLAS 13 TeV data suggests that resummation resolves most, but not all, of the tension; the remaining discrepancy is a candidate signal of underestimated experimental correlations or of non-perturbative effects, separable only by an exact resummed-grid implementation.","One could test the correlation model directly: if scale variations are decoupled between low- and high-pT regions, the low-pT deterioration seen in the paper should largely disappear, which would mean the current theory covariance matrix is overcorrelated.","A closure test using resummed pseudodata would establish whether the K-factor method biases PDFs; the paper's approximation could be validated or falsified without waiting for a full RadISH-PineAPPL interface."],"forward_implications":["Future global PDF analyses should include resummed Z-pT predictions while retaining the 30 GeV cut; otherwise the 13 TeV ATLAS data would distort PDFs and degrade fit quality.","The gluon PDF in the region relevant for gluon-fusion Higgs production is stabilized, so precision predictions for the Higgs cross section are unlikely to be significantly altered by the new Z-pT data once resummation is included.","Numerically stable NNLOjet predictions allow removal of the extra 1% uncorrelated uncertainty previously assigned to 8 TeV Z-pT data, without significant PDF distortion.","Extending fits to pT below about 30 GeV is not yet justified; the current correlation model for theoretical uncertainties is too restrictive in the low-pT region.","Determination of alpha_s from global fits could benefit from resummed low-pT data, given the sensitivity of the pT peak to the strong coupling."],"fun_headline_variants":["Resummation halves 13 TeV Z-data tension in PDF fits","Small-pT resummation key for Z-pT data consistency","Resummation curbs Z-induced gluon distortions in PDFs","Z-boson resummation stabilizes gluon in PDF fits","Resummation essential for ATLAS Z-pT PDF compatibility"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The central conclusion depends on the approximation that resummation can be applied as a bin-by-bin multiplicative K-factor computed with a fixed PDF set, without reweighting individual partonic channels; if partonic-channel dependence matters, the improved compatibility of the 13 TeV data could be an artifact.","fun_headline_variants_meta":{"raw":{"variants":["Resummation halves 13 TeV Z-data tension in PDF fits","Small-pT resummation key for Z-pT data consistency","Resummation curbs Z-induced gluon distortions in PDFs","Z-boson resummation stabilizes gluon in PDF fits","Resummation essential for ATLAS Z-pT PDF compatibility"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00026,"raw_usage":{"total_tokens":1495,"prompt_tokens":883,"completion_tokens":612,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":627,"completion_tokens_details":{"reasoning_tokens":520}},"tokens_in":627,"tokens_out":612,"duration_ms":6797,"temperature":1.0,"reasoning_tokens":520,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T13:10:04.075437+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the Z-pT spectrum with an exact resummed interpolation grid (e.g., interfacing RadISH with PineAPPL) for the ATLAS 13 TeV kinematics and compare the resulting fit chi2 and PDFs with the K-factor-based fit; if the exact fit shows chi2 per point above 6 or restores the fixed-order PDF distortions, the paper's central claim fails.","supporting_citations":[],"review_version":1}