{"id":"0cd0ada8-d37e-4fb8-a177-b651067f75c8","arxiv_id":"2412.21165","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":4,"one_line_summary":"A complete NNLO QCD fit to LHC and HERA dijet data yields alpha_s(m_Z) = 0.1178 with a total uncertainty of 0.0022 and tests the running coupling from 7 GeV up to 7 TeV.","lead":"This paper extracts the strong force coupling constant, alpha_s, from high-energy dijet data at the LHC and HERA colliders using next-to-next-to-leading-order QCD predictions, finding alpha_s(m_Z) = 0.1178, consistent with the world average. It is the first extraction that reaches dijet scales up to 7 TeV and tests the running of alpha_s over more than three orders of magnitude in energy.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No dedicated uncertainty for the 5-flavor decoupling scheme at scales far above m_top is assigned; a non-negligible top-quark contribution to the multi-TeV dijet bins would bias the high-scale running points.","rationale":"The strongest, most defensible reading of the paper is that it is a careful, well-documented determination of alpha_s(m_Z) from NNLO dijet cross sections with full color. The central value 0.1178 +/- 0.0022 is consistent with the world average, the combined fits have chi2/ndof near unity, and the cross-checks with alternative PDFs and data variants are extensive. The running claim, however, is the most novel part and the most dependent on the theoretical scheme. The authors use n_f = 5 throughout, including far above m_top, and explicitly state that no top-quark effects are included. This is a legitimate scheme choice for a 5-flavor calculation, but it is not an exact description of the full SM at 7 TeV: real top-pair production is not a power-suppressed correction at scales far above m_top, and it is not included in the massless 5-flavor NNLO matrix elements. Because no uncertainty is assigned for this omission, the quoted uncertainties on the 4.88 and 7.04 TeV points do not cover this potential bias. A Monte Carlo estimate of the ttbar contribution in the exact fiducial bins can settle whether the concern is numerically relevant. The reader's verdict already conditions on the scheme treatment and on the non-public code, so this stress-test does not change the verdict. The concern is load-bearing because if the ttbar fraction is non-negligible, the multi-TeV running points - the paper's main novelty - would be biased even though the low-scale alpha_s(m_Z) extraction remains intact.","tokens_in":20405,"tokens_out":19321,"duration_ms":228721,"concrete_test":"Generate pp -> ttbar + jets at 13 TeV with an NLO MC (e.g., Powheg-Box + Pythia8 or MadGraph5_aMC@NLO) and apply the exact fiducial selection of the paper (anti-kT R = 0.8, |y*| < 2.0 or y_max < 2.0 as appropriate, y_b < 1.0, and the m_jj binning of Table I). Compute the fraction of selected events in the m_jj bins corresponding to mu_R = 4.88 TeV and 7.04 TeV. If this top-quark fraction is below about 0.5% compared with the quoted 1-2% experimental precision in those bins, the decoupling approximation is adequate and the scheme concern does not land; if it is above a few percent, the high-scale alpha_s points and the RGE-running claim require revision.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central alpha_s(m_Z) result is robust, but the headline running claim (Fig. 2 and Table III) rests on treating the 5-flavor massless scheme, with the top quark in the decoupling limit, as exact at all scales up to mu_R = 7.04 TeV. Appendix A states that alpha_s evolution is performed with n_f = 5 'in particular also beyond the top-quark mass threshold' and that the NNLO calculation 'does not include top-quark effects.' This is consistent up to power corrections m_t^2/mu_R^2, but at mu_R >> m_t real top-pair production contributes to the inclusive dijet final state (via t -> bW -> jets) and is absent from the 5-flavor massless matrix elements. No uncertainty is assigned for this omission. The affected points are precisely those carrying the multi-TeV reach: mu_R = 4.88 TeV (alpha_s(m_Z) = 0.1185 +/- 0.0031 +/- 0.0034) and mu_R = 7.04 TeV (0.1232 +/- 0.0128 +/- 0.0037). A few-percent top-quark contamination in these bins would shift alpha_s(m_Z) by an amount comparable to the quoted scale uncertainties and could change the apparent consistency with the RGE. The low-scale 7.4 GeV point is also explicitly acknowledged as below the 2m_b threshold (Section V), so the 'spanning more than three orders of magnitude' claim includes a point the authors themselves regard as at the edge of validity.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript determines the strong coupling constant and tests its running using inclusive dijet cross sections from ATLAS and CMS at 7, 8 and 13 TeV, together with HERA dijet data, all confronted with complete (full-color) NNLO QCD predictions obtained with NNLOJET and stored in APPLfast grids. A fit to the five LHC data sets (367 points after rapidity cuts) yields alpha_s(m_Z) = 0.1178 +/- 0.0014 (fit,PDF) +/- 0.0001 (mu0) +/- 0.0017 (mu_R,mu_F) with chi2/ndof = 0.92, and the combined LHC+HERA fit (486 points) gives 0.1180 +/- 0.0024 total with chi2/ndof = 0.88. In a second step the data are grouped into 20 renormalization-scale intervals and one alpha_s(m_Z) parameter is fitted per interval, giving the running from mu_R = 7.4 GeV to 7.04 TeV; the resulting values are claimed to agree with the RGE expectation, which the authors present as a test of the QCD running over almost three orders of magnitude. The appendix contains a consistency study with seven PDF sets, split by rapidity and mass bins, and the full correlation matrix of the 20 scale-interval parameters.","tokens_in":20624,"tokens_out":30577,"duration_ms":302965,"significance":"The central alpha_s(m_Z) result is credible and well cross-checked: variants of the fit (CMS 2D vs 3D data, individual vs combined data sets, seven PDF choices) return values between 0.1172 and 0.1181, all consistent with the world average, and the global fit quality (chi2/ndof = 0.88-0.92) is good. The paper appears to deliver the first full-color NNLO alpha_s extraction from LHC dijet data and extends the scale reach well beyond previous LHC jet determinations, with competitive precision (about 1-2% per point) in the 0.25-3.5 TeV interval. The explicit reporting of the correlations of the 20 fitted alpha_s(m_Z) parameters (Table V), the tests with seven PDF sets, and the careful separation of fit/PDF, mu0, and scale uncertainties are strengths. The headline multi-TeV and wide-scale running claims, however, rest on the highest- and lowest-scale points, where the paper itself flags validity issues (five-flavor scheme below 2m_b and top decoupling far above m_t); these claims need additional support before the paper can be accepted as stated.","major_comments":[{"comment":"The multi-TeV running claim rests on the two highest-scale entries of Table III, but the theoretical framework there treats the top quark in the decoupling limit: Appendix A states that alpha_s evolution is performed with n_f = 5 'in particular also beyond the top-quark mass threshold' and that the NNLO calculation 'does not include top-quark effects.' In the inclusive dijet phase space at m_jj of several TeV, top-pair production followed by jetty W decays can populate the selected final state, and because the partonic luminosity and the 2->2 matrix elements are of the same order as for massless dijet production at these x values, the relative contamination is not obviously negligible. No estimate or uncertainty for this effect is given; a few-percent top-quark contribution would shift the alpha_s(m_Z) values in the 4.88 and 7.04 TeV rows by amounts comparable to their quoted (fit,PDF) and scale uncertainties (0.0031 and 0.0034, and 0.0128 and 0.0037, respectively). I request that the authors either (i) quantify the top-quark-initiated contribution to the selected bins, e.g., with a parton-level t tbar + jets calculation in the experimental phase space, and subtract it or assign an uncertainty, or (ii) state the multi-TeV determination with an explicit caveat limiting the running claim to scales where the decoupling approximation is safe.","section":"V, Table III (rows 4880, 7040), Appendix A"},{"comment":"The RGE test is currently asserted visually ('excellent agreement') without a quantitative statistic. The authors' own machinery provides the natural test: comparing the one-parameter fit of Section IV (chi2/ndof = 0.88 over 486 points) with the 20-parameter per-interval fit of Section V, the difference chi2(1 param) - chi2(20 params) on 19 degrees of freedom directly tests the running hypothesis and should be reported. In addition, the claim that 'the assumption of the QCD running enters in each interval only within a very limited range' is hard to reconcile with the fact that each fitted alpha_s(m_Z) is converted to alpha_s(mu_R) by evolving over the full interval from m_Z to mu_R; the actual test is the consistency of the 20 extracted alpha_s(m_Z) values, and this should be stated explicitly. Since the (mu0) and (mu_R,mu_F) uncertainties are declared fully correlated across intervals (Appendix E), the cleanest statement would be a chi2 of the 20 alpha_s(m_Z) values against the world average, evaluated with the full covariance, with an indication of what such a test can exclude given the correlations in Table V.","section":"V, Table III, Table V, Fig. 2"},{"comment":"The nominal selection cuts y* < 2.0 and y_b < 1.0 were adopted after the consistency study in the supplementary material showed elevated chi2/ndof at large rapidity, and the text states the cuts were chosen to 'avoid a possible bias' and to 'reduce some moderate tensions.' Because this is a data-driven selection that removes about 26% of the LHC data points, the quoted central value and uncertainty could carry a selection bias, and no demonstration of stability under the cut choice is provided. I ask the authors to report alpha_s(m_Z) for at least one tighter (e.g., y* < 1.5) and one looser (e.g., y* < 2.5) selection, or to include the high-y bins with the per-bin alpha_s values shown; this is needed to confirm that the result is stable at the level of the quoted total uncertainty of 0.0022.","section":"II and Supplementary material (Fig. 4, Sec. E.1)"}],"minor_comments":[{"comment":"The claim of spanning 'more than three orders of magnitude' is not strictly met by the reported endpoints: 7.04 TeV / 7.4 GeV = 951, which is below 1000, and the 7.4 GeV point is itself flagged in the text as below the 2m_b threshold. Recommend rewording to 'nearly three orders of magnitude' and stating the range without the 7.4 GeV point (from about 10 GeV) when the scheme-validity caveat is applied.","section":"Abstract and Section V"},{"comment":"The sentence 'the assumption of the QCD running enters in each interval only within a very limited range' is misleading for the multi-TeV intervals, since the RGE is used to relate alpha_s(m_Z) to alpha_s(mu_R) across the full scale range; please clarify the wording.","section":"V"},{"comment":"The bin-to-bin correlation of 0.5 assumed for the non-perturbative uncertainties is adopted from Ref. [42] without a sensitivity study; please state the effect of this assumption on the fit uncertainty (e.g., repeats with correlation 0 and 1).","section":"Appendix B"},{"comment":"Typos and referencing: 'with with n_f=5' (Appendix A), 'mutli-TeV' (Abstract), 'previosuly' (Supplementary material). Also, Ref. [41] is a self-reference to the inline supplementary material and should be replaced by a direct citation of the appendix.","section":"Appendix A, Abstract, Supplementary material"},{"comment":"The captions describe the shaded/hatched band as 'the value of alpha_s(m_Z) from LHC dijet data and its running'; since the upper panels show alpha_s(mu_R), please clarify which uncertainty is propagated to the band and how the correlation of the scale uncertainty is treated.","section":"Fig. 2 and Fig. 3 captions"},{"comment":"The 7.04 TeV row has a (fit,PDF) uncertainty of 0.0128 on alpha_s(m_Z) and about 6% relative uncertainty on alpha_s(mu_R); a sentence stating this precision would prevent the title's 'precise determination up to the multi-TeV range' from being overread as applying uniformly to all scales.","section":"Table III"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is a good fit for a high-energy physics journal, and the central alpha_s(m_Z) value will be useful regardless of the running claim. The main risk is the multi-TeV part of the running claim, which depends on the 4.88 and 7.04 TeV points and on the five-flavor decoupling scheme; I recommend that the authors quantify the top-quark contamination or explicitly restrict the claim. The self-citation to Ref. [41] should be cleaned up. There is a concurrent CMS measurement (Ref. [50]) already acknowledged in the Note added; no novelty concern."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is the first full-color NNLO extraction of alpha_s from LHC dijet data, and it delivers a genuinely new kinematic reach: alpha_s points up to 7 TeV, plus a HERA+LHC running test spanning 7 GeV to 7 TeV. The central value alpha_s(mZ)=0.1178(22) is consistent with the world average, the fits have chi2/ndof near 0.9, and the consistency checks are extensive: multiple PDF sets, mu0 variations, 2D/3D CMS variants, per-bin rapidity studies, and a useful correlation table. The machinery is standard but applied carefully, and the alpha_s(mZ) result is robust to the variations shown. Credit is due for the honest treatment of the low-scale 7.4 GeV point, which the authors themselves flag as below the 2m_b threshold and at the edge of the nf=5 scheme's validity.\n\nThe weak spots are real but not load-bearing. The stress-test worry about top-quark decoupling at scales far above m_top is, I think, overstated: top-pair production contributes negligibly to inclusive dijet bins at these multiplicities and mass scales, and the 7 TeV point has a ~13% total uncertainty, so a few-percent scheme effect would not change the running picture. Still, the authors should add a sentence quantifying why the decoupling approximation is safe at mu_R >> m_t, rather than leaving it as an implicit assumption. The rapidity cuts were informed by the consistency study, which is a post-hoc selection, but the cuts have independent motivation (reduced PDF sensitivity, tracking coverage) and the resulting chi2 is excellent, so I do not see a bias beyond what the PDF and scale uncertainties already cover. The upper panel of Fig. 2 does display alpha_s(mu_R) computed with the very RGE under test, so the impressive-looking agreement is partly built in; the lower panel and Table III are the more honest test, and they look fine. The absence of public code is a limitation, but it is standard for NNLOJET-based analyses and not a reason to reject.\n\nThis paper deserves a serious referee. The central extraction is credible, the running test is the widest single-process check to date, and the authors have done the needed internal-consistency homework. I would cite it and would bring it to reading group.","headline":"A careful, genuinely new full-color NNLO alpha_s extraction from LHC dijets with a wide running test; the central value holds up, and the main soft spots are scheme-approximation and selection-choice concerns rather than flaws in the fit.","tokens_in":21417,"tokens_out":2056,"would_cite":true,"duration_ms":24028,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["12.38.Qk","13.87.-a"],"model":"deepseek-v4-flash","headline":"Dijet cross sections from the LHC and HERA yield alpha_s(mZ)=0.1178±0.0022 and, for the first time, probe the strong coupling up to 7 TeV in agreement with QCD running.","keywords":["strong coupling constant","alpha_s running","dijet production","NNLO QCD","renormalization group equation","LHC","HERA","parton distribution functions"],"falsifier":"Recompute the multi-TeV $\\alpha_s(\\mu_R)$ values using a scheme that includes top-quark running effects (for example, a matched six-flavor evolution) or assign a dedicated uncertainty for the decoupling approximation by comparing schemes; if the resulting $\\alpha_s(\\mu_R)$ points shift beyond the quoted uncertainties or acquire a trend with $\\mu_R$, the claimed verification of the renormalization group running is not robust.","tokens_in":20079,"feed_emoji":"⚛️","tokens_out":14179,"duration_ms":125340,"temperature":0.7,"pith_summary":"The paper extracts the strong coupling constant of QCD from dijet cross sections measured at the LHC and at HERA, using next-to-next-to-leading-order (NNLO) predictions that include all subleading color contributions for the first time in a jet-based determination. From the five LHC data sets alone it obtains $\\alpha_s(m_Z) = 0.1178 \\pm 0.0022$, and adding HERA dijet data gives $0.1180$ with similar total uncertainty, both consistent with the world average. By fitting separate $\\alpha_s(m_Z)$ parameters in twenty intervals of the renormalization scale, the analysis probes the running of the coupling from about 7 GeV to 7 TeV — the first determination up to 7 TeV — and finds agreement with the three-loop renormalization group equation across more than three orders of magnitude. The result matters because it tests the asymptotic behavior of QCD at the highest energy scales where the strong coupling has been measured.","feed_headline":"Strong coupling probed up to 7 TeV for the first time","feed_subtitle":"Dijet data from LHC and HERA confirm the renormalization group running of alpha_s from 7 GeV to 7 TeV.","key_machinery":"The load-bearing object is the complete next-to-next-to-leading-order QCD calculation for inclusive dijet production, including all subleading color contributions, together with an interpolation grid that stores the hard-scattering coefficients independently of the value of $\\alpha_s$ and of the parton distribution functions. The renormalization scale is identified with the dijet invariant mass $m_{jj}$ at the LHC and with $\\sqrt{Q^2 + \\langle p_T \\rangle_{1,2}^2}$ at HERA; these same scales group the data into twenty intervals, and the fit assigns one $\\alpha_s(m_Z)$ parameter to each interval while the three-loop running in the modified minimal-subtraction scheme is used to evolve to $\\mu_R$. A single least-squares fit over all intervals exploits correlations among experimental and PDF uncertainties, yielding both the individual $\\alpha_s(m_Z)$ values and their full correlation matrix. The top quark is treated in the decoupling limit with five active flavors throughout.","core_discovery":"The paper's central claim is that complete NNLO QCD predictions for dijet production, evaluated from 367 LHC data points (and 486 when HERA data are included), determine $\\alpha_s(m_Z) = 0.1178 \\pm 0.0022$ from the LHC subset and $0.1180$ from the combined fit, both consistent with the world average of $0.1180 \\pm 0.0010$. It further claims that a fit with one $\\alpha_s(m_Z)$ parameter per renormalization-scale interval yields values of $\\alpha_s(\\mu_R)$ that follow the three-loop renormalization group running from $\\mu_R = 7.4$ GeV to $\\mu_R = 7.04$ TeV, with no significant deviation in any interval. The authors present this as the first determination of $\\alpha_s$ at scales up to 7 TeV and as the first use of full-colour NNLO dijet predictions, including subleading color contributions, in an $\\alpha_s$ extraction from LHC jet data.","pith_inferences":["A natural next step would be to apply the same scale-binned fitting technique to inclusive jet, three-jet, or event-shape measurements at the LHC; those observables have different scale choices and could provide independent confirmation of the 7 TeV point, which currently carries the largest uncertainty.","The published correlation matrix for the twenty $\\alpha_s$ parameters would allow future global analyses, such as PDF determinations, to reuse these results as external constraints without re-fitting the raw dijet cross sections.","Because the 7.4 GeV interval lies below the two-bottom-quark threshold, a four-flavor calculation for that single interval would test whether the five-flavor prediction biases the low-scale end of the reported running.","If future high-luminosity data reveal a deviation above 1 TeV, separating it from large-$x$ gluon PDF effects will be the principal challenge; the rapidity cuts in this analysis already reduce that entanglement, but a dedicated multi-TeV PDF study would sharpen any such claim."],"forward_implications":["The combined value $\\alpha_s(m_Z) = 0.1180$ from LHC and HERA dijets adds an independent, high-scale constraint on the strong coupling with precision comparable to the world average.","The measured running from 7 GeV to 7 TeV extends the direct test of QCD's renormalization group equation beyond any previous single analysis, covering the multi-TeV region where the coupling is smallest.","Including subleading color contributions in the NNLO predictions is shown to be feasible and relevant for $\\alpha_s$ determinations from jet data, moving beyond the leading-color approximation used in earlier LHC jet fits.","The consistency between HERA and LHC dijet data in a single correlated fit ($\\chi^2/\\mathrm{ndof} = 0.88$) strengthens confidence in the NNLO predictions and in the parton distribution functions used."],"supporting_citations":[{"why":"Provides the next-to-next-to-leading-order dijet matrix elements including all subleading color contributions, the key ingredient that makes this the first full-colour NNLO $\\alpha_s$ determination from LHC jet data.","marker":"[7–9]"},{"why":"Supplies the 7 TeV double-differential dijet cross sections from the LHC used in the fit.","marker":"[10]"},{"why":"Supplies the 13 TeV double-differential dijet cross sections from the LHC used in the fit.","marker":"[11]"},{"why":"Supplies the 7 TeV dijet cross sections in the $m_{jj}$–$y_{\\max}$ binning.","marker":"[12]"},{"why":"Supplies the 8 TeV triple-differential dijet cross sections used in the fit.","marker":"[13]"},{"why":"Supplies the 13 TeV dijet cross sections in double- and triple-differential forms; the double-differential variant is the nominal data set.","marker":"[14]"},{"why":"Provides the HERA dijet analysis method and the scale choice $\\mu^2=Q^2+\\langle p_T\\rangle^2$ used for the electron–proton data.","marker":"[23]"},{"why":"Provides the combined proton parton distribution functions and their uncertainty used throughout the fit.","marker":"[39]"}],"fun_headline_variants":["Alpha_s measured up to 7 TeV via NNLO dijets","First strong coupling extraction at 7 TeV scales","Dijet data confirm QCD running from 7 GeV to 7 TeV","NNLO dijets yield alpha_s = 0.1178 ± 0.0022"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The result stands on the assumption that the five-flavor massless scheme, with the top quark decoupled and bottom quarks treated as massless, is accurate at every fitted scale from about 7 GeV to 7 TeV.","fun_headline_variants_meta":{"raw":{"variants":["Alpha_s measured up to 7 TeV via NNLO dijets","First strong coupling extraction at 7 TeV scales","Dijet data confirm QCD running from 7 GeV to 7 TeV","NNLO dijets yield alpha_s = 0.1178 ± 0.0022"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000634,"raw_usage":{"total_tokens":2959,"prompt_tokens":1015,"completion_tokens":1944,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":631,"completion_tokens_details":{"reasoning_tokens":1861}},"tokens_in":631,"tokens_out":1944,"duration_ms":13939,"temperature":1.0,"reasoning_tokens":1861,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T23:02:00.571992+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recompute the multi-TeV $\\alpha_s(\\mu_R)$ values using a scheme that includes top-quark running effects (for example, a matched six-flavor evolution) or assign a dedicated uncertainty for the decoupling approximation by comparing schemes; if the resulting $\\alpha_s(\\mu_R)$ points shift beyond the quoted uncertainties or acquire a trend with $\\mu_R$, the claimed verification of the renormalization group running is not robust.","supporting_citations":[{"cited_title":"Measurement of the triple-differential dijet cross section in proton-proton collisions at sqrt(s) = 8 TeV and constraints on parton distribution functions","cited_arxiv_id":"1705.02628","evidence_quote":"Supplies the 8 TeV triple-differential dijet cross sections used in the fit."}],"review_version":1}