{"id":"4f325196-7915-4318-8dc7-2a395d043c36","arxiv_id":"2509.19038","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":9,"one_line_summary":"First search for electroweak ttWj production at 13 TeV sets observed (expected) 95% CL upper limits of 251 fb (230 fb) versus the SM prediction of 47.7 fb and constrains the O_Ht and O_HQ^(1) EFT operators.","lead":"ATLAS researchers searched for a rare process, electroweak top-quark pair production together with a W boson and a jet, using 140 inverse femtobarns of LHC data. They found no excess, set an upper limit about five times the predicted rate, and placed new constraints on two effective field theory parameters.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Unvalidated shape of the |Δη_j| discriminant for ttW QCD background could bias the extracted ttWjEW signal strength; the dedicated control regions constrain only the integral, not the SR bin-by-bin shape.","rationale":"The reader's weakest_assumption points to the MC modeling of |Δη_j|; I agree. The analysis is well executed, with seven CRs and a thorough systematic treatment, and the null result is consistent with the SM. The main concern is that the shape of the discriminating variable is not directly cross-checked in data, so the limit is contingent on the MC shape. However, given the small expected signal (15.7 events vs 379 total SR data) and the already large systematic uncertainties, a modest shape mis-modeling would not alter the qualitative conclusion of consistency with the SM. The EFT interpretation uses additional handles (ttZ, pT(l0) bins) and is not the central claim. Thus the concern does not invalidate acceptance; it could be addressed in a revision with a simple differential validation, but the current result stands.","tokens_in":58741,"tokens_out":10331,"duration_ms":108473,"concrete_test":"Split the 4J-hiΔη CR (which has the same |Δη_j|>2 requirement as the SR but orthogonal H_T, pT(l0), pT(l1), m_ll selection) into the same three |Δη_j| bins used in the SR, and compare data to the post-fit background-only prediction with the ttW QCD normalization constrained elsewhere. If the data-to-prediction ratio is consistent across bins, the shape is validated. Alternatively, perform a spurious-signal test: generate pseudo-experiments from the background-only hypothesis with the ttW QCD |Δη_j| template varied within the Sherpa/MadGraph and Powheg/Herwig differences, and check that the fitted mu distribution is centered at zero and that the 95% upper limit shifts by less than ~20% relative to the nominal.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The cross-section limit and mu extraction rely on the three-bin |Δη_j| distribution (SR, Table 1, Section 8) to separate ttWjEW from ttW QCD. The ttW QCD normalization is free, constrained by 4J-loΔη and 4J-hiΔη CRs (Table 2), but these are binned in channel/charge/b-tag, not in |Δη_j|. Thus the shape of |Δη_j| inside the SR is taken entirely from the Sherpa NLO prediction. The assigned modeling uncertainties (Section 7) compare Sherpa vs MadGraph5_aMC@NLO+Pythia8 and Powheg+Pythia8 vs Powheg+Herwig7; both alternatives share common NLO matching assumptions and are not validated against data in a Δη-binned control region. If the true ttW QCD |Δη_j| shape is harder or softer than predicted, the fitted mu and the resulting 251 fb (230 fb) upper limit would shift beyond the quoted 2.4/3.0 uncertainty, since the background normalization would absorb part of the shape difference. This is the weakest link in the central measurement.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper reports a search for electroweak ttbarWj production at sqrt(s)=13 TeV using 140 fb^-1 of ATLAS Run 2 data. The analysis selects same-charge dilepton events with at least four jets and one b-tag, and exploits the forward-jet pseudorapidity difference |Δη_j| to separate the electroweak ttbarWj signal from the dominant QCD ttbarW background. A binned profile-likelihood fit is performed over three signal-region bins and seven control regions, with data-driven estimates for fake leptons (matrix method), charge flips, and photon conversions. The observed (expected) 95% CL upper limit on the ttbarWj_EW cross section is 251 fb (230 fb) versus a Standard Model prediction of 47.7 fb; the fitted signal strength is mu = 0.9 +2.4/-3.0. The paper also presents EFT limits on the Wilson coefficients c_Ht and c_HQ^(1) from ttbarWj and ttbarZ production, using pseudo-dataset confidence intervals to account for non-Wilks behavior.","tokens_in":59023,"tokens_out":8731,"duration_ms":70335,"significance":"If the results are correct, this is the first direct search for the tW-scattering vertex and provides the first constraints on O_Ht and O_HQ^(1) from ttbarWj production. The analysis is technically thorough: it uses the full Run 2 dataset, a seven-control-region fit with data-driven background estimates and a full systematic uncertainty evaluation, and it correctly handles the breakdown of Wilks' theorem in the EFT interpretation with pseudo-dataset thresholds. The energy-dependent sensitivity study is a useful case study for multi-process, multi-operator EFT interpretations. However, three load-bearing points need clarification or additional validation: the data/MC validation of the |Δη_j| shape for the dominant ttW QCD background, the parameterization of per-bin EFT normalization factors in Section 9, and an apparent inconsistency between the text and Table 6 regarding whether the SM point is contained in the EFT confidence intervals.","major_comments":[{"comment":"The signal extraction relies on the three-bin |Δη_j| distribution in the SR (after the |Δη_j|>2.0 requirement), while the ttW QCD normalization is constrained by the 4J-loΔη and 4J-hiΔη control regions, which are binned in channel, lepton charge, and b-jet multiplicity but not in |Δη_j|. The shape of the dominant ttW QCD background across the three SR bins is therefore taken entirely from the Sherpa NLO prediction, and the modeling uncertainties in Section 7 compare generators that share common NLO matching assumptions. No data/MC agreement plot binned in |Δη_j| is shown. A mismatch in the predicted |Δη_j| slope would be partially absorbed by the free ttW QCD normalization and could bias mu_ttWjEW and the quoted limit. Please add a Δη-binned validation region orthogonal to the SR, or otherwise demonstrate that the assigned shape uncertainties cover the observed data/MC differences in a |","section":"Sections 5 and 8; Tables 2 and 4"},{"comment":"The parameterization of the EFT signal is not clear. The text states that when an EFT operator is considered, the floating normalization factor for the process is replaced by a dedicated cross-section uncertainty nuisance parameter, and \"a new free-floating parameter is introduced to parameterize the effect of the EFT operator on the ttWjEW (ttZ) normalization in each bin of the analysis.\" If each analysis bin has an independent, unconstrained normalization for the EFT contribution, then the shape information in pT(l0) and in the CRs is not used, and the energy-dependent sensitivity claimed in Figure 11 could not arise. If these per-bin parameters are actually constrained (e.g., by the MC prediction as a function of the Wilson coefficient), the constraints must be specified. Please clarify the exact role of these per-bin parameters and how the Wilson coefficient is extracted given their","section":"Section 9"},{"comment":"The text says: \"In the most complex fit ... the 95% confidence interval contains the SM expectation.\" This appears to be contradicted by Table 6, which for the simultaneous ttWjEW+ttZ fit gives c_Ht 95% CI = [−10,−0.96] ∪ [1.5,7.3] and c_HQ = [−12,0.33] ∪ [3.4,12]. The c_Ht interval excludes zero, so the SM point (c_Ht=c_HQ=0) is not contained in the marginal intervals. If the marginal intervals are correct, then the statement that the interval contains the SM expectation is wrong, and the paper would be reporting a 95% exclusion of the SM in the most complex EFT fit. This discrepancy must be resolved — either the text or the table is erroneous, and the conclusion about consistency with the SM needs to be stated accurately.","section":"Section 9 and Table 6"},{"comment":"The EFT samples used for the interpretation are generated at LO with MadGraph and processed with the AtlFast II fast detector simulation, while the nominal signal and background samples use full Geant4 simulation. No systematic uncertainty is described for the difference between fast and full simulation in the EFT signal shapes. Since the EFT limits depend on the pT(l0) distribution and on migration between bins, the possibility of a fast-simulation bias in acceptance or shape should be quantified or otherwise justified. At minimum, the paper should state whether a fast/full comparison was performed and what its impact is.","section":"Sections 3 and 9"}],"minor_comments":[{"comment":"The row for the signal region appears garbled: \"|Δη_j|—>2.0ÍpT(j)—>250GeV\" should be split into two separate entries with proper formatting.","section":"Table 1"},{"comment":"The definition of Δη_j is awkward: \"the difference in pseudorapidity between the most forward non-b-tagged jet and the jet with which it produces the highest invariant mass.\" It would be clearer to specify whether the invariant mass is computed with the most forward jet or with each pair of jets, and to define the pairing prescription explicitly.","section":"Section 5"},{"comment":"The phrase \"considering the impact of both O_Ht or O_HQ\" should read \"both O_Ht and O_HQ\".","section":"Section 9"},{"comment":"The signal is shown multiplied by 20, but no numerical scale or legend entry indicates the factor for the dashed line. Please add a note in the caption or legend.","section":"Figure 4"}],"recommendation":"major_revision","confidential_remarks":"The paper is a workmanlike ATLAS search with a potentially interesting first result on tW scattering and new EFT constraints. The main issues are not obvious fakes or circular reasoning, but they are load-bearing to the central claims: the unvalidated |Δη_j| shape for ttW QCD, the unclear EFT per-bin parameterization, and the text/table inconsistency about whether the SM point is excluded. These are fixable with further analysis or clarification, but they need to be addressed before the paper can be accepted."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First thing you should know: this is the first direct search for electroweak ttWj production, and the result is a clean null. The observed limit is 251 fb against a 47.7 fb SM prediction, about a factor of five above the SM. The analysis is solid, and the EFT interpretation is the most interesting part.\n\nWhat's genuinely new: the process contains a tW-scattering vertex, and this is the first experimental probe of it. The EFT case study shows that the energy-dependent sensitivity of ttWj_EW breaks the degeneracy between O_Ht and O_HQ that plagues ttZ-only fits. They also do the right statistical thing by using pseudo-datasets for EFT intervals, because the likelihood is non-Gaussian. The analysis itself is careful: seven control regions, data-driven fake estimates, profile likelihood. Nothing sloppy here.\n\nThe soft spot is the one flagged in the stress test. Signal is separated from ttW QCD by a three-bin |Δη_j| distribution in the SR. The ttW QCD CRs are binned in channel, charge, and b-tag, not in Δη; they only split at |Δη_j|=2.0. So the three-bin shape inside the SR comes entirely from Sherpa NLO, with generator-comparison systematics but no data validation in Δη. That is a legitimate concern. It's not fatal: the limit is loose enough that a shape miscalibration would shift the bound rather than produce a false signal, and the ttW QCD modeling uncertainty (±1.2 on mu) is one of the largest. But if I were refereeing, I'd ask for a Δη-binned validation in a CR, or at least a demonstration that the fitted mu doesn't move when the SR binning is changed.\n\nOne thing the reader's summary omitted: the EFT fits actually prefer non-zero Wilson coefficients, driven by a small excess in the highest pT bin in the SR. The 95% CL interval contains the SM, so it's not evidence, but the paper reports it honestly and it's worth knowing.\n\nWho should read it: anyone doing top EFT or rare-process searches. It's a good template and a necessary first measurement. I'd send it to a serious referee.","headline":"First direct search for ttWj_EW; a clean null result with a solid EFT interpretation, though the discriminating-variable shape for the dominant background relies on unvalidated MC against data in Δη.","tokens_in":59518,"tokens_out":3069,"would_cite":true,"duration_ms":26784,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["14.65.Ha"],"model":"deepseek-v4-flash","headline":"This paper reports the first direct search for electroweak ttWj production, which probes the tW-scattering vertex, setting an observed 95% CL upper limit on the cross section of 251 fb (expected 230 fb), consistent with the Standard Model p","keywords":["electroweak ttW production","tW scattering","forward jet","same-charge dilepton","SMEFT","effective field theory","top-quark couplings","LHC"],"falsifier":"A comparison of the observed |Δη_j| distribution in a high-statistics QCD ttW control region with the Sherpa prediction (or with an alternative generator such as an NLO merging prediction) would falsify the shape assumption if a significant discrepancy appeared. Alternatively, a reanalysis of the same data using a fully independent discriminating variable — for example, the transverse mass of the ttW system or the jet multiplicity — that yields a signal strength inconsistent with 0.9 would undermine the claimed consistency with the Standard Model.","tokens_in":58590,"feed_emoji":"⚛️","tokens_out":6709,"duration_ms":49162,"temperature":0.7,"pith_summary":"The paper attempts to establish electroweak production of a top-antitop pair with a W boson and an extra jet (ttWj_EW) as a measurable process that directly probes the tW-scattering vertex, the interaction of a top quark with a W boson, a part of the Standard Model never before isolated. Analyzing 140 fb^-1 of 13 TeV LHC data in same-charge dilepton events with a forward jet, the search finds a signal strength mu = 0.9 +2.4/-3.0, consistent with the Standard Model, and sets an observed 95% CL upper limit of 251 fb on the cross section versus the SM prediction of 47.7 fb. The paper further claims the first constraints on the dimension-six SMEFT operators O_Ht and O_HQ^(1) from this process, exploiting the quadratic energy growth of their contributions to lift a degeneracy between them that exists in ttZ-only interpretations. If correct, this establishes a new window into electroweak symmetry breaking and a template for combined multi-process EFT analyses.","feed_headline":"First tW-scattering probe sets 251 fb limit","feed_subtitle":"No excess over the Standard Model; new bounds on top-quark EFT couplings.","key_machinery":"|Δη_j|, the absolute difference in pseudorapidity between the most forward non-b-tagged jet and the jet with which it forms the highest invariant mass, is the central discriminating variable: electroweak-produced events contain a spectator jet at higher pseudorapidity than the QCD ttW background, so the shape of this distribution separates signal from background in a binned profile-likelihood fit. For the EFT interpretation, the signal region is re-binned in the leading-lepton pT to exploit the quadratic energy growth of the O_Ht and O_HQ^(1) contributions, and confidence intervals are derived from pseudo-datasets to correct for the breakdown of Wilks' theorem.","core_discovery":"The central claim is that the tW-scattering vertex can be probed directly in electroweak ttWj production, and that this process can be extracted from the much larger QCD ttW background using the pseudorapidity difference |Δη_j| between the most forward non-b-tagged jet and the jet with which it forms the highest invariant mass. The measured signal strength is 0.9 +2.4/-3.0, and the observed (expected) 95% CL upper limit on the cross section is 251 fb (230 fb), compared with the Standard Model prediction of 47.7 fb. This is the first direct bound on the tW-scattering vertex. Additionally, the analysis sets limits on the SMEFT operators O_Ht and O_HQ^(1), which modify the top-quark couplings t","pith_inferences":["Editorial inference: the forward-jet pseudorapidity technique may carry over to other electroweak top-quark processes (e.g., tqZ, tqH with a spectator jet), where a similar forward jet separates the EW component from the QCD one.","Editorial inference: a dedicated measurement of the |Δη_j| shape in a high-statistics ttW-enriched region, comparing Sherpa with an alternative NLO merging prediction, would directly test the load-bearing modeling assumption.","Editorial inference: the observed consistency with the SM suggests that the quadratic EFT terms are currently unconstrained; with more data, a full SMEFT global fit including ttWj_EW, ttZ, and ttH could break remaining flat directions among top-quark operators."],"forward_implications":["The first direct bound on the tW-scattering vertex means future LHC data can turn an excess or deficit here into a concrete test of new physics tied to electroweak symmetry breaking.","Combining the ttWj_EW and ttZ channels removes a degeneracy between O_Ht and O_HQ^(1), so the two operators can be constrained separately rather than only in combination.","The consistency with the Standard Model supports current higher-order predictions for ttW production, tempering the long-standing tension between direct ttW measurements and theory.","The strong energy growth in high-pT bins establishes that the sensitivity of the search comes disproportionately from boosted events, motivating targeted high-energy triggers and analyses."],"fun_headline_variants":["First direct tW-scattering probe yields 251 fb limit","ATLAS places first direct bound on tW-scattering vertex","No excess in electroweak ttWj; 95% CL limit 251 fb","ATLAS sets limits on top-quark EFT operators via ttWj","Measurements of electroweak ttWj find no deviation from SM"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The result hinges on the Monte Carlo prediction for the shape of |Δη_j| in the QCD ttW background; if the forward-jet pseudorapidity distribution is mis-modeled in the signal region, the fitted signal strength and the upper limit would be biased.","fun_headline_variants_meta":{"raw":{"variants":["First direct tW-scattering probe yields 251 fb limit","ATLAS places first direct bound on tW-scattering vertex","No excess in electroweak ttWj; 95% CL limit 251 fb","ATLAS sets limits on top-quark EFT operators via ttWj","Measurements of electroweak ttWj find no deviation from SM"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000821,"raw_usage":{"total_tokens":3510,"prompt_tokens":907,"completion_tokens":2603,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":651,"completion_tokens_details":{"reasoning_tokens":2508}},"tokens_in":651,"tokens_out":2603,"duration_ms":15247,"temperature":1.0,"reasoning_tokens":2508,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-04T15:36:53.743694+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A comparison of the observed |Δη_j| distribution in a high-statistics QCD ttW control region with the Sherpa prediction (or with an alternative generator such as an NLO merging prediction) would falsify the shape assumption if a significant discrepancy appeared. Alternatively, a reanalysis of the same data using a fully independent discriminating variable — for example, the transverse mass of the ttW system or the jet multiplicity — that yields a signal strength inconsistent with 0.9 would undermine the claimed consistency with the Standard Model.","supporting_citations":[],"review_version":1}