{"id":"7e392b42-0cec-4a27-ad43-0b51c3b1b9b4","arxiv_id":"1909.00490","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Evaluating the b2 and b4 angular observables in the ttbar h rest frame, instead of the lab frame, reduces by about 250 inverse femtobarns the integrated luminosity needed to exclude a pure CP-odd top-quark Yukawa coupling.","lead":"This paper proposes measuring angular correlations of the top, the antitop, and the Higgs boson in their own center-of-mass frame to distinguish a standard Higgs-top coupling from a CP-odd one. In a simulated LHC analysis, this choice saves roughly 250 inverse femtobarns of luminosity compared with lab-frame variables at 90 percent confidence.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Missing closure test: the claimed ~250 fb^-1 gain from the ttbar h rest frame over the lab frame rests on untested reconstruction assumptions; the paper's own pT(h)>200 GeV caveat is the explicit weak point.","rationale":"After reading the full text, the central claim is indeed the luminosity improvement from using the ttbar h rest frame. The load-bearing condition is that the reconstructed rest-frame observables preserve (or enhance) the parton-level shape difference between scalar and pseudoscalar signals. This condition is not directly tested anywhere in the paper: Section IV describes the reconstruction but provides no validation against generator-level truth, and Figures 11-12 show only reconstructed distributions. The reader's weakest assumption identified this reconstruction fidelity issue, including the pT(h)>200 GeV caveat from Section V; I agree with that assessment. My concern sharpens it: the comparison of lab vs rest frame is not apples-to-apples in terms of sensitivity to the kinematic fit, because only the rest-frame variables require the boost. A biased fit could artificially enlarge the separation. The proposed truth-boost test is a minimal check that would settle this. I do not see an internal inconsistency in the parton-level/NLO part of the paper, and the qualitative effect is plausible; however, the quantitative 250 fb^-1 number is not yet fully supported. Hence the verdict remains CONDITIONAL, unchanged from the reader.","tokens_in":19701,"tokens_out":11539,"duration_ms":121010,"concrete_test":"Use the existing MadGraph + Delphes samples. For each event passing the selection, recompute b2(t,tbar) and b4(t,tbar) in the ttbar h frame using the generator-level (truth) ttbar h four-momentum to perform the boost on the reconstructed object momenta, while keeping everything else (selection, object momenta, likelihood construction) identical. Recompute the CL vs luminosity curves. If the ~250 fb^-1 advantage of the rest frame over the lab frame persists with the truth boost, the reconstruction is not the source of the gain; if the gap shrinks or reverses, the headline claim is an artifact of the kinematic fit. A complementary check: overlay reconstructed vs generator-level b2 distributions for the scalar and pseudoscalar signals to quantify the smearing in both frames.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claim, in the abstract and Section V, is that evaluating b2 and b4 in the ttbar h center-of-mass frame cuts the luminosity needed to exclude a pure CP-odd coupling by a few hundred fb^-1 (specifically roughly 250 fb^-1 for b2) relative to the lab frame. The comparison is made at reconstruction level after the dileptonic selection (at least four jets, at least three b-tags) and a full kinematic fit described in Section IV. The rest-frame observables (Eqs. 7-8) require boosting by the four-momentum of the reconstructed ttbar h system, which depends on the fit's neutrino solutions and jet-to-parton assignments. Lab-frame observables use the same reconstructed objects but do not depend on the boost. The paper shows only reconstructed distributions (Figures 11-12) and contains no closure test: no comparison of reconstructed b2/b4 with generator-level values, no migration or response matrices, and no check that the kinematic fit does not preferentially reconstruct events so as to artificially increase the apparent scalar-vs-pseudoscalar separation in the rest frame. Because the fit is a biased estimator, the rest-frame advantage could be in part an artifact of reconstruction rather than a property of the underlying dynamics. The paper itself flags a related limitation in Section V: the resolved h->bb identification degrades for pT(h) greater than about 200 GeV, so the analyzed events are a kinematic subset, and the unqualified few-hundred-fb^-1 saving may not hold inclusively. Without a truth-level closure test, the quantitative claim is not fully supported.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript proposes to evaluate the CP-sensitive angular observables b2 and b4, introduced earlier for the laboratory frame, in the center-of-mass frame of the ttbar h system, in order to probe a possible CP-odd component of the top-quark Yukawa coupling. The signal is parameterized by Eq. (1) with a mixing angle alpha. Parton-level distributions at NLO in QCD are computed with MadGraph5_aMC@NLO, followed by a parton-shower study and a full case study in the dileptonic ttbar h (h to bb) final state, using Delphes fast simulation and a kinematic fit that reconstructs the neutrinos and the ttbar h system. Expected CLs for excluding a pure CP-odd coupling versus the SM scalar hypothesis are evaluated from binned shape distributions with a Poisson likelihood-ratio test as a function of integrated luminosity. The central claim is that evaluating b2 in the ttbar h rest frame reduces the luminosity needed for a 90% exclusion by roughly 250 inverse femtobarns compared with the laboratory frame, and more generally that the rest-frame choice improves sensitivity by a few hundred inverse femtobarns; the paper also finds that top-quark and Higgs pT distributions add no discrimination beyond counting.","tokens_in":20003,"tokens_out":6059,"duration_ms":54743,"significance":"Should the quantitative gain survive reconstruction-level validation, the paper's proposal is valuable: it is an inexpensive, well-defined modification of an existing analysis strategy that could improve the HL-LHC reach for a CP-odd top Yukawa component. The calculation chain is internally consistent and uses standard, publicly documented tools (MadGraph5_aMC@NLO, Pythia6, Delphes, MadAnalysis 5), and the CL procedure is a closed-loop Monte Carlo projection with no parameter fitted to data. The NLO and shower-level shape differences between scalar and pseudoscalar signals are clearly presented, and the claim is falsifiable in the sense that the CL curves are a definite quantitative prediction. The main residual risk concerns the reconstruction-level comparison, which currently lacks closure checks; this is a quantitative validation issue rather than a conceptual error. The overall scope is modest but appropriate for a phenomenological journal.","major_comments":[{"comment":"The central claim—that b2 needs roughly 250 fb^-1 less luminosity at 90% CL when evaluated in the ttbar h rest frame than in the laboratory frame—is obtained from reconstructed distributions (Figures 11–12) and the resulting CL curves (Figure 13), but the paper does not provide any closure test of the reconstruction chain for these observables. There is no comparison of reconstructed b2/b4 with generator-level values, no migration or response matrix, and no check that the kinematic fit (with its neutrino-solution and jet-assignment choices, described only by reference to [54–56]) does not preferentially select events in a way that artificially enlarges the scalar-versus-pseudoscalar separation in the rest frame. Because the rest-frame observables require boosting by the reconstructed ttbar h four-momentum, whereas the lab-frame observables do not, a biased fit could in principle produce part of the claimed gain. The authors should add resolution and migration studies, or an explicit fit-bias check, before the luminosity saving is quoted as a property of the ttbar h rest frame.","section":"Section IV / Figures 11–13"},{"comment":"The quantitative luminosity comparison considers only statistical uncertainties, as stated in Section V ('Only statistical uncertainties are considered'), yet the abstract and conclusions state the few-hundred-fb^-1 saving without that qualification. At the luminosities where the 90% exclusion is projected (roughly 1000–3000 fb^-1), systematic uncertainties on jet energy scale, b-tagging, background normalizations, and signal shapes would typically dominate, and a differential shape comparison between the two frames could easily lose part of the apparent advantage. The paper should either add a crude systematic treatment (for example, nuisance-parameter smearing of the template shapes) or clearly qualify the headline number as a statistical-only projection.","section":"Section V, CL results"},{"comment":"The paper itself flags that the analysis targets events in which the b quarks from the Higgs decay result in two resolved small-radius jets, and that for pT(h) above about 200 GeV the fraction of events failing this requirement becomes significant. This means the quoted gain applies to a kinematic subset of the ttbar h phase space, while the abstract and conclusions present the luminosity reduction without this caveat. The scope of the claim should be stated explicitly in the abstract or in the conclusions, ideally with a remark on whether boosted Higgs reconstruction techniques could extend the gain.","section":"Section V, last paragraph"},{"comment":"The theoretical-uncertainty estimate varies only the factorization scale in the range {mu0,f/2, 2 mu0,f} around Eq. (3); the renormalization scale is not varied. The paper acknowledges this for the cross-section uncertainty, but the same limitation applies to the shape comparisons in Figures 1–5 that motivate the rest-frame choice, and the NLO/LO k-factor variations in Figure 5 show sizable shape distortions. A comment on the expected sensitivity to mu_r variation, or at least a justification for varying only mu_f, would strengthen the motivation for the rest-frame observables.","section":"Section II, scale uncertainty"}],"minor_comments":[{"comment":"The PDF set is called NNPDF2.3 in Section III and NN23LO1 in Section IV; the two labels should be reconciled for clarity.","section":"Sections III and IV"},{"comment":"There is a typo in the introduction: 'posibile' should read 'possible'.","section":"Section I"},{"comment":"In the printed panels, the curves for b4 and for the |Delta eta| observables are difficult to distinguish from one another; different line styles or a more detailed legend would improve readability.","section":"Figures 13–14"},{"comment":"The decomposition in Eq. (6), and its stated bin-by-bin generalization, implicitly assumes the absence of scalar-pseudoscalar interference; a one-sentence derivation or citation would avoid the appearance of an unstated assumption.","section":"Section II, Eq. (6)"},{"comment":"Given that the kinematic-fit details are deferred to references [54–56], the absence of reconstruction-level validation in the present manuscript is more consequential than it would otherwise be; even a short appendix with reconstructed-versus-truth distributions would make the paper self-contained for the central comparison.","section":"Section IV"}],"recommendation":"major_revision","confidential_remarks":"This is a competent phenomenological study with a clear, falsifiable claim. The gating issue is the missing reconstruction-level closure test for the central luminosity comparison between the ttbar h rest frame and the laboratory frame; if the authors can supply it, or alternatively soften the headline claim to a statistical-only, selection-dependent projection, the paper would be publishable. I also note the 2019 submission date; the analysis appears internally consistent and does not duplicate other work, but the heavy reliance on the authors' own earlier papers for the full analysis chain leaves the current manuscript's validation content thin for a claim that is quantitative in nature."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nI read 1909.00490. The useful new thing is taking the Gunion-He observables b2 and b4, previously defined in the lab frame, and evaluating them in the ttbar-h rest frame. The paper shows, with an internally consistent NLO+shower+Delphes+kinematic-fit chain, that the rest-frame versions separate a scalar from a pseudoscalar top Yukawa coupling more cleanly than the lab-frame versions. That direction is probably right; the parton-level and shower-level distributions in Figs. 9-10 show a real separation, and the reconstruction-level CL curves in Fig. 13 align with it. If the central claim holds, the b2 rest-frame version saves roughly 250 fb^-1 at 90% CL in the dilepton channel. That is a useful, practical input for an HL-LHC analysis, even if it does not resolve any deep puzzle.\n\nThe soft spot is the one the stress-test flags, and it is real. The rest-frame boost depends on the four-momentum of the reconstructed ttbar-h system, which comes from a kinematic fit with neutrino solutions and jet-to-parton assignments. The paper shows reconstructed distributions but never a closure test -- no truth-level comparison, no migration matrix, no check of whether the fit biases the scalar/pseudoscalar separation. A biased fit can manufacture or inflate an apparent improvement. The paper's own pT(h)>200 GeV caveat acknowledges that the resolved-jet analysis covers only a kinematic subset, so the unqualified saving is an upper bound in spirit.\n\nTwo smaller issues. The scale uncertainty is estimated only by varying the factorization scale, and the CLs include statistical uncertainties only. That is fine for a first sensitivity projection, but it means the 250/fb number is not a robust prediction. Also, the recipe for arbitrary alpha -- combining scalar and pseudoscalar distributions bin-by-bin as in Eq. (6) -- assumes no interference contribution in each bin. That may hold by CP symmetry for these observables, but it is not argued.\n\nTo its credit, the paper states these limitations openly, and the analysis is a clean, reproducible-in-principle phenomenological study. No code or data are released, which is a pity but not a fatal flaw for a projection of this kind.\n\nThis should go to peer review. The missing closure test is exactly what a good referee would ask for, and it is checkable. I would likely cite the paper for the rest-frame suggestion, not for the headline luminosity saving. Worth a reading-group slot on the strength of the idea alone.","headline":"The rest-frame b2/b4 observables are a genuinely useful idea, and the direction of the improvement is almost certainly right, but the headline 250/fb saving is not yet nailed down because the reconstruction chain is missing a closure test.","tokens_in":20549,"tokens_out":3429,"would_cite":true,"duration_ms":32052,"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":"Evaluating two angular observables in the ttbar-h rest frame reduces the LHC luminosity needed to exclude a purely CP-odd (pseudoscalar) top-quark Yukawa coupling by roughly 250 inverse femtobarns at 90% confidence.","keywords":["top-quark Yukawa coupling","CP violation","pseudoscalar Higgs component","t-tbar-h rest frame","angular observables b2 b4","LHC","dileptonic final state","Higgs boson CP properties"],"falsifier":"Recompute the expected exclusion confidence levels using only events with Higgs transverse momentum above about 200 GeV, where the $h\\to b\\bar b$ jets merge, or switch to a boosted-jet tagging technique; if the roughly 250 inverse femtobarns advantage of the $t\\bar{t}h$ rest frame vanishes or inverts, the gain is an artefact of the resolved-jet reconstruction, while if it survives, the paper's central claim is confirmed in the regime it explicitly left uncovered.","tokens_in":19516,"feed_emoji":"⚛️","tokens_out":8869,"duration_ms":71190,"temperature":0.7,"pith_summary":"The paper argues that the rest frame of the $t\\bar{t}h$ system is a substantially better place to measure the CP properties of the top-quark Yukawa coupling than the laboratory frame. It defines two angular observables, $b_2$ and $b_4$, that project the $t$, $\\bar{t}$, and $h$ momenta onto the beam axis and transverse plane, and shows at next-to-leading order in QCD that their distributions clearly separate a pure scalar (Standard Model) coupling from a pure pseudoscalar one when evaluated in the $t\\bar{t}h$ rest frame. Using dileptonic $t\\bar{t}h$ events with $h\\to b\\bar b$, a fast detector simulation, and a full kinematic fit, the paper reports that the $b_2$ observable requires about 250 inverse femtobarns less luminosity to exclude the pure CP-odd hypothesis at 90% confidence than the same observable measured in the lab frame. It also finds that the Higgs and top transverse-momentum spectra carry no more discriminating information than a simple counting experiment.","feed_headline":"Boosting to the ttbar-h rest frame sharpens Higgs CP test by 250 fb^-1","feed_subtitle":"Angular observables in the ttbar-h rest frame save about 250 fb^-1 of LHC data for a 90% exclusion.","key_machinery":"The load-bearing objects are the angular observables $b_2(i,j) = ((\\vec p_i\\times \\hat k_z)\\cdot(\\vec p_j\\times \\hat k_z))/(|\\vec p_i||\\vec p_j|)$ and $b_4(i,j) = p_{i,z}p_{j,z}/(|\\vec p_i||\\vec p_j|)$, where $i,j$ run over $t$, $\\bar t$, $h$ and $\\hat k_z$ is the beam axis; they measure the transverse-dot and longitudinal correlations of the heavy final-state momenta. The paper's proposal is to evaluate them in the $t\\bar{t}h$ center-of-mass frame, which requires full four-momentum reconstruction of the three heavy particles. This frame change concentrates the spin-correlation information that distinguishes a scalar from a pseudoscalar coupling: in the rest frame the $b_2(t,\\bar t)$ distribution is single-peaked for the SM scalar and clearly different for the pseudoscalar, whereas the lab-frame shapes are much closer. The machinery that turns this shape difference into a number is a binned likelihood-ratio test: 100,000 Poisson pseudo-experiments per luminosity point yield an expected confidence level for excluding the pure CP-odd scenario as a function of integrated luminosity, with only statistical uncertainties.","core_discovery":"The paper's central claim is that the two observables $b_2(i,j)$ and $b_4(i,j)$, defined from the transverse and longitudinal projections of the $t$, $\\bar{t}$, and $h$ momenta relative to the beam axis, become markedly more discriminating between a CP-even and a CP-odd top-quark Yukawa coupling when they are evaluated in the rest frame of the $t\\bar{t}h$ system rather than in the laboratory frame. At parton level, with next-to-leading-order QCD corrections, the normalized $b_2(t,\\bar t)$ and $b_4(t,\\bar t)$ distributions show clear shape differences between the pure scalar and pure pseudoscalar cases only in the $t\\bar{t}h$ frame, not in the lab frame. After parton showering, a dileptonic selection with at least four jets and at least three b-tags, fast detector simulation, and a full kinematic fit, the reconstructed distributions still show the separation. The paper quantifies the gain through expected confidence levels: for the $b_2$ observable, reaching the 90% exclusion of the pure CP-odd hypothesis needs roughly 250 inverse femtobarns less integrated luminosity when the observable is evaluated in the $t\\bar{t}h$ rest frame than in the lab frame. It also reports that the transverse-momentum spectra of the top quarks and the Higgs boson provide no better discrimination than a counting experiment.","pith_inferences":["This result suggests a general recipe: for any process producing three or more on-shell heavy particles, spin-sensitive angular variables should be constructed in the process rest frame, where the boost of the initial state no longer dilutes the angular correlations.","Because the paper provides the interpolation rule $\\sigma_\\alpha = \\sigma_{\\rm scalar}\\cos^2\\alpha + \\sigma_{\\rm pseudoscalar}\\sin^2\\alpha$, the same machinery could be extended from a binary scalar-versus-pseudoscalar exclusion to a direct fit of the mixing angle $\\alpha$, including a measurement of the sign of the CP-odd component.","A dedicated study with a full, Geant4-based detector simulation and systematic uncertainties would test whether the roughly 250 $\\mathrm{fb}^{-1}$ advantage survives more realistic reconstruction, including jet energy scale and b-tagging efficiencies."],"forward_implications":["The same rest-frame evaluation can be applied to the $b_4$ observable and to other angular correlations, so the sensitivity gain is not limited to a single variable.","Combining the dileptonic channel with the single-lepton $t\\bar{t}h$ final state should reduce the luminosity required for a given exclusion level by roughly a factor of five relative to the dileptonic analysis alone.","The transverse momentum distributions of the top quarks and Higgs boson add no discriminating power beyond the total cross-section count, so future direct CP measurements should prioritize angular observables over $p_T$ spectra.","The quoted luminosity saving applies to events where the Higgs decay $b$ quarks are resolved as two small-radius jets; boosted Higgs bosons with $p_T(h)$ above about 200 GeV require separate treatment and are not covered by the gain."],"supporting_citations":[{"why":"introduced the $b_2$ and $b_4$ angular observables that this paper redefines in the $t\\bar{t}h$ rest frame.","marker":"[41]"},{"why":"provided the lab-frame NLO and NLL differential distributions and total cross sections for $t\\bar{t}h$ that serve as the baseline.","marker":"[40]"},{"why":"the Monte Carlo generator used for NLO event generation of the signal and background samples.","marker":"[49]"},{"why":"supplied the implementation of the scalar-plus-pseudoscalar top Yukawa coupling used in the calculation.","marker":"[50]"},{"why":"the dileptonic $t\\bar{t}h$ analysis with full kinematic reconstruction that the case study follows.","marker":"[54-56]"},{"why":"the fast detector simulation producing the reconstruction-level distributions used for the CL study.","marker":"[61]"},{"why":"the HL-LHC study quoted for the factor-five sensitivity gain expected from adding single-lepton channels.","marker":"[65]"}],"fun_headline_variants":["ttbar-h rest frame saves 250 fb^-1 for Higgs CP test","Rest-frame observables cut LHC luminosity for CP test by 250 fb^-1","Kinematic frame boosts Higgs CP sensitivity, saves 250 fb^-1","CP test in ttbar-h frame needs 250 fb^-1 less luminosity"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The quantitative luminosity gain assumes that the full four-momentum reconstruction of the $t$, the $\\bar t$, and the $h$ in the $t\\bar{t}h$ rest frame preserves enough of the parton-level shape difference after detector simulation, the dileptonic selection, and the kinematic fit; the paper itself notes that for Higgs transverse momentum above about 200 GeV, the resolved-jet $h\\to b\\bar b$ identification degrades, so the quoted gain does not cover the boosted regime.","fun_headline_variants_meta":{"raw":{"variants":["ttbar-h rest frame saves 250 fb^-1 for Higgs CP test","Rest-frame observables cut LHC luminosity for CP test by 250 fb^-1","Kinematic frame boosts Higgs CP sensitivity, saves 250 fb^-1","CP test in ttbar-h frame needs 250 fb^-1 less luminosity"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001311,"raw_usage":{"total_tokens":5504,"prompt_tokens":1266,"completion_tokens":4238,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":882,"completion_tokens_details":{"reasoning_tokens":4154}},"tokens_in":882,"tokens_out":4238,"duration_ms":27261,"temperature":1.0,"reasoning_tokens":4154,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T05:51:59.775615+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recompute the expected exclusion confidence levels using only events with Higgs transverse momentum above about 200 GeV, where the $h\\to b\\bar b$ jets merge, or switch to a boosted-jet tagging technique; if the roughly 250 inverse femtobarns advantage of the $t\\bar{t}h$ rest frame vanishes or inverts, the gain is an artefact of the resolved-jet reconstruction, while if it survives, the paper's central claim is confirmed in the regime it explicitly left uncovered.","supporting_citations":[{"cited_title":"Pseudoscalar couplings in $t \\bar{t} H$ production at NLO+NLL accuracy","cited_arxiv_id":"1707.01803","evidence_quote":"provided the lab-frame NLO and NLL differential distributions and total cross sections for $t\\bar{t}h$ that serve as the baseline."}],"review_version":1}