{"id":"9a99b541-efc6-4562-bb3a-548b5324d756","arxiv_id":"1303.3758","paper_version":3,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Full GEANT4 simulation of a 500 GeV CLIC detector gives an expected top quark mass precision near 100 MeV, with 34 MeV statistical from a threshold scan and the remainder dominated by theory and luminosity-spectrum systematics.","lead":"A full detector simulation shows that a CLIC-style linear collider could measure the top quark mass to roughly 100 MeV, both by reconstructing decay products at 500 GeV and by scanning the top pair production threshold. The threshold result is the one delivered in a theoretically well-defined mass scheme, which is what makes it usable for Standard Model stability calculations.","discovery_kind":"unclear","skeptic_critique":{"model":"selfhosted/deepseek-v4.1-flash","headline":"The threshold-scan budget rests on a flat signal-efficiency and background model measured only at 352 GeV (§6.2); the 34 MeV statistical figure is a toy closure that cannot absorb efficiency variation across 344–353 GeV, so the 'below 100 MeV' total is not yet established.","rationale":"The paper is a careful, honest simulation study. The above-threshold result (80 MeV statistical, Table 2) is well supported by the fit and the explicit jet-energy-scale studies, and the authors correctly decline to claim a theory-usable mass there. The CLIC-versus-ILC comparison (§6.4) is a genuine result and is handled fairly: the 20% statistical advantage of ILC is shown to be immaterial once systematics are included. The authors are also forthright about the unresolved invariant-mass-to-short-distance-mass conversion. The conditional verdict is right, and the reader's weakest_assumption — the flat efficiency/background extrapolation from 352 GeV — is the correct target. I would sharpen two points. First, the 34 MeV is a closure of the toy: because the data are generated with the same epsilon and background used in the templates, the study cannot reveal a bias from those assumptions, and the 'no bias' demonstration in Figure 6 is guaranteed by construction. Second, the acknowledged luminosity-spectrum term (75 MeV at 20% RMS, §6.3.2) is quantitatively the largest single entry in the budget and is admittedly not fully studied; in quadrature with 34 MeV it already gives about 82 MeV, so the 'below 100 MeV' conclusion holds only if the true spectrum uncertainty is at or below about 20% RMS. The paper should either propagate a realistic spectrum uncertainty or state the conclusion as conditional on it. Neither gap overturns the physics message — a linear collider reaches the 100 MeV scale and systematics, not statistics, set the floor — but both should be closed before the 34 MeV or the 'below 100 MeV' total is used as a design target. I therefore keep CONDITIONAL.","tokens_in":19090,"tokens_out":9895,"duration_ms":84677,"concrete_test":"Re-run the existing GEANT4 + particle-flow + selection chain on signal (and the dominant WW/ZZ/qqbar backgrounds) at three additional energies spanning the scan, e.g. 346, 349 and 351 GeV, and measure the selection efficiency and residual background at each. Regenerate the ten toy data points using the energy-dependent efficiency and background, then repeat the 2D template fit. If the fitted 1S mass shifts by more than about 15 MeV, or the statistical error grows beyond about 40 MeV, the flat-efficiency assumption is load-bearing and the 34 MeV figure is not robust. A cheaper first pass: vary epsilon by plus or minus 5% linearly across the scan in the existing toy and report the induced mass shift.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing number is the 34 MeV statistical precision of the 1S-mass fit (Table 3), which underpins the abstract's 'total uncertainties of around 100 MeV' (abstract:s4). That number comes from a toy study in which the ten 'data points' are generated as NNLO cross section times a single selection efficiency (70.2%) plus a single background level (73 fb), both measured only at 352 GeV (§6.2: 'adding background events assuming a constant level over the considered energy range of 10 GeV'). The fit templates use the same constant efficiency and background, so the toy is a closure: it recovers the input mass by construction, and the quoted 34 MeV reflects only Poisson fluctuations of that model, not any bias from the model itself. Near threshold the ttbar kinematics change rapidly — at 344–347 GeV the tops are nearly at rest, jets are soft and close, and b-tagging and kinematic-fit success plausibly differ from 352 GeV — so a real experiment extracting sigma = N/(epsilon L) with a constant epsilon would distort the turn-on shape and shift the fitted mass. The paper does not quantify this. The acknowledged luminosity-spectrum term (§6.3.2: 75 MeV for a 20% RMS variation, 'a full investigation has not yet been performed') is larger still and compounds the issue; the conclusion 'below 100 MeV' (§7) is therefore conditional on two unmeasured modeling inputs.","agreement_with_reader":"agree"},"referee_report":{"model":"selfhosted/deepseek-v4.1-flash","summary":"The manuscript reports a full-simulation study of the expected precision on the top quark mass at a 500 GeV stage of CLIC, using GEANT4 with the CLIC_ILD detector model, particle-flow reconstruction, and overlaid γγ→hadrons beam-induced background at realistic levels. Two routes are pursued on the same reconstructed event sample: (i) direct reconstruction of the invariant mass of the top decay products at √s = 500 GeV, with a kinematic fit imposing equal top masses and W-mass constraints, followed by a binned-likelihood background rejection and an unbinned fit of a Breit-Wigner convolved with a three-Gaussian resolution plus a high-mass tail, giving a combined statistical uncertainty of 80 MeV for 100 fb⁻¹ (Table 2); and (ii) a ten-point threshold scan between 344 and 353 GeV with 10 fb⁻¹ per point, in which the 1S mass and αs are extracted simultaneously from a two-dimensional template fit to NNLO TOPPIK cross sections folded with ISR and the CLIC luminosity spectrum, giving 34 MeV statistical and 5–8 MeV theory normalization on the mass (Table 3). The threshold analysis is repeated with the ILC luminosity spectrum, yielding 27 MeV, and the paper concludes that a total uncertainty below 100 MeV is reachable in a well-defined mass scheme.","tokens_in":19311,"tokens_out":5180,"duration_ms":68826,"significance":"The study is a careful and unusually complete feasibility estimate: it is not a generator-level exercise but a full chain of GEANT4 simulation, beam-induced background overlay (300 bunch crossings of γγ→hadrons), particle-flow reconstruction, flavour tagging, kinematic fitting, and likelihood-based background rejection, applied consistently to signal and to the WW, ZZ, qq̄ and WWZ backgrounds listed in Table 1. The comparison of the CLIC and ILC luminosity spectra within the same analysis chain is a genuinely useful addition, and the paper is candid that the direct-reconstruction result is a PYTHIA generator mass whose conversion to a short-distance scheme is not under control, which is exactly the right caveat to state. The authors also acknowledge, rather than hide, that the luminosity-spectrum uncertainty has not been fully investigated and that colour-reconnection effects are outside the scope of the paper. If the threshold-scan number holds up, it is a meaningful statement about the physics reach of a staged linear collider, and the paper's framing of the two methods as complementary — the threshold measurement constraining the non-perturbative interpretation of the invariant-","major_comments":[{"comment":"The ten pseudo-data points are built from a single signal selection efficiency of 70.2% and a single residual background cross section of 73 fb, both determined from fully simulated events at 352 GeV and then applied unchanged over the whole 344–353 GeV range ('adding background events assuming a constant level over the considered energy range of 10 GeV'). Near threshold the top quarks are nearly at rest, jets are soft and overlapping, and the b-tagging and kinematic-fit success rates plausibly differ from their 352 GeV values; an efficiency that varies with energy distorts the shape of the cross-section turn-on, which is precisely the observable that carries the mass sensitivity. Because the same constant efficiency and background enter the fit templates, the toy cannot reveal this effect. The 34 MeV of Table 3 is therefore conditional on an unquantified modelling assumption. The authors should either simulate two or three additional energy points to bound the variation of efficiency and background across the scan, or demonstrate by an explicit distortion test (e.g. injecting a few-percent linear or threshold-shaped efficiency slope) that the fitted mass is insensitive at the quoted precision.","section":"§6.2","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The paper states that a 20% uncertainty on the RMS width of the main luminosity peak produces a top mass uncertainty of approximately 75 MeV, that 'a full investigation has not yet been performed', and then concludes in §7 that the total uncertainty is 'below 100 MeV'. The luminosity-spectrum term alone is more than twice the 34 MeV statistical error and, added in quadrature with the 5–8 MeV theory normalization and the 18 MeV background-subtraction term, leaves very little headroom under the quoted ceiling. The conclusion should be supported by an explicit error-budget table in which each contribution is listed with its assumed input and combined in quadrature, and the abstract and §7 should say 'expected to be below 100 MeV under the assumed luminosity-spectrum control' rather than presenting the number as achieved. As written, the abstract sentence 'total uncertainties of around 100 MeV are achieved' overstates what the body of the paper establishes.","section":"§6.3.2 and §7","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The 80 MeV combined statistical precision of Table 2 is smaller than every jet-energy-scale bias quoted in §5.2: a 2% shift of all jets gives 200 MeV and a 2% shift of b-jets alone gives 350 MeV. The paper argues that in-situ calibration from the reconstructed W mass and from Z→bb̄ in di-boson and radiative-return events will bring the jet energy scale to better than 1% and the resulting top mass systematics below 50 MeV, but this is asserted rather than demonstrated, and no event yield or expected calibration precision is shown. Since the central claim of the 500 GeV analysis is an experimental uncertainty 'around 100 MeV', the calibration performance is load-bearing. The authors should present the required jet-energy-scale precision explicitly as a requirement on the detector and reconstruction, state the assumed integrated luminosity and event sample on which the calibration would rest, and make clear in the abstract and conclusions that the 80 MeV is a statistical figure whose systematic counterpart depends on achieving that calibration.","section":"§5.2","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The statement that 'the fit itself does not introduce a bias on the results', illustrated by the 5000-trial closure of Fig. 6, holds only within the model that generated the pseudo-data: the same TOPPIK cross section, the same ISR folding, the same luminosity spectrum, and the same constant efficiency and background are used for both the pseudo-data and the templates. This is a legitimate and standard sensitivity estimate, but it should be labelled as such, because a reader could otherwise read the closure as evidence that the extracted 1S mass is unbiased against realistic mismodelling. A short explicit sentence in §6.2 or §6.3 stating that the toy is a closure and that modelling biases are addressed separately (or not at all) would remove the ambiguity.","section":"§6.3 and Fig. 6","authors_acknowledge":null,"authors_acknowledge_p":null}],"minor_comments":[{"comment":"The word 'non-perturbartive' appears in the sentence beginning 'Given the influence of non-perturbartive effects on the invariant mass'; it should read 'non-perturbative'.","section":"§1","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"Reference [14] (SiD) is given as arXiv:0911.006, which is not a valid identifier; the correct number is arXiv:0911.0006.","section":"References","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The bias test using a 175 GeV / 1.5 GeV sample is described as showing deviations 'less than the statistical uncertainty', but the text also notes that the statistical uncertainty on the resolution parameters is comparable to the overall statistical uncertainty. The test therefore has limited discriminating power, and the paper should say so rather than concluding that 'there is no indication for a bias' without qualification.","section":"§5.2","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The semi-leptonic mass, 174.293 ± 0.137 GeV, deviates from the generator value by about 2σ. The paper notes the deviation but does not discuss whether it points to an inadequacy of the fit function (for example the single-Gaussian tail or the fixed resolution parameters) or is simply a fluctuation; a brief statement would help the reader judge the reliability of the quoted central value and error.","section":"Table 2","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The fit success rates quoted in §4.5 (37% fully-hadronic, 47% semi-leptonic at 500 GeV) and in §6 (95% at threshold) refer to different fit-quality criteria, since at threshold converged fits are accepted regardless of fit probability. This is explained, but the two numbers are far enough apart that a one-sentence reminder at the point where the 95% figure is used would prevent confusion.","section":"§4.5 and §6","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"It should be stated explicitly whether y_sig and y_bgd in Eq. (7) are event yields or normalizations, and whether the tail fraction f and the tail Gaussian parameters are fixed from the signal-only sample or floated in the final fit. The text implies they are fixed, but the reader should not have to infer this.","section":"Eqs. (5)–(7)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The background-mismodelling study quotes an 18 MeV shift in the top mass for a 5% variation in the subtracted background. It should be stated whether this is to be interpreted as a systematic uncertainty (i.e. 5% is the expected precision on the background) or as a sensitivity coefficient; the two readings imply different entries in the error budget.","section":"§6.3.2","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The expected number of selected signal and background events per scan point is not given. Providing these numbers, even in a small table, would let the reader judge the Poisson behaviour of Fig. 5 directly and would make the statistical precision of Table 3 easier to verify.","section":"§6.2","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The figure shows normalized distributions for the fully-hadronic and semi-leptonic W candidates, but the vertical axis label and the legend do not make clear that the semi-leptonic entry includes the neutrino-related smearing discussed in §4.4; a short addition to the caption would help.","section":"Fig. 1","authors_acknowledge":null,"authors_acknowledge_p":null}],"recommendation":"major_revision","confidential_remarks":"The paper is a solid and honest feasibility study, and the direct-reconstruction part is essentially complete for its genre. My concern is concentrated on the threshold-scan number, which is the one the abstract leads with and the one most likely to be quoted. The flat-efficiency assumption of §6.2 is not mentioned anywhere as a caveat, and the luminosity-spectrum term of §6.3.2 is acknowledged but not carried into the §7 conclusion. Both are fixable with modest additional simulation and an explicit budget table, and I would expect the revised paper to be publishable. I would also encourage the authors to keep the ILC comparison, which is a useful service to the community and is currently buried at the end of §6."},"author_rebuttal":null,"desk_editor":{"model":"selfhosted/deepseek-v4.1-flash","letter":"Two things to know. The 500 GeV direct-reconstruction result (80 MeV statistical, Table 2) is properly supported and honestly framed. The threshold number (34 MeV, Table 3) is a sensitivity estimate from a toy in which the pseudo-data are generated with the same efficiency and background the fit templates use, so it recovers the input mass by construction and reflects Poisson fluctuations only. That is fine as far as it goes, but it is not a test of the model, and the abstract's 'total uncertainties of around 100 MeV' reads as if it were. The one modeling input the authors do not flag is the flat efficiency: 70.2% signal and 73 fb residual background, both measured at 352 GeV, applied unchanged across 344 to 353 GeV (Section 6.2). Near threshold the tops are nearly at rest, jets are soft and close together, and b-tagging and kinematic-fit success will move with energy. A few percent variation distorts the turn-on and shifts the fitted mass and alpha_s. Worth asking for a coarse efficiency-versus-energy scan, or at least a statement of how much variation the fit tolerates. The luminosity-spectrum term (75 MeV for a 20% RMS change, Section 6.3.2) is larger and is acknowledged, so the 'below 100 MeV' total is conditional on two inputs, one admitted and one not. Note also that the classifier found no stated limitations in the abstract; the body is considerably more candid than that, and the refusal to call the invariant-mass value a theory-usable mass is exactly right. What is genuinely new is the execution: full GEANT4 CLIC_ILD with overlaid photon-photon hadron background at the CLIC rate, particle flow, kinematic fit, flavour tagging, likelihood rejection. And the CLIC-versus-ILC luminosity-spectrum comparison (Tables 3 and 4) is a real result, handled well: the 20% statistical edge at ILC is shown to be irrelevant once systematics dominate. That is the useful message for the design community. Send it out. The physics message, that a linear collider reaches the 100 MeV scale and that systematics set the floor, survives; the specific threshold number needs the efficiency assumption tested before anyone uses it as a design target.","headline":"A careful full-simulation CLIC study whose above-threshold result holds up; the 34 MeV threshold number is a toy closure resting on a flat efficiency measured at one energy, and that gap is the one the authors don't flag.","tokens_in":20031,"tokens_out":1330,"would_cite":true,"duration_ms":15301,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"pith_extraction":{"msc":[],"pacs":["14.65.Ha","29.20.Ej"],"model":"selfhosted/deepseek-v4.1-flash","headline":"A simulated 500 GeV CLIC stage could pin the top quark mass to roughly 100 MeV, and the threshold scan does it in a mass scheme that theory can actually interpret.","keywords":["top quark mass","CLIC","threshold scan","1S mass scheme","kinematic fitting","b-tagging","linear collider","strong coupling constant"],"falsifier":"Run the full detector simulation and event selection at each of the ten scan energies rather than only at 352 GeV: if the selection efficiency or the residual background varies by more than a few percent across the 10 GeV range, the fitted cross-section shape distorts and the quoted 34 MeV statistical mass error and the extracted $alpha_{s}$ both shift.","tokens_in":20361,"feed_emoji":"⚛️","tokens_out":2604,"duration_ms":38662,"temperature":0.7,"pith_summary":"The paper asks how precisely the top quark mass could be measured at a first-stage linear electron-positron collider built on CLIC technology, running at 500 GeV and around the top-pair threshold near 350 GeV, with 100 inverse femtobarns of data. Two routes are simulated in full: reconstructing the invariant mass of the top decay products above threshold, and scanning the rise of the top-pair cross section across the threshold. The direct reconstruction reaches a combined statistical precision of 80 MeV, but that number is a parameter of the PYTHIA event generator, and the authors say plainly that the theoretical bridge from a generator mass to a well-defined mass scheme is not yet built to match it. The threshold scan reaches 34 MeV statistical on the mass, simultaneously with the strong coupling at 0.0009, and because the cross-section shape is computed in the 1S mass scheme with next-to-next-to-leading-order QCD, the result lands in a scheme that can be translated into other conventions. The paper's real message is thus a comparison of two kinds of precision: experimental precision that theory cannot yet absorb, and a slower measurement whose precision theory can absorb.","feed_headline":"Top mass to 100 MeV at a 500 GeV linear collider","feed_subtitle":"Simulated CLIC data show a threshold scan pins the top mass in a well-defined scheme, not a generator number.","key_machinery":"Two pieces of machinery carry the argument. The first is the kinematic fit: it enforces three-dimensional energy and momentum conservation including the beam crossing angle, fixes both $W$ masses to 80.4 GeV, and forces the top and anti-top masses to be equal, varying jet and lepton four-momenta within their measured resolutions; this both sharpens the mass distribution and rejects more than 97 percent of the $qbar{q}$ background. The second is the threshold template fit: the top-pair cross section near threshold is computed with the TOPPIK code at next-to-next-to-leading order in the 1S mass scheme, folded with initial-state radiation and with the machine's luminosity spectrum, and the simulated data points are compared against a grid of templates in mass and $alpha_{s}$ spaced by 50 MeV and 0.0007 respectively. The luminosity spectrum is the pivot of the comparison with ILC: beamstrahlung smears the threshold rise, so the shape of that spectrum is itself a systematic input to the mass.","core_discovery":"The central claim is that a 500 GeV CLIC stage with 100 inverse femtobarns delivers a top quark mass with a total uncertainty around or below 100 MeV by two independent routes, and that the threshold route is the one with a defensible theoretical interpretation. Above threshold, a kinematic fit imposing energy-momentum conservation, the nominal $W$ mass of 80.4 GeV, and equal masses for top and anti-top, followed by a binned-likelihood background rejection, yields clean fully-hadronic and semi-leptonic samples; an unbinned fit of a Breit-Wigner convolved with a three-Gaussian resolution function and a background tail gives $m_{t} = 174.133$ GeV with a combined statistical error of 80 MeV and a width of $1.55$ GeV with an error of 220 MeV. At threshold, ten energy points spaced by 1 GeV from 344 to 353 GeV, 10 inverse femtobarns each, are fitted against NNLO cross-section templates in the 1S scheme, giving a statistical mass error of 34 MeV and a strong coupling error of 0.0009, with a theory normalization uncertainty worth only 5 to 8 MeV on the mass. Repeating the threshold analysis with the ILC luminosity spectrum, which has a sharper main peak, tightens the statistical error to 27 MeV, a difference far below the systematic uncertainties.","pith_inferences":["The paper's two numbers point in opposite directions for strategy: the 80 MeV invariant-mass result is cheaper in luminosity but currently unusable for precision physics, while the 34 MeV threshold result is only meaningful if the machine's luminosity spectrum is known to better than roughly 20 percent in RMS width, so accelerator diagnostics become a physics deliverable.","If the flat-efficiency assumption across the scan range fails, the distortion enters as a shape change in the cross-section rise, which is degenerate with a mass shift; a single additional simulated point at 344 GeV would test this cheaply.","The 350 MeV shift from a 2 percent b-jet energy scale error, versus under 100 MeV from light jets, suggests the b-jet calibration programme (using $Z \\to b\\bar{b}$ in di-boson and radiative-return events) is the binding experimental systematic for the above-threshold measurement.","A future comparison of a threshold-scan 1S mass with a hadron-collider Monte Carlo mass would test the size of the non-perturbative correction directly, rather than inferring it from generator comparisons."],"forward_implications":["A threshold scan at a linear collider would deliver the top mass in the 1S scheme, which can be converted into other mass conventions, unlike the generator mass extracted from invariant-mass reconstruction.","The same scan determines the strong coupling to about 0.0009, comparable to the current world-average uncertainty, so $alpha_{s}$ need not be taken as an external input.","Because the direct-reconstruction and threshold masses would be measured at the same machine, their difference becomes a handle on non-perturbative and colour-reconnection effects in the invariant-mass interpretation.","The comparison of CLIC and ILC luminosity spectra shows the sharper ILC peak buys only about 20 percent in statistical precision, so the choice between technologies is not decided by the threshold-scan mass error.","A ten-point scan of 100 inverse femtobarns total would already be limited by systematics rather than statistics, with the luminosity-spectrum shape the largest single worry.","A six-point scan from 344 to 349 GeV, 60 inverse femtobarns, degrades the mass error to 40 MeV and the coupling error to 0.0013, showing the region above threshold matters mainly for $alpha_{s}$."],"supporting_citations":[{"why":"Supplies the 1S top mass scheme and the TOPPIK threshold cross-section calculation on which the whole threshold-scan result rests.","marker":"[8]"},{"why":"Provides the next-to-next-to-leading-order threshold cross-section calculation used to build the fit templates.","marker":"[38]"},{"why":"The earlier TESLA threshold-scan study whose two-part strategy (efficiency from simulation, cross section from theory) this paper follows.","marker":"[37]"},{"why":"The kinematic fitting method that supplies the constrained event reconstruction and much of the background rejection.","marker":"[23]"},{"why":"The flavour-tagging package that identifies b-jets and enables the jet-to-top assignment.","marker":"[22]"},{"why":"The particle-flow reconstruction and beam-induced background suppression cuts applied to the simulated events.","marker":"[18]"},{"why":"The CLIC detector concept and beam-induced background model, including the $e^+e^-$ pair and two-photon overlay assumptions.","marker":"[12]"},{"why":"The YFS treatment of initial-state radiation folded into the threshold cross section.","marker":"[39]"},{"why":"The source of the estimated theory normalization uncertainty on the threshold cross section, quoted at 1 and 3 percent.","marker":"[40]"}],"fun_headline_variants":["Threshold scan pins top mass near 100 MeV at CLIC","CLIC simulation gives 100 MeV top mass by two routes","Direct top mass at CLIC reaches 100 MeV, theory lags","ILC peak cuts top mass statistical error to 27 MeV","Threshold scan yields top mass in a defined scheme at CLIC"],"cache_read_input_tokens":0,"weakest_assumption_plain":"The threshold-scan precision rests on the assumption that the signal selection efficiency of 70.2 percent and the residual background cross section of 73 femtobarns, both measured from fully simulated events at a single energy of 352 GeV, stay constant across the entire 344 to 353 GeV scan range, even though near threshold the event kinematics and the kinematic-fit success rate change quickly with energy.","fun_headline_variants_meta":{"raw":{"variants":["Threshold scan pins top mass near 100 MeV at CLIC","CLIC simulation gives 100 MeV top mass by two routes","Direct top mass at CLIC reaches 100 MeV, theory lags","ILC peak cuts top mass statistical error to 27 MeV","Threshold scan yields top mass in a defined scheme at CLIC"]},"model":"selfhosted/deepseek-v4.1-flash","effort":"low","attempts":1,"cost_usd":0.0,"raw_usage":{"total_tokens":1881,"prompt_tokens":982,"completion_tokens":899,"prompt_tokens_details":null,"completion_tokens_details":{"reasoning_tokens":824}},"tokens_in":982,"tokens_out":899,"duration_ms":16117,"image_pages":[],"temperature":1.0,"reasoning_budget":50,"reasoning_tokens":824,"transport_provider":"selfhosted","cache_read_input_tokens":0,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-11-10T12:28:12.706515+00:00","model_set":{"reader":"selfhosted/deepseek-v4.1-flash"},"falsifier":"Run the full detector simulation and event selection at each of the ten scan energies rather than only at 352 GeV: if the selection efficiency or the residual background varies by more than a few percent across the 10 GeV range, the fitted cross-section shape distorts and the quoted 34 MeV statistical mass error and the extracted $alpha_{s}$ both shift.","supporting_citations":[{"cited_title":"Phys. Rev. D60, 114027","cited_arxiv_id":null,"evidence_quote":"Supplies the 1S top mass scheme and the TOPPIK threshold cross-section calculation on which the whole threshold-scan result rests."},{"cited_title":"Phys. Rev. D58, 114023","cited_arxiv_id":null,"evidence_quote":"Provides the next-to-next-to-leading-order threshold cross-section calculation used to build the fit templates."},{"cited_title":"Eur. Phys. J. C27, 49","cited_arxiv_id":null,"evidence_quote":"The earlier TESLA threshold-scan study whose two-part strategy (efficiency from simulation, cross section from theory) this paper follows."},{"cited_title":"Nucl. Instrum. Meth. A624, 184","cited_arxiv_id":null,"evidence_quote":"The kinematic fitting method that supplies the constrained event reconstruction and much of the background rejection."},{"cited_title":"et al., Nucl. Instrum. Meth. A610, 573","cited_arxiv_id":null,"evidence_quote":"The flavour-tagging package that identifies b-jets and enables the jet-to-top assignment."},{"cited_title":"Nucl. Instrum. Meth. A700, 153","cited_arxiv_id":null,"evidence_quote":"The particle-flow reconstruction and beam-induced background suppression cuts applied to the simulated events."},{"cited_title":"Physics and Detectors at CLIC: CLIC Conceptual Design Report","cited_arxiv_id":"1202.5940","evidence_quote":"The CLIC detector concept and beam-induced background model, including the $e^+e^-$ pair and two-photon overlay assumptions."},{"cited_title":"Z. Phys. C49, 577","cited_arxiv_id":null,"evidence_quote":"The YFS treatment of initial-state radiation folded into the threshold cross section."},{"cited_title":"NNLL top-antitop production at threshold","cited_arxiv_id":"1111.4486","evidence_quote":"The source of the estimated theory normalization uncertainty on the threshold cross section, quoted at 1 and 3 percent."}],"review_version":1}