{"id":"a38f6f7f-9399-4536-afec-84d1fd60d314","arxiv_id":"1606.04833","paper_version":2,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"low","formal_verification":"none","parameter_count":5,"one_line_summary":"With 3.2 fb^-1 of 13 TeV data, ATLAS finds no diboson resonance and excludes a scalar singlet below 2650 GeV, a heavy vector triplet below 2600 GeV, and a bulk Randall-Sundrum graviton below 1100 GeV at 95% CL.","lead":"ATLAS searched 3.2 inverse femtobarns of 13 TeV proton collisions for heavy particles decaying into pairs of W and Z bosons, combining four decay channels. No excess appeared, and the resulting mass limits (up to 2.65 TeV for a scalar, 2.6 TeV for a vector triplet, 1.1 TeV for a graviton) are the strongest of their kind at the time and set the benchmark that later Run 2 searches had to beat.","discovery_kind":"unclear","skeptic_critique":{"model":"selfhosted/deepseek-v4.1-flash","headline":"The HVT model-B exclusion at 2600 GeV rests on rescaling model-A (Drell-Yan) samples; model B (gV=3) suppresses fermion couplings and enhances VBF, yet no VBF sample or acceptance comparison is shown and no systematic is assigned.","rationale":"The paper is a careful, conventional ATLAS resonance search, and its qualitative conclusion — no significant deviations from the Standard Model, with 95% CL limits set on three benchmark scenarios — is well supported. The statistical machinery (profile-likelihood-ratio test statistic, asymptotic CLs validated against pseudo-experiments, control regions folded into the combined fit) is standard and correctly applied, and Table 5 shows data and prediction agreeing in every control and signal region. The reader's weakest assumption, the four-parameter multi-jet background form of Eq. (1) in the qqqq channel, is a real concern, but the paper validates it with simulated dijet events and shifted-mass sidebands and reports that the modelling uncertainty is negligible compared to the statistical uncertainty in the fit parameters. I would not overturn the verdict on that basis. The concern I raise is narrower and more specific: the model-B HVT limit at 2600 GeV is obtained by rescaling model-A Drell-Yan samples, and the paper asserts without demonstration that the kinematic distributions are similar. Because model B suppresses the fermion coupling and enhances the VBF coupling, the production mechanism is expected to shift toward VBF at high mass, and the acceptance for VBF events is not shown to match the DY acceptance. The paper assigns no systematic to this rescaling. This does not invalidate the qualitative claim that limits are set, but it means the specific 2600 GeV number carries an unquantified approximation that the other two headline numbers do not share to the same degree. The reader noted this rescaling in the rationale but dismissed it as not load-bearing; I think it is the most specific unmanaged approximation affecting a headline number, which is why I mark agreement as partial rather than full. The verdict should remain ACCEPT because the central claim survives, but the 2600 GeV exclusion should be read with the caveat that the VBF acceptance has not been demonstrated to match the DY acceptance used to derive it.","tokens_in":52465,"tokens_out":14279,"duration_ms":312130,"concrete_test":"Generate HVT model-B signal events including the VBF production channel (qq -> qq W'/Z') with MadGraph5_aMC@NLO, pass them through the full event selection of Section 5, and compare the acceptance times efficiency to that of the rescaled model-A DY samples at 2.0 and 2.6 TeV. If the VBF acceptance differs from the DY acceptance by more than the quoted systematic uncertainties (or by more than about 20%), the 2600 GeV model-B exclusion is not established at the stated precision and the number should be requoted with that spread.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 3 states that the model-B (gV=3) HVT interpretation 'uses the model A signal samples rescaled to the predicted cross-sections from model-B,' justified by the assertion that 'the kinematic distributions relevant to this search are very similar.' That assertion is not demonstrated. In the HVT parameterisation the Drell-Yan coupling scales as g^2 cF/gV and the VBF coupling as gV cH; going from model A (gV=1) to model B (gV=3) suppresses DY by roughly a factor of nine and enhances VBF by roughly a factor of nine, so model B is expected to be VBF-dominated at high mass. VBF events carry two forward tagging quarks and a resonance with a different rapidity and pT distribution from DY events. The analysis selections that shape the acceptance — pT,J/m > 0.4, pT,lv/m > 0.4, |yJ1 - yJ2| < 1.2 in qqqq, and the lepton and missing-energy requirements in the semileptonic channels — can respond differently to that topology. The paper assigns no systematic uncertainty to the rescaling and shows no VBF sample. If the VBF acceptance differs from the DY acceptance by more than the quoted uncertainties, the 2600 GeV number — one of the three headline exclusions — is not established at the stated precision. This is a more specific and less-managed gap than the qqqq background-shape concern the reader flags, which the paper at least validates with sidebands and simulation.","agreement_with_reader":"partial"},"referee_report":{"model":"selfhosted/deepseek-v4.1-flash","summary":"The paper reports a combined search for heavy narrow-width resonances decaying to WW, WZ, or ZZ in the ννqq, ℓνqq, ℓℓqq, and qqqq final states, using 3.2 fb⁻¹ of pp collisions at √s = 13 TeV recorded by ATLAS in 2015. The discriminating variables are invariant or transverse masses, and the four channels are combined in a profile-likelihood fit with data-driven background estimates and control regions. No significant deviations from the Standard Model are observed. Limits at 95% CL are set on three benchmark scenarios: a CP-even scalar singlet (Unsuppressed scenario, excluded below 2650 GeV), a heavy vector triplet (HVT model B with gV = 3, excluded below 2600 GeV), and a bulk Randall–Sundrum graviton (κ/M̄Pl = 1, excluded below 1100 GeV). The results extend the reach of the 8 TeV searches.","tokens_in":52839,"tokens_out":13724,"duration_ms":306941,"significance":"The result is significant for the field. It is the first 13 TeV combination of the four diboson final states in ATLAS, and the quoted mass exclusions are the strongest contemporaneous limits on these three benchmark scenarios. The analysis is technically solid: the profile-likelihood-ratio test statistic (Eq. 3) and the CLs prescription are standard and correctly applied, the asymptotic approximation is validated against pseudo-experiments to better than 10%, the dominant backgrounds are constrained with dedicated control regions folded into the combined fit, and the qqqq background parameterisation is validated with simulation and data sidebands. The paper is also transparent about its model assumptions, stating explicitly that off-shell and interference effects are not considered and that no combination with the 8 TeV results has been performed. These are acknowledged limitations, not new findings.","major_comments":[{"comment":"Section 3 states that the HVT model-B (gV = 3) interpretation 'uses the model A signal samples rescaled to the predicted cross-sections from model-B,' justified by the assertion that 'the kinematic distributions relevant to this search are very similar.' That assertion is not demonstrated. In the HVT parameterisation the Drell–Yan coupling scales as g²cF/gV while the VBF coupling scales as gVcH, so going from model A (gV = 1) to model B (gV = 3) suppresses Drell–Yan and enhances VBF, and model B is expected to have a larger VBF fraction at high mass. The selections that shape the acceptance — pT,J/m > 0.4, pT,ℓν/m > 0.4, |yJ1 − yJ2| < 1.2 in qqqq, and the lepton and missing-energy requirements in the semileptonic channels — can respond differently to the VBF topology, which carries two forward tagging quarks and a resonance with a different rapidity and pT distribution. No VBF sample comparison and no systematic uncertainty for the rescaling are shown. Because the 2600 GeV model-B exclusion is one of the three headline numbers, the paper should either show that the acceptance for a VBF-produced triplet is close to that for the Drell–Yan sample or assign a systematic uncertainty to the rescaling. The total uncertainty is dominated by the large-R jet scale (about 35%), so the effect may be subdominant, but it should be quantified rather than asserted.","section":"Section 3","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The qqqq background is modelled by the four-parameter function of Eq. (1), whose shape parameters p2 and p3 are fitted to the same mJJ spectrum used to set the limit. The validation described in Section 6 shifts the large-R jet mass window to 110–140 GeV or 50–65 GeV, so it tests the shape function in a different selection rather than in the signal region itself. The paper does not report a goodness-of-fit p-value for Eq. (1) in the signal region, nor an estimate of the bias in the fitted signal yield from the choice of functional form. Since the qqqq channel drives the high-mass reach of the scalar and graviton limits, the paper should report the goodness of fit and a quantitative estimate of the fit bias, for example from pseudo-experiments or from the validation regions. A smooth four-parameter form cannot absorb a narrow resonance, so the concern is broad mismodelling rather than signal-like bumps, but the check is cheap and would strengthen the result.","section":"Section 6, Eq. (1)","authors_acknowledge":null,"authors_acknowledge_p":null}],"minor_comments":[{"comment":"The abstract quotes three mass exclusions without naming the benchmark scenarios, whereas the conclusion specifies the Unsuppressed scalar, HVT model B with gV = 3, and κ/M̄Pl = 1. Please add the benchmark labels to the abstract so the three numbers are unambiguous.","section":"Abstract","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The W+jets row lists four entries (797, 54, 208, 397) for a table with seven columns, and the caption states that the W+jets background in the Z+jets control region is negligible while the row appears to contain a value in that column. Please check the table against the caption and supply or confirm the missing entries.","section":"Table 5","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The 2.5σ excess at 1.6 TeV is quoted as a local significance. Please give the global p-value, or state explicitly that the excess is not significant after accounting for the look-elsewhere effect over the scanned mass range.","section":"Section 9","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The sentence 'Points inside the ellipse, but where the absolute values of the cH/Λ and c3/Λ parameters are larger than at the exclusion contour, are considered to be excluded' is hard to parse; please state the exclusion direction more directly.","section":"Figure 5 caption","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The text gives c3 = (1/4π)² for the NDA scenario while the Table 1 caption writes c3 = 1/16π²; these are equal, but the notation should be harmonised.","section":"Section 3 and Table 1","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The cuts pT,J/m > 0.4 and pT,ℓν/m > 0.4 use m without definition in that sentence; please state that m is the corresponding diboson invariant mass.","section":"Section 5","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The paper states that PDF-induced uncertainties are not evaluated for the qqqq channel because they are subdominant; please quantify the expected size or cite the study that establishes the subdominance.","section":"Section 7","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The abstract's 'significantly extend the previous limits' compares limits obtained for different benchmark models (EGM W′ at 8 TeV versus HVT model B at 13 TeV); the text handles this carefully in Section 9, but the abstract could be more precise.","section":"Abstract and Section 9","authors_acknowledge":null,"authors_acknowledge_p":null}],"recommendation":"minor_revision","confidential_remarks":"The paper is a strong candidate for the journal. The two major points are requests for quantification of assumptions the paper already states, not indications that the central result is wrong: the HVT model-B rescaling is a known approximation in the HVT literature, and the qqqq background fit is standard practice in dijet resonance searches. The requested checks are local and should be straightforward to supply."},"author_rebuttal":null,"desk_editor":{"model":"selfhosted/deepseek-v4.1-flash","letter":"This is the first 13 TeV ATLAS diboson resonance search, and it does what it says: four channels combined, 3.2 fb^-1, exclusions at 2650 GeV (scalar singlet, Unsuppressed), 2600 GeV (HVT model B), and 1100 GeV (bulk graviton). Those numbers are new and they roughly double the 8 TeV reach. The machinery is inherited from the Run 1 combination and the substructure programme, and it is applied carefully: profile-likelihood CLs, asymptotic validation to 10%, control regions folded into the fit, Table 5 showing data and prediction agreeing everywhere. The 2.5 sigma local excess at 1.6 TeV is labelled local, which is the right call. The paper also states plainly that no 8 TeV combination was performed, which is honest bookkeeping rather than a hidden gap. The soft spots are the standard ones for this kind of analysis, and neither is load-bearing. The qqqq multi-jet background is fitted in situ with Eq. (1) and validated only through simulation and shifted-mass sidebands; that is the assumption the high-mass scalar and graviton limits lean on hardest, but it is well-managed and the sidebands are sensible. The stress test flags the HVT model-B rescaling: model A samples rescaled to model-B cross-sections, with the paper asserting the kinematic distributions are 'very similar' and assigning no systematic. That is a real gap worth quantifying. But the specific mechanism the stress test proposes, that model B is VBF-dominated because the DY coupling scales as 1/gV and the VBF coupling as gV, is not clearly right for the standard HVT implementation, where the triplet is generated via Drell-Yan in both benchmarks and VBF is a small correction. The paper would be stronger with an explicit acceptance comparison or a systematic on the rescaling, but I would not call the 2600 GeV number unestablished on this basis. Who should read it: anyone tracking diboson resonance limits, and anyone building the next round of 13 TeV searches who wants the channel definitions and background strategy. It is a reference paper, not a methods paper. Recommendation: send it to referees. It is a careful, honest search and the exclusions will be cited.","headline":"A solid, standard ATLAS diboson resonance search on the first 3.2 fb^-1 at 13 TeV; the exclusions are real and the soft spots are the usual ones, neither load-bearing.","tokens_in":53571,"tokens_out":3378,"would_cite":true,"duration_ms":78821,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"pith_extraction":{"msc":[],"pacs":["12.60.-i","13.85.Rm","14.70.-e"],"model":"selfhosted/deepseek-v4.1-flash","headline":"ATLAS finds no diboson resonances in the first 13 TeV data, and converts the null result into mass exclusions reaching 2.65 TeV for a scalar singlet, 2.6 TeV for a heavy vector triplet, and 1.1 TeV for a warped graviton.","keywords":["diboson resonances","ATLAS","LHC 13 TeV","jet substructure","boosted W/Z bosons","heavy vector triplet","Randall-Sundrum graviton","CLs limits"],"falsifier":"A bump in the $m_{JJ}$ distribution between 1 and 3.5 TeV that survives the boson-tagging and track-multiplicity requirements, or a data-driven demonstration that the multi-jet background has structure inside that window which the 110–140 GeV and 50–65 GeV sidebands do not capture, would break the quoted $qqqq$ limits and with them the combined scalar and graviton exclusions.","tokens_in":53723,"feed_emoji":"⚛️","tokens_out":3102,"duration_ms":62735,"temperature":0.7,"pith_summary":"The paper asks whether the 13 TeV proton–proton collisions recorded by ATLAS in 2015 contain a heavy particle that decays into a pair of W or Z bosons. It answers by combining four mutually exclusive event selections — $qqqq$, $qqqq$ with missing energy ($qqqq$ replaced by $qqqq$; specifically $qqqq$, $qqqq$, $qqqq$, $qqqq$ meaning fully hadronic, missing-transverse-momentum, one-lepton and two-lepton plus hadronic-jet final states) — each of which reconstructs at least one highly boosted W or Z as a single large-radius jet and identifies it through its mass and two-prong substructure. No bump appears above the smoothly falling Standard Model backgrounds in any channel, so the paper turns the absence into 95% confidence-level upper limits on production cross-section times branching ratio as a function of resonance mass. Read against three named benchmark models, those limits translate into concrete exclusions: a CP-even scalar singlet below 2650 GeV in the Unsuppressed scenario, a heavy vector-boson triplet in model-B with $g_V = 3$ below 2600 GeV, and a bulk Randall–Sundrum graviton with $c = 1$ below 1100 GeV. The value to a reader is that these are the reach statements that matter for model builders: the 8 TeV run excluded the corresponding EGM $W'$ only below 1.81 TeV and the bulk graviton only below 810 GeV, so the same logic applied to 3.2 fb$^{-1}$ at higher energy roughly doubles the graviton reach and pushes the vector-triplet reach past 2.5 TeV.","feed_headline":"No diboson bumps: new particles pushed above 2.6 TeV","feed_subtitle":"ATLAS's first 13 TeV data rule out a heavy vector triplet below 2600 GeV and a warped graviton below 1100 GeV.","key_machinery":"Three pieces of machinery carry the argument. First, boosted-boson tagging: hadronically decaying W and Z bosons are reconstructed as trimmed large-radius jets with $p_T > 200$ GeV, selected by a mass window of $pm 15$ GeV around the measured W (83.2 GeV) or Z (93.4 GeV) peak together with a $p_T$-dependent cut on the two-prong observable $D_2^{beta=1}$, tuned to about 50% efficiency on longitudinally polarised bosons while rejecting more than 90% of quark and gluon jets. Second, channel-specific background estimation: the semileptonic channels use simulated W+jets, Z+jets, $tbar{t}$ and diboson shapes with normalisations constrained by dedicated control regions inside the same fit, while the fully hadronic channel, dominated by multi-jet events, models the $m_{JJ}$ spectrum with the four-parameter form $f(x) = N(1-x)^{p_2+xi p_3} x^{p_3}$ with $x = m_{JJ}/sqrt{s}$, validated on simulated dijet samples and on data sidebands with the boson mass window shifted to 110–140 GeV or 50–65 GeV. Third, the statistical combination: a product of Poisson likelihoods across channels and discriminant bins, with nuisance parameters for jet energy and mass scale and resolution, background modelling, luminosity and signal acceptance, turned into limits through the asymptotic distributions of the profile-likelihood-ratio test statistic.","core_discovery":"The central claim is a set of exclusions, not a detection. Combining the $qqqq$, $qqqq$, $qqqq$ and $qqqq$ channels in a profile-likelihood fit with the modified-frequentist $CL_s$ prescription, ATLAS observes no significant deviation from the Standard Model; the largest local excess is 2.5 standard deviations, in the scalar selection near 1.6 TeV, and it is not claimed as evidence. The quoted 95% CL lower mass limits are 2650 GeV for the scalar singlet in the Unsuppressed benchmark, 2600 GeV for the HVT model-B triplet with $g_V = 3$, and 1100 GeV for the bulk Randall–Sundrum graviton with $c = 1$. The paper also reports model-independent limits on cross-section times branching ratio into $WW$ and $ZZ$ as a function of mass, and recasts them as exclusion contours in the scalar coupling plane $(c_H/$, $c_3/$) and in the HVT coupling plane $(g^2 c_F/g_V, g_V c_H)$, so that a reader with a different parameter choice can read off whether their point survives.","pith_inferences":["The exclusion pattern is driven less by statistics than by kinematics: because a 2.6 TeV resonance produces W and Z bosons with transverse momenta of order a TeV, the boosted-jet technique is what converts a modest 3.2 fb$^{-1}$ into TeV-scale reach, so further gains should come from tagger efficiency at the highest $p_T$ rather than from luminosity alone.","The three benchmark exclusions are not equally robust: the scalar and graviton limits lean on the fully hadronic channel at high mass, where the background is a fitted functional form rather than a simulated or data-driven template, whereas the vector-triplet limit draws on all four channels. A reader comparing the three numbers should weight them differently.","Because the limits are quoted for narrow resonances, a broad or strongly interfering resonance of the same mass could hide inside these bounds; the coupling-plane contours already hint at where the narrow-width assumption starts to bite through the 5% width ellipse.","The absence of any excess in the $qqqq$ channel at 13 TeV, where the 8 TeV combination had a 2.5 standard deviation bump near 2 TeV, is itself informative: it suggests that earlier feature was a fluctuation rather than a signal."],"forward_implications":["Any model predicting a colour-singlet resonance that decays to $WW$, $WZ$ or $ZZ$ with a narrow width and a mass below roughly 2.6 TeV now has to survive these limits or explain why its production or branching ratio falls below them.","The bulk Randall–Sundrum graviton with $c = 1$ is pushed above 1.1 TeV, roughly a factor 1.4 beyond the 8 TeV exclusion of 810 GeV, which tightens the curvature-scale window that warped-extra-dimension model builders can still occupy.","The published cross-section limits as a function of mass and the coupling-plane contours let anyone re-derive an exclusion for a benchmark the paper did not tabulate, without redoing the analysis.","The 2.5 standard deviation local excess near 1.6 TeV in the scalar selection becomes a concrete target: more luminosity at 13 TeV either confirms a bump there or removes it.","The demonstrated combination of fully hadronic and semileptonic diboson channels at 13 TeV sets the template that later, larger datasets extend simply by rescaling the same selections."],"supporting_citations":[{"why":"Supplies the CP-even scalar singlet model whose Unsuppressed benchmark yields the 2650 GeV exclusion.","marker":"[16]"},{"why":"Defines the heavy vector triplet parameterisation and its model-A and model-B benchmarks against which the 2600 GeV limit is quoted.","marker":"[17]"},{"why":"Provides the bulk Randall–Sundrum graviton model with SM fields in the extra dimension, the basis for the 1100 GeV exclusion.","marker":"[19]"},{"why":"The 8 TeV ATLAS diboson combination whose limits (1.81 TeV EGM $W'$, 810 GeV bulk graviton) are the baseline this analysis extends.","marker":"[31]"},{"why":"Defines the $D_2^{beta=1}$ two-prong substructure observable used by the boson tagger that makes the boosted W/Z reconstruction possible.","marker":"[82]"},{"why":"Provides the 13 TeV identification and calibration of boosted hadronically decaying W and Z bosons used throughout the event selection.","marker":"[14]"},{"why":"Supplies the large-radius jet energy and mass scale and resolution corrections, the dominant systematic uncertainty in the fit.","marker":"[80]"},{"why":"Gives the asymptotic formulae for the profile-likelihood-ratio test statistic used to compute p-values and limits.","marker":"[95]"},{"why":"Defines the $CL_s$ prescription applied to set the 95% confidence-level upper limits on cross-section times branching ratio.","marker":"[96]"}],"fun_headline_variants":["ATLAS diboson search finds no significant deviation from the Standard Model","Excluded below 2650 GeV: scalar singlet in ATLAS 13 TeV diboson search","Heavy vector triplet excluded below 2600 GeV; graviton below 1100 GeV","Largest local excess is 2.5 sigma near 1.6 TeV, not claimed as evidence","Review flag: channel list repeats qqqq four times; exclusions unchanged"],"cache_read_input_tokens":0,"weakest_assumption_plain":"The high-mass reach rests on the fully hadronic channel, and that channel assumes the entire multi-jet background follows one smooth four-parameter curve fitted to the very mass spectrum used to set the limit, with only simulated dijet events and shifted-mass sidebands as validation.","fun_headline_variants_meta":{"raw":{"variants":["ATLAS diboson search finds no significant deviation from the Standard Model","Excluded below 2650 GeV: scalar singlet in ATLAS 13 TeV diboson search","Heavy vector triplet excluded below 2600 GeV; graviton below 1100 GeV","Largest local excess is 2.5 sigma near 1.6 TeV, not claimed as evidence","Review flag: channel list repeats qqqq four times; exclusions unchanged"]},"model":"selfhosted/deepseek-v4.1-flash","effort":"low","attempts":1,"cost_usd":0.0,"raw_usage":{"total_tokens":1980,"prompt_tokens":1066,"completion_tokens":914,"prompt_tokens_details":null,"completion_tokens_details":{"reasoning_tokens":810}},"tokens_in":1066,"tokens_out":914,"duration_ms":15973,"image_pages":[],"temperature":1.0,"reasoning_budget":50,"reasoning_tokens":810,"transport_provider":"selfhosted","cache_read_input_tokens":0,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-11-09T14:21:17.578522+00:00","model_set":{"reader":"selfhosted/deepseek-v4.1-flash"},"falsifier":"A bump in the $m_{JJ}$ distribution between 1 and 3.5 TeV that survives the boson-tagging and track-multiplicity requirements, or a data-driven demonstration that the multi-jet background has structure inside that window which the 110–140 GeV and 50–65 GeV sidebands do not capture, would break the quoted $qqqq$ limits and with them the combined scalar and graviton exclusions.","supporting_citations":[{"cited_title":"What is the gamma gamma resonance at 750 GeV?","cited_arxiv_id":"1512.04933","evidence_quote":"Supplies the CP-even scalar singlet model whose Unsuppressed benchmark yields the 2650 GeV exclusion."},{"cited_title":"Heavy Vector Triplets: Bridging Theory and Data","cited_arxiv_id":"1402.4431","evidence_quote":"Defines the heavy vector triplet parameterisation and its model-A and model-B benchmarks against which the 2600 GeV limit is quoted."},{"cited_title":"Combination of searches for $WW$, $WZ$, and $ZZ$ resonances in $pp$ collisions at $\\sqrt{s} = 8$ TeV with the ATLAS detector","cited_arxiv_id":"1512.05099","evidence_quote":"The 8 TeV ATLAS diboson combination whose limits (1.81 TeV EGM $W'$, 810 GeV bulk graviton) are the baseline this analysis extends."},{"cited_title":"Power Counting to Better Jet Observables","cited_arxiv_id":"1409.6298","evidence_quote":"Defines the $D_2^{beta=1}$ two-prong substructure observable used by the boson tagger that makes the boosted W/Z reconstruction possible."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the 13 TeV identification and calibration of boosted hadronically decaying W and Z bosons used throughout the event selection."},{"cited_title":"Performance of jet substructure techniques for large-R jets in proton-proton collisions at sqrt(s) = 7 TeV using the ATLAS detector","cited_arxiv_id":"1306.4945","evidence_quote":"Supplies the large-radius jet energy and mass scale and resolution corrections, the dominant systematic uncertainty in the fit."},{"cited_title":"Asymptotic formulae for likelihood-based tests of new physics","cited_arxiv_id":"1007.1727","evidence_quote":"Gives the asymptotic formulae for the profile-likelihood-ratio test statistic used to compute p-values and limits."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the $CL_s$ prescription applied to set the 95% confidence-level upper limits on cross-section times branching ratio."}],"review_version":1}