{"id":"56fe4b3e-8398-4d42-b7fd-d926c66056ba","arxiv_id":"2511.16394","paper_version":4,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"Using low-momentum muon pairs and lepton-plus-track events, CMS finds no new physics and sets new exclusion limits on compressed Higgsinos, reaching 115 GeV at a 3.5 GeV splitting.","lead":"This CMS analysis searches for Higgsino particle pairs that decay into very soft muons or an electron plus a track, aided by large missing energy, in 137 inverse femtobarns of 13 TeV proton-proton data. No new-physics excess is found, and the new technique extends sensitivity into compressed-mass regions, excluding Higgsinos up to 115 GeV at 3.5 GeV splitting.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Jetty BDT-shape extrapolation in Eq. (1) is the fragile link: a high-BDT cTF mismatch could shift the 2.2σ excess and the 115 GeV limit; simulation closure may not cover it.","rationale":"I read the paper as a standard CMS search whose central claim is a null result with new exclusion limits in compressed-Higgsino parameter space. The statistical procedure, signal model, and systematic treatment are internally consistent, and the paper openly reports the 2.2σ local excess and the marginal role of the exclusive-track category. The weakest point is indeed the Jetty transfer-factor extrapolation: Eq. (1) assumes shape invariance of the BDT distribution between the isolation sideband and the main band, with cTF measured only in BDT<0. This is the same concern identified by the Reader. The paper mitigates it with simulation closure tests and a 8–22% shape systematic, but the highest-BDT bins are the ones that drive the observed excess and the limit. A concrete numerical test on the public HEPData yields would settle whether this assumption is numerically important. Because the concern is already partially covered by assigned systematics and does not make the paper internally invalid, I would not change the ACCEPT verdict, only note that the limit’s robustness in the most sensitive bin depends on an extrapolation that is not independently verifiable from the text alone.","tokens_in":38561,"tokens_out":10866,"duration_ms":105818,"concrete_test":"Using the public HEPData yields, recompute the 95% CLs limit after scaling the Jetty prediction in the highest dimuon BDT bin (0.5–1.0) by the quoted +22% shape systematic, and separately by shifting cTF_Phase1 by ±1σ. If the maximum excluded Higgsino mass changes by more than ~10 GeV or the local significance of the Phase-1 excess drops below 2σ, the Eq. (1) BDT-shape extrapolation is decisive; if the limit is stable, the concern does not land.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The dominant background in the dimuon category is the Jetty background, estimated with Eq. (1): a single transfer factor cTF measured in the BDT<0 region is applied to the isolation-sideband BDT shape to predict all BDT>0 signal-region bins. The central result—no significant excess and the 95% CL exclusion reaching 115 GeV—rests on the assumption that the BDT shape of the Jetty background is the same in the isolation sideband and the main band. The validation in Section 4.3 is a simulation closure test: a first-order polynomial fit to the simulated ratio versus BDT is consistent with unity, and the resulting shape systematic grows from 8% to 22% in the highest-BDT bins. However, the most sensitive dimuon SR, where the 2.2σ local excess appears, is precisely the bin where this systematic is largest and where the simulated sample is thinnest. A physical violation of the shape assumption—for example, a change in the heavy-flavor/light-parton composition of misidentified leptons with BDT, or a data/MC difference in low-pT nonprompt muon rates—would directly shift the background prediction in the SRs that drive the exclusion contour. The paper provides no data-driven closure of the BDT-shape extrapolation in the SR region, such as an alternative isolation definition or a same-charge dimuon control region. This is a standard limitation rather than an internal inconsistency, but it is the load-bearing assumption for both the null result and the headline limit.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper reports a CMS search for compressed-spectrum Higgsino pair production using 137 fb^-1 of 13 TeV pp collisions. It targets final states with large missing transverse momentum and either two low-momentum muons or a lepton plus an isolated track, using BDT discriminants to define signal regions. Backgrounds are estimated with data-driven methods: the dominant Jetty background is extrapolated from an isolation sideband using a transfer factor measured in a BDT<0 region (Eq. 1), the tau-tau background uses simulation corrected in a control region, and the lepton+track category uses a same-charge control region. No significant deviation from the SM background is observed; the largest local excess is 2.2 sigma in the most sensitive dimuon signal region of the Phase 1 sample. The analysis sets 95% CL exclusion limits in the Higgsino mass versus mass-splitting plane, probing neutralino mass differences down to 1.5 GeV for a Higgsino mass of 100 GeV and excluding Higgsino masses up to 115 GeV at a mass difference of 3.5 GeV.","tokens_in":38966,"tokens_out":6872,"duration_ms":66465,"significance":"If the result is correct, the analysis fills a previously unexplored gap in compressed-electroweakino searches between soft-lepton and disappearing-track signatures. The paper is careful and transparent: it reports the 2.2 sigma excess honestly, assigns systematic uncertainties that grow from 8% to 22% toward the high-BDT bins where the excess appears, provides simulation closure tests for the background methods, and makes tabulated results available on HEPData. The statistical framework is standard for LHC searches, and the central null result is stated with appropriate caution. The main weakness is the reliance on the Jetty BDT-shape extrapolation in Eq. (1), but the paper addresses this with a simulation closure test and a shape systematic; a same-charge dilepton control region is also mentioned, though not described in detail.","major_comments":[],"minor_comments":[{"comment":"The notation N^SR_sideband(x) is confusing: the superscript 'SR' appears to label the sideband, not the signal region. Please rename this to something like N_isoSideband(x) to make clear that the BDT template is taken from the isolation sideband and that the transfer factor is measured in the BDT<0 normalization region.","section":"Section 4.1, Eq. (1)"},{"comment":"The text states that the systematic uncertainty is assessed using discrepancies in a same-charge dilepton control region, but this validation is not described. If this CR is used to test the Jetty BDT-shape extrapolation at high BDT, a brief description and the observed agreement (or limit) would strengthen the paper. If it serves a different purpose, please state what it validates explicitly.","section":"Section 4.3"},{"comment":"The paper quotes a 2.2 sigma local significance but does not give a global significance or mention the trials factor. Please either provide the global significance or state explicitly that only the local significance is quoted.","section":"Section 5"},{"comment":"The phrase 'No selected muon or electron may lie within Delta R of 0.01 of the track' should read 'within Delta R < 0.01' to avoid ambiguity.","section":"Section 3"},{"comment":"The Phase 1 tau-tau transfer factor has a 90% relative uncertainty. The text notes the impact on sensitivity is negligible, but a short explanation of how such a large uncertainty is handled in the likelihood and why it does not affect the final limits would be helpful.","section":"Table 2 / Section 4.1"}],"recommendation":"minor_revision","confidential_remarks":"This is a solid and careful experimental paper. The only substantive concern is the Jetty BDT-shape extrapolation in Eq. (1), which is the most fragile part of the background model and sits exactly where the 2.2 sigma excess appears. The paper already contains a same-charge dilepton CR and a simulation closure test, but the CR is not described; adding a sentence or two about what it validates would fully address the concern. I do not see a need for another full round of review."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: this is a solid, honestly reported search that does what it says. It extends compressed-Higgsino sensitivity by roughly 0.5–2 GeV below the previous soft-lepton reach, using muons with pT down to 2 GeV and a lepton-plus-exclusive-track category. The analysis is clearly written, the background model is conventional, and the 2.2 sigma local excess is reported without spin. That alone merits a serious referee.\n\nWhat's genuinely new: first use of sub-3 GeV muons combined with a lepton-plus-track category, explicitly targeting the gap left by Ref. [33] (which loses sensitivity below about 3 GeV splitting). The paper is appropriately self-critical that the track category contributes only marginally to the final result. The exclusion limits and the HEPData record give the community something reusable.\n\nThe soft spot is exactly the one the stress test flags: Eq. (1) extrapolates the Jetty background from the isolation sideband into the signal region using a single transfer factor measured at BDT<0, with closure confirmed only in simulation. The shape systematic grows to 22% in the highest-BDT bins, and that is where the observed excess sits. This is a real limitation, but it is a standard one. The method is widely used, the systematic is treated as correlated, and the observed limit is weaker than expected, so a background misestimate in that direction would not produce an over-optimistic exclusion. A same-charge dimuon control region would be a useful additional cross-check, but its absence does not undermine the central null result. The tau-tau transfer factor has a 90% uncertainty in Phase 1 but negligible impact on sensitivity.\n\nThis paper is for SUSY phenomenologists and experimental colleagues tracking compressed-spectrum searches. It is incremental but solid. I see no fatal flaws. Send it to review; it will pass with minor revisions. The main thing I would ask the referee to examine is the transfer-factor closure and whether an alternative isolation definition can confirm the highest-BDT bins.","headline":"Competent, honest CMS search that extends compressed-Higgsino coverage by about 0.5–2 GeV; the single real weak link is the standard simulation-validated background extrapolation.","tokens_in":39422,"tokens_out":2488,"would_cite":true,"duration_ms":25454,"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":"The paper claims that a 137 fb−1 search for compressed-spectrum Higgsinos sees no beyond-standard-model excess, probing neutralino mass differences down to 1.5 GeV and excluding Higgsino masses up to 115 GeV.","keywords":["Higgsino","compressed supersymmetry","soft leptons","missing transverse momentum","boosted decision trees","electroweakino","dark matter candidate","LHC physics"],"falsifier":"A concrete check: using the full simulation, compute the sideband-to-main-band transfer factor separately in each BDT bin above 0.4 and see whether it drifts outside the assigned 8–22% systematic uncertainty. Alternatively, with the roughly 300 fb−1 already available from Run 3, if the tightest dimuon signal region shows an excess over the background prediction of more than 3 standard deviations, the conclusion that the 2.2σ deviation is a fluctuation would be wrong.","tokens_in":38448,"feed_emoji":"⚛️","tokens_out":7498,"duration_ms":74778,"temperature":0.7,"pith_summary":"Pair production of Higgsinos—supersymmetric fermions that are natural dark-matter candidates—is hardest to see when the two neutralino states are almost degenerate: the decay leptons are then extremely soft and usually thrown away by lepton triggers or isolation. This paper targets exactly that corner, using 137 fb−1 of 13 TeV proton-proton collisions from the CMS detector, selecting events with large missing transverse momentum plus either two low-momentum muons or one lepton paired with an isolated, unmatched track. It is the first search to use muons with transverse momentum down to 2 GeV together with a track-recovery category in this final state. The data match the standard-model background across the signal regions; the most notable deviation is a 2.2σ excess in the tightest dimuon bin. Interpreting the null result in a compressed Higgsino model, the paper excludes chargino and neutralino states up to a Higgsino mass of 115 GeV and reaches neutralino mass differences as low as 1.5 GeV at a 100 GeV Higgsino.","feed_headline":"No Higgsinos seen down to 1.5 GeV splitting","feed_subtitle":"A soft-lepton, lepton-plus-track search closes the compressed-spectrum gap where previous limits lost all sensitivity.","key_machinery":"The central object is the compressed Higgsino simplified model, with four nearly degenerate electroweakino states: two neutralinos separated by a mass difference of 1–10 GeV and a chargino separated from the lightest neutralino by half that amount, with the heavier neutralino decaying through a virtual Z to two same-flavor leptons. The analysis machinery consists of three channels—dimuon, muon+exclusive track, and electron+exclusive track—plus a custom jet-based isolation that preserves leptons down to 2 GeV, a track-picking boosted decision tree (BDT) that selects the track most likely to be the lost lepton, and an event-level BDT whose output defines the signal bins. The dominant jet-relat","core_discovery":"On the paper's own terms, the discovery is a null result that maps a previously blank region of the compressed-supersymmetry plane. The signal process is pair production of a heavier neutralino together with either a lighter neutralino or a chargino, followed by the decay of the heavier neutralino to the lighter one plus a virtual Z boson that decays to two same-flavor leptons. By lowering the muon transverse-momentum threshold to 2 GeV, allowing muon pairs with very small angular separation, and adding a lepton-plus-exclusive-track category to recover one lepton missed by identification, the analysis gains sensitivity to neutralino mass splittings down to 1.5 GeV. No significant deviation f","pith_inferences":["The transfer-factor extrapolation is the point I would stress in a follow-up: the background prediction in the signal bins rests on the assumption that the jet-related background's BDT shape is identical in the isolation sideband and the main band, so a direct measurement of the transfer factor in the highest BDT bins, using more simulation or an independent control region, would be the most natur","The result mildly tightens the case against very light natural supersymmetry: with Higgsino masses near 100 GeV excluded for splittings of a few GeV, the simplest fine-tuning-motivated corners must move to heavier masses, larger splittings, or different production assumptions.","The same low-momentum lepton-plus-track strategy should translate directly to the larger Run 3 dataset; with two to three times the integrated luminosity, a signal just below today's limits would either appear as a more than 3σ excess in the tightest bins or extend the exclusion by tens of GeV.","The published limits can be reinterpreted in other models with inelastic dark matter or near-degenerate electroweak doublets, since the final-state signature depends mainly on the mass spectrum and the virtual-Z decay, not on supersymmetry details."],"forward_implications":["If the null result is right, the search closes the sensitivity gap between soft-lepton and disappearing-track searches for compressed Higgsinos, reaching neutralino mass splittings of 1.5 GeV where previous lepton-based searches lost sensitivity below about 3 GeV.","Chargino masses up to about 115 GeV—and Higgsino masses near 100 GeV with splittings down to 1.5 GeV—are excluded at 95% confidence level, tightening the allowed parameter space for natural supersymmetry and electroweak-multiplet dark matter.","The lepton+exclusive-track category, which recovers up to 50% of leptons that would otherwise be missed, demonstrates a reusable strategy for other searches at the kinematic edge.","The observed 2.2σ excess in the tightest dimuon bin means that the highest-BDT bins deserve scrutiny with more data; a real signal would show up as a growing deviation in this bin.","The exclusion strength depends on the assumed 5% branching fraction for the heavier neutralino decaying to two muons or two electrons plus the lightest neutralino; a larger branching fraction would widen the excluded region."],"fun_headline_variants":["No Higgsinos down to 1.5 GeV mass splitting","Tightest split yet: Higgsino search finds nothing at 1.5 GeV","Soft-lepton search excludes Higgsinos down to 1.5 GeV split","CMS probes compressed Higgsinos: null result to 1.5 GeV"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The background estimate assumes that the ratio of jet-related background production in the isolation sideband versus the main band, calibrated in the BDT<0 region, stays the same in the high-BDT signal bins where the signal would appear.","fun_headline_variants_meta":{"raw":{"variants":["No Higgsinos down to 1.5 GeV mass splitting","Tightest split yet: Higgsino search finds nothing at 1.5 GeV","Soft-lepton search excludes Higgsinos down to 1.5 GeV split","CMS probes compressed Higgsinos: null result to 1.5 GeV"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000818,"raw_usage":{"total_tokens":3465,"prompt_tokens":838,"completion_tokens":2627,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":582,"completion_tokens_details":{"reasoning_tokens":2558}},"tokens_in":582,"tokens_out":2627,"duration_ms":16433,"temperature":1.0,"reasoning_tokens":2558,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T21:06:46.133807+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A concrete check: using the full simulation, compute the sideband-to-main-band transfer factor separately in each BDT bin above 0.4 and see whether it drifts outside the assigned 8–22% systematic uncertainty. Alternatively, with the roughly 300 fb−1 already available from Run 3, if the tightest dimuon signal region shows an excess over the background prediction of more than 3 standard deviations, the conclusion that the 2.2σ deviation is a fluctuation would be wrong.","supporting_citations":[],"review_version":2}