REVIEW 3 major objections 8 minor 6 references
Reviewed by Pith at T0; open to challenge.
T0 means a machine referee read the full paper against a public rubric. The mark states how deep the mechanical check went, never who wrote it. the ladder, T0–T4 →
T0 review · glm-5.2
LHC has not ruled out light supersymmetry, study finds
2026-07-09 08:42 UTC pith:TIEJEYLT
load-bearing objection SModelS reproduces ATLAS pMSSM EWKino exclusions for bino-like LSPs but under-excludes by ~46% for non-bino LSPs; the gap is partially explained but not fully quantified. the 3 major comments →
On the coverage of electroweak-inos within the pMSSM with SModelS -- a comparison with the ATLAS pMSSM study
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
Core claim
SModelS v3.0 reproduces ATLAS's internal pMSSM electroweak-ino constraints with good agreement, and extending to the full analysis database with statistical combination of up to 37 uncorrelated ATLAS and CMS analyses increases exclusion coverage by 20% observed (35% expected) relative to the original ATLAS study. Despite this, light electroweak-inos below the TeV scale remain viable, especially higgsino-like LSP scenarios.
What carries the argument
SModelS v3.0 matches full-model BLM signatures to a database of simplified-model (SMS) experimental results, computes an r-value (predicted signal cross section divided by the 95% CL upper limit) for each analysis, and statistically combines uncorrelated analyses into a single likelihood per parameter point.
Load-bearing premise
The paper assumes that SModelS's simplified-model matching faithfully reproduces the full detector-level simulation used by ATLAS, but does not quantify the systematic uncertainty introduced by the SMS approximation across the full parameter space.
What would settle it
A specific pMSSM parameter point that SModelS declares excluded (r > 1) but that full detector-level simulation shows is allowed, or vice versa, would indicate a breakdown of the simplified-model matching in that parameter region.
If this is right
- Light electroweak-inos, particularly higgsino-like LSPs below about 700 GeV in chargino mass, remain experimentally viable and are primary targets for LHC Run 3 searches.
- Statistical combination of uncorrelated analyses across experiments is not optional but necessary for maximizing SUSY parameter-space coverage; single-analysis reinterpretation leaves significant loopholes.
- Public release of efficiency maps for all relevant simplified-model topologies, especially for three-lepton and disappearing-track channels, is the limiting factor for faithful reinterpretation and should be prioritized by experiments.
- Naive luminosity rescaling suggests LHC Run 3 results will probe the yellow-to-red r-value region of currently allowed parameter space, but a substantial region of light higgsinos may persist beyond Run 3.
Where Pith is reading between the lines
- If the SMS-to-full-simulation matching introduces composition-dependent biases, the 20-35% improvement from analysis combination could be unevenly distributed across pMSSM regions, over-excluding some and under-excluding others in ways that distort the surviving parameter-space map.
- The identified coverage gaps (disappearing tracks at 36 vs 136 fb^-1, missing Wh+MET and ZZ+MET maps, unmatched multi-step decays) suggest that the true exclusion coverage with complete inputs could be meaningfully higher than even the combined result presented here.
- The persistence of light higgsino parameter space after aggressive combination implies that future exclusion will depend less on luminosity and more on dedicated compressed-spectrum search strategies targeting small mass gaps.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. This paper compares the exclusion coverage of electroweak-ino (EWKino) parameter space in the pMSSM as obtained by SModelS v3.0 against the official ATLAS pMSSM study (arXiv:2402.01392). Using the publicly released ATLAS scan data (8,953 points after filtering), the author computes NLO cross sections with Resummino and applies SModelS constraints from individual analyses, the full database, and statistical combinations of uncorrelated analyses. The paper identifies specific causes for discrepancies between SModelS and ATLAS (e.g., the disappearing track analysis being at 36 fb$^{-1}$ in SModelS vs. 136 fb$^{-1}$ in ATLAS, missing efficiency maps for certain topologies, and unmatched multi-step decays). The main quantitative claims are: (i) SModelS shows 'good agreement' with ATLAS, and (ii) extending to the full database with analysis combination yields a 20% (35% expected) increase in exclusion coverage. The paper concludes that light EWKinos, particularly higgsinos, remain viable.
Significance. The paper provides a useful and timely validation of SModelS v3.0 against a full-experiment reinterpretation, which is valuable for the BSM phenomenology community that relies on SModelS for rapid parameter-space scans. The identification of specific coverage gaps (disappearing tracks luminosity, missing $Wh$+MET and $ZZ$+MET efficiency maps for SUSY-2019-09, unmatched multi-step decays for SUSY-2018-05) is constructive and actionable. The analysis-combination methodology builds on prior SModelS collaboration work (Ref. 4) and demonstrates a concrete, quantified gain in exclusion coverage. The physics conclusion about viable light higgsinos is of broad interest. The study is reproducible in principle, as it uses publicly released ATLAS scan data and a public tool.
major comments (3)
- Table 2 reveals a substantial asymmetry in the SModelS-vs-ATLAS comparison that is not adequately discussed in the text. For bino-like LSPs (3,034 points), SModelS with ATLAS EWKino results excludes 575 (19%) vs. ATLAS's 529 (17%) — a slight over-exclusion consistent with 'good agreement.' However, for non-bino LSPs (5,919 points, 66% of the dataset), SModelS excludes only 687 (12%) vs. ATLAS's 1,271 (21%) — a 46% relative shortfall. Even with the full database and analysis combination (row 5), SModelS reaches 1,184 (20%), still below ATLAS's 21% for this category. The text in §2.3 attributes the under-exclusion to three causes (disappearing tracks at 36 vs. 136 fb$^{-1}$ accounting for ~60% of the gap, missing $Wh$+MET/$ZZ$+MET efficiency maps, and unmatched multi-step decays), but no quantitative breakdown is provided: the reader cannot determine how many of the 584 under-excluded non-
- The 'good agreement' claim in §2.3 and the abstract is not supported by the data for the majority of the parameter space. The abstract states 'good agreement' without qualification, while Table 2 shows this holds only for bino-like LSPs. For non-bino LSPs (66% of the dataset), the agreement is poor at the level of individual analyses and only reaches parity with ATLAS after the full database plus analysis combination — and even then, only because ATLAS's own exclusion for non-bino LSPs is 21% vs. SModelS's 20%, not because the individual-analysis comparison shows agreement. The abstract and §2.3 should explicitly state the bino/non-bino asymmetry and qualify the 'good agreement' claim accordingly. As written, the claim is misleading for a reader who examines only the abstract or the summary in §2.3.
- The conclusion that 'light EWKinos below the TeV scale remain viable' (§3, §4) is drawn from the analysis-combination results (Fig. 3). However, for non-bino LSPs — which include the higgsino LSP scenarios central to this conclusion — SModelS under-excludes relative to ATLAS even after analysis combination (Table 2, row 5: 20% vs. 21%). More critically, Fig. 3 incorporates the ATLAS 140 fb$^{-1}$ disappearing track results (as noted in the caption footnote), which are not part of the SModelS database and are applied as an external post-hoc cut. The paper should clarify exactly how this external constraint is applied in Fig. 3, whether it affects the allowed-point classification, and how sensitive the 'light higgsinos remain viable' conclusion is to this external input. Without this clarification, it is unclear whether the viable region in Fig. 3 reflects SModelS's own constraining power.
minor comments (8)
- §2.2: The filtering criteria are stated, but the number of points removed at each step is not given. Providing the breakdown (e.g., N filtered by ATLAS, N removed for BR($h$→inv), N removed for $m_A$ limits, N removed for the SPheno bug) would improve transparency.
- Table 1: The '3 leptons on-shell / off-shell' row for SUSY-2019-09 lists 'YES' but the text in §2.3 states that $Wh$+MET and $ZZ$+MET efficiency maps are missing for this analysis. The table should note this partial implementation, or the text should clarify which efficiency maps are and are not available.
- §3, Fig. 2 caption: The figure shows 'EWKino results in SModelS' (black) and 'full SModelS database' (light blue), but the 'full DB with analysis combination' (blue) is also shown. The caption should explicitly state what each colored line represents, as the current description is incomplete.
- §3: The 20% (35% expected) increase in exclusion is stated without specifying the baseline. Is this relative to the ATLAS EWKino-only SModelS result (row 2 of Table 2), or relative to ATLAS's own exclusion (row 1)? The baseline should be stated explicitly.
- Fig. 3: The color scale for the combined $r$-value is not shown or described in sufficient detail. The reader needs to know the range of $r$ values and how they map to the color code to interpret the 'naïve rescaling' argument about Run 3 sensitivity.
- §2.3: The phrase 'about 60% of the under-excluded points' is attributed to the disappearing track luminosity gap. It is unclear whether this refers to 60% of all under-excluded points or 60% of under-excluded points in a specific mass range. This should be clarified.
- The reference to Ref. 3 (the companion paper arXiv:2512.14502) is cited as the source of the study setup. The relationship between this paper and Ref. 3 should be clarified — is this a summary of Ref. 3, or does it contain new results not in Ref. 3?
- §3, footnote c: The footnote text is truncated or awkwardly placed. It states that points allowed by the combination but excluded by the ATLAS 136 fb$^{-1}$ disappearing track analysis have been removed, but this information is critical to the interpretation of Fig. 3 and should be stated more prominently in the main text.
Simulated Author's Rebuttal
We thank the referee for a careful reading and constructive comments. The referee raises three major points, all of which we find legitimate. We agree that the bino/non-bino asymmetry in Table 2 must be discussed quantitatively, that the 'good agreement' claim needs qualification, and that the role of the external disappearing-track constraint in Figure 3 requires clarification. We outline below our planned revisions for each point.
read point-by-point responses
-
Referee: Table 2 reveals a substantial asymmetry in the SModelS-vs-ATLAS comparison that is not adequately discussed in the text. For bino-like LSPs, SModelS shows slight over-exclusion, but for non-bino LSPs (66% of the dataset), SModelS excludes only 12% vs ATLAS's 21% — a 46% relative shortfall. The text attributes the under-exclusion to three causes but provides no quantitative breakdown of how many of the under-excluded non-bino points are accounted for by each cause.
Authors: The referee is correct that the bino/non-bino asymmetry is substantial and that the text does not provide a quantitative breakdown of the under-exclusion causes for non-bino LSPs. We will revise the manuscript to include this breakdown. Specifically, we can quantify the contributions as follows: approximately 60% of the under-excluded non-bino points are attributable to the disappearing-track analysis being at 36 fb^{-1} in SModelS vs 136 fb^{-1} in ATLAS (this affects primarily wino-like LSPs with small mass splittings). A further fraction is due to missing Wh+MET and ZZ+MET efficiency maps in SUSY-2019-09, which particularly impacts higgsino-like LSP scenarios. The remaining gap comes from unmatched multi-step decays in SUSY-2018-05. We will add a dedicated paragraph (or table) in Section 2.3 giving the point counts for each cause, so the reader can see explicitly how the 584-point shortfall (1271 - 687) decomposes. We note that a precise one-to-one assignment is not always possible because some points are affected by more than one cause simultaneously; we will be transparent about this overlap. revision: yes
-
Referee: The 'good agreement' claim in the abstract and Section 2.3 is not supported by the data for the majority of the parameter space. For non-bino LSPs (66% of the dataset), the agreement is poor at the level of individual analyses and only reaches parity with ATLAS after the full database plus analysis combination. The abstract and Section 2.3 should explicitly state the bino/non-bino asymmetry and qualify the 'good agreement' claim accordingly.
Authors: We agree. The unqualified 'good agreement' claim in the abstract is misleading for non-bino LSPs, which constitute the majority of the dataset. We will revise the abstract to state that good agreement is observed for bino-like LSPs, while for non-bino (wino/higgsino) LSPs the individual-analysis comparison shows a significant under-exclusion that is largely recovered only when the full database and analysis combination are employed. We will make a corresponding revision in Section 2.3, adding an explicit discussion of the asymmetry visible in Table 2. This will include the numbers the referee cites (19% vs 17% for bino; 12% vs 21% for non-bino at the individual-analysis level; 20% vs 21% after full combination). We believe this makes the presentation honest and transparent. revision: yes
-
Referee: The conclusion that 'light EWKinos below the TeV scale remain viable' is drawn from analysis-combination results (Fig. 3), but for non-bino LSPs SModelS under-excludes relative to ATLAS even after analysis combination. More critically, Fig. 3 incorporates the ATLAS 140 fb^{-1} disappearing track results, which are not part of the SModelS database and are applied as an external post-hoc cut. The paper should clarify exactly how this external constraint is applied, whether it affects the allowed-point classification, and how sensitive the 'light higgsinos remain viable' conclusion is to this external input.
Authors: This is a fair and important point. We will clarify in the revised text that the ATLAS 140 fb^{-1} disappearing-track result is applied as an external post-hoc exclusion: any point that is allowed by the SModelS analysis combination but was excluded by ATLAS's 136 fb^{-1} disappearing-track analysis is removed from the allowed set shown in colour in Figure 3 (this is already noted in the footnote, but we agree it needs to be stated more prominently in the main text). We will also add a discussion of sensitivity: the disappearing-track constraint primarily removes compressed wino-like LSP points with small chargino-neutralino mass splittings. For higgsino-like LSPs, which are central to the 'light EWKinos remain viable' conclusion, the disappearing-track analysis has limited sensitivity because the chargino lifetime is shorter and the track length is typically below detection thresholds. Therefore, the viability conclusion for light higgsinos does not critically depend on this external input. We will state this explicitly. We also note that the under-exclusion of non-bino LSPs relative to ATLAS (20% vs 21% after combination) means that the allowed region in Figure 3 is, if anything, slightly conservative — SModelS's own constraining power alone would leave a few additional points allowed, and the external disappearing-track cut compensates for part of this gap. revision: yes
Circularity Check
No circularity found; the paper is a validation study against an external benchmark with no self-referential derivation chain.
full rationale
The paper's central derivation chain is straightforward and non-circular: (1) it takes ATLAS's publicly released pMSSM scan data and exclusion constraints (ref 2, ATLAS collaboration) as an external benchmark; (2) it runs the independently developed SModelS v3.0 tool (ref 1) on the same scan points; (3) it compares the exclusion counts numerically (Table 2, Fig. 1); (4) it extends to the full database and applies a statistical combination method (ref 4, Altakach et al.) to quantify coverage gains. No step reduces to its inputs by construction. The self-citations (refs 3 and 5 by Constantin et al.) are companion/related works — ref 3 appears to be the full-length version of this conference contribution, and ref 5 is a separate physics-interpretation paper — but neither is load-bearing for the comparison methodology or the numerical results. The 'good agreement' and '20% increase' claims are direct numerical outputs of running an independent tool on external data, not fitted parameters renamed as predictions. The SModelS SMS-matching approximation is a methodological limitation (correctness risk), not a circularity: the tool's exclusions are computed from experimental efficiency maps and cross sections, not defined in terms of the ATLAS results they are compared against.
Axiom & Free-Parameter Ledger
free parameters (1)
- None introduced by this paper =
N/A
axioms (3)
- domain assumption Simplified Model Spectra (SMS) results from ATLAS and CMS faithfully approximate the full detector-level simulation for the pMSSM topologies considered.
- domain assumption The analyses selected for combination are uncorrelated.
- standard math Resummino NLO cross sections and NLLfast NNLL gluino cross sections are accurate for the pMSSM parameter space scanned.
read the original abstract
The ATLAS collaboration has recently performed a vast scan of the phenomenological Minimal Supersymmetric Standard Model (pMSSM) with a focus on the electroweak-ino sector, and analysed how their Run 2 searches for electroweak production of supersymmetric (SUSY) particles constrain this dataset. All the SLHA files from the scan as well as the constraints from the eight individual searches considered by ATLAS were made publicly available. We use this material to study how well the ATLAS constraints can be reproduced with SModelS v3.0. Moreover, we explore how the picture changes when also including CMS results, and what can be gained by the statistical combination of analyses. Finally, we discuss the part of parameter space with light electroweak-inos that remains valid despite the stringent LHC limits. Our results underscore the need of a broad, multifaceted approach for maximising sensitivity and closing loopholes in the extensive SUSY parameter space.
Figures
Reference graph
Works this paper leans on
-
[1]
Aad, G. and others. ATLAS Run 2 searches for electroweak production of supersymmetric particles interpreted within the pMSSM. JHEP. 2024. doi:10.1007/JHEP05(2024)106. arXiv:2402.01392
work page internal anchor Pith review Pith/arXiv arXiv doi:10.1007/jhep05(2024)106 2024
-
[2]
SModelS v3: Going Beyond Z2 Topologies
Altakach, M.M. and others. SModelS v3: going beyond Z _ 2 topologies. JHEP. 2024. doi:10.1007/JHEP11(2024)074. arXiv:2409.12942
work page internal anchor Pith review Pith/arXiv arXiv doi:10.1007/jhep11(2024)074 2024
-
[3]
SModelS v2.3: enabling global likelihood analyses
Altakach, M.M. and others. SModelS v2.3: Enabling global likelihood analyses. SciPost Phys. 2023. doi:10.21468/SciPostPhys.15.5.185. arXiv:2306.17676
work page internal anchor Pith review Pith/arXiv arXiv doi:10.21468/scipostphys.15.5.185 2023
-
[4]
Global LHC constraints on electroweak-inos with SModelS v2.3
Altakach, M.M. and others. Global LHC constraints on electroweak-inos with SModelS v2.3. SciPost Phys. 2024. doi:10.21468/SciPostPhys.16.4.101. arXiv:2312.16635
work page internal anchor Pith review Pith/arXiv arXiv doi:10.21468/scipostphys.16.4.101 2024
-
[5]
Constantin, L. and others. On the coverage of electroweak-inos within the pMSSM with SModelS -- a comparison with the ATLAS pMSSM study. arXiv:2512.14502. 2025. arXiv:2512.14502
work page internal anchor Pith review Pith/arXiv arXiv 2025
-
[6]
Constantin, L. and Kraml, S. and Mahmoudi, F. The LHC has ruled out supersymmetry really?. Nucl. Phys. B. 2025. doi:10.1016/j.nuclphysb.2025.117012. arXiv:2505.11251
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.