{"id":"003698d8-4a36-43c1-be3a-875d49f98f0f","arxiv_id":"2411.09652","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":9,"one_line_summary":"High-energy equal-mass black hole scattering data show that L-resummation, weob, and SEOB-PM models all develop pathological behavior, with NR-calibrated pseudo-5PM corrections and Padé-resummed EOB potentials offering partial improvement.","lead":"New numerical relativity simulations of high-energy black hole scattering are used to compare three resummation schemes for post-Minkowskian scattering angles. All schemes fail at high energies, and NR-informed pseudo-5PM terms improve agreement only when the critical angular momentum is well estimated.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The paper's central comparison rests on new NR data at Γ4–Γ7 that are not validated by resolution or extraction-window convergence tests; if those data carry systematic errors, the claim that all resummations fail at high energies could be an artifact.","rationale":"The reader's weakest_assumption focuses on the partial 5PM term being representative of the full 5PM order. That is a valid secondary caveat, and the authors themselves flag it in Sec. IVC3 and the conclusions. However, it does not bear on the paper's primary empirical claim that all three resummations fail at high energies: that claim is based on the 2PM/3PM/4PM predictions, which do not involve the partial 5PM term, and on the NR data themselves. The SEOB-PM model is shown only up to 4PM, and the L-resummed and weob 4PM curves already exhibit the divergent behaviour. Thus even if the full 5PM term were completely different, the conclusion that the currently available resummations are unreliable at high energies would stand.\n\nThe load-bearing element for that conclusion is the reliability of the new NR scattering angles at Γ4–Γ7. The paper presents these as the new data, but provides no convergence study or independent cross-check. The quoted error bars only capture the spread from polynomial-order variation in the extrapolation, not the dominant systematic errors (resolution, finite extraction radius, initial-data effects). The manuscript does include a junk-radiation analysis (App. A) that shows those effects are small, but it does not address the discretization or extrapolation systematic. Given that the near-plunge points show extraction errors of order tens of degrees, a detailed convergence and cross-validation study is necessary to establish that the high-energy comparison is well grounded.\n\nMy recommendation is therefore unchanged: the paper merits a conditional acceptance, with the condition being that the authors provide (or reference) a resolution and extraction-window convergence study for at least the new high-energy runs, and ideally an independent-code comparison for one configuration. This is consistent with the reader's CONDITIONAL verdict, although I differ as to which assumption is the most load-bearing.","tokens_in":43427,"tokens_out":15092,"duration_ms":142498,"concrete_test":"Re-run one high-energy case (e.g., Γ7, b=5.5M) with the Einstein Toolkit at three grid resolutions (e.g., standard, 1/2, and 2 times the nominal spacing), and repeat the extraction with the outgoing window changed from [30,180]M to [50,150]M and [40,250]M. If the resulting θNR shifts by more than the quoted sub-degree error bars or by an amount comparable to the analytical-model residuals (tens of degrees), the paper's central comparison is not robust. In addition, cross-check one Γ7 configuration with an independent NR code (e.g., SpEC or bamps) to rule out code-dependent bias.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The manuscript's central claim, that all three resummation schemes become unreliable at high energies, is established by comparing analytical predictions to NR scattering angles at Γ4–Γ7. The error bars quoted in Tables II–VIII are derived solely from the variation of the polynomial order used to extrapolate the puncture trajectories in 1/r (Sec. IIB). No convergence test with respect to grid resolution, extraction radius, or gauge choice is reported for the new runs, and no independent-code verification is offered. The delicate nature of the extraction is visible in the near-plunge data points, which carry asymmetric errors as large as −11° (Table II) and −16° (Table VIII) and are excluded from the quantitative analysis by an unspecified boundary rule. Since the model residuals against NR are tens to hundreds of degrees, modest systematic errors in the reported angles would not change the qualitative conclusion; however, a bias of that size cannot be excluded from the information presented. If the systematic error in the high-energy extraction is comparable to the residuals, the claimed 'PM hierarchical shifts and divergences' would be a property of the data analysis rather than of the physical models.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper extends the existing NR catalogue of equal-mass, non-spinning binary black-hole scattering by adding four new energy sets (Γ4–Γ7, up to γ ≈ 1.96) and re-extracting the low-energy data with a common polynomial-extrapolation pipeline (Sec. II, App. B). It then compares the scattering-angle predictions of three resummation schemes — the L-resummation [82], the w^eob EOB-potential resummation [82], and the SEOB-PM model [84], together with the partial 5PM (conservative, 1GSF) information of Eq. (3.15) — against the NR angles (Sec. IV). The central findings are that all three schemes develop PM-order-hierarchy inversions and divergent or unphysical behaviour in the high-energy regime, that NR-calibrated pseudo-5PM coefficients (θ5,I, θ5,II and w5,I–III) can repair the agreement in-sample at varying levels, and that Padé resummation of the EOB radial potentials is a promising, proof-of-principle remedy (Secs. IVB–IVC). The paper explicitly flags the effective nature of the pseudo-5PM terms and the incompleteness of the 5PM input.","tokens_in":43709,"tokens_out":15580,"duration_ms":144522,"significance":"The high-energy scattering regime is directly relevant to using PM information in the next generation of EOB and phenomenological waveform models, and this paper provides the first systematic, cross-model stress test of the three main resummation proposals at γ up to 1.96. If the central claim holds, it is an important caution: naive inclusion of higher PM orders (and the associated radiative terms) does not improve, and can degrade, strong-field predictions, so resummation strategy — not just PM order — is the key design choice. The manuscript ships the full NR data tables (App. B), uses an extraction pipeline that follows the established references [81,83], compares models in a transparent, side-by-side way, and is explicit about the main caveats (partial 5PM content, effective nature of the calibrated coefficients, proof-of-principle status of the Padé analysis). These are genuine strengths that make the paper useful even where the conclusions are provisional.","major_comments":[{"comment":"The new Γ4–Γ7 data carry the paper's central empirical claim, but their quoted error bars are derived solely from the variation of the polynomial order in the 1/r puncture-trajectory extrapolation (Sec. IIB). No resolution-convergence study, extraction-radius or fit-window variation, gauge-dependence test, or independent-code comparison is reported for the new runs, and Sec. IIA gives the differencing stencil and gauge choices but no grid resolutions or refinement-level parameters. The agreement with [83] at lower energies provides some pipeline validation, but it does not by itself establish the accuracy of the new high-energy runs. The qualitative conclusion that the models fail by tens to hundreds of degrees at the highest energy would survive modest systematic errors, but the claimed sub-degree precision of many entries in Tables II–VIII is not verifiable from the information given, and the near-plunge points already carry −11° to −16° asymmetric errors (Tables II and VIII). Please add convergence/consistency checks for representative runs (including at least one strong-field point per new energy), report the grid parameters, and state whether the extrapolation-window choice (r ∈ [20,90]M and [30,180]M) is stable at Γ4–Γ7.","section":"IIB, Tables II–VIII"},{"comment":"The abstract's statement that “including higher-order information improves the agreement” rests on two legs, and both are weaker than the sentence suggests. First, the only genuine 5PM input, θ5,cons^{1GSF}, is the conservative first-order-self-force piece; the radiative and 2GSF contributions are unknown, a point the authors themselves make in Sec. IVC3, so the partial-5PM comparisons in Figs. 3, 4, 8 and 14 do not yet characterise the full 5PM order. Second, the pseudo-5PM coefficients θ5,I, θ5,II, w5,I and w5,III are calibrated against the same NR data sets against which the resulting models are then evaluated (Figs. 7, 12, 13), so the reported improvements — e.g., residuals ∼1° at Γ1 and the Padé gains at Γ7 — are in-sample demonstrations of the flexibility of the templates rather than evidence of predictive superiority. The body is appropriately cautious (“effective parameters”, “proof-of-principle”), but the abstract and Sec. V should either carry the same caveats explicitly or be supported by an out-of-sample test (e.g., calibrating on Γ4–Γ6 and evaluating at Γ7).","section":"IIIB Eq. (3.15); IVB2–IVB3, IVC2–IVC3"},{"comment":"The reconstructed “NR potential” wNR is not a direct measurement: it is obtained by fitting the three-parameter L-resummed template of Eq. (4.24) (ℓ0,NR, θ5,NR, θ6,NR) to the scattering angles and then Abel-inverting the fitted, extrapolated curve with Eq. (4.23). The comparison in Fig. 10 between wNR and the w^eob potentials is therefore only as model-independent as that template; at high energies the fitted curve extrapolates beyond the measured ℓ-range and into the near-plunge region, and this extrapolation uncertainty is not propagated. Part of the reported discrepancy between wNR and the PM-expanded potentials could in principle reflect template bias rather than a genuine failure of the w^eob potentials. Please state this limitation when interpreting Fig. 10, or test the sensitivity of wNR to the assumed functional form.","section":"IVC2, Eq. (4.23)"}],"minor_comments":[{"comment":"The row “1.07727 6.80 5.440 214.3263 +0.8919 −0.0007” lists a value of Γ that is inconsistent with the other rows of the table; it should read Γ = 1.07277.","section":"Table V"},{"comment":"In the conclusions, “psuedo-5PM” should read “pseudo-5PM” (the misspelling appears twice).","section":"V"},{"comment":"The rule for flagging data with an asterisk is described only as “due to unbound/plunge uncertainty”; please quantify the selection criterion and state explicitly whether the excluded points affect any of the reported fits.","section":"App. B"},{"comment":"The figure compares Cauchy estimates of ℓ0 with NR-derived values, but the procedure for extracting ℓ0 from the NR data and the handling of its uncertainties are not described in Sec. II; a sentence specifying the procedure (or a pointer to the [82] prescription) would help reproducibility.","section":"Fig. 5"},{"comment":"The arXiv abstract and the main-text abstract differ in the phrasing of the central claim (“Each model is shown to demonstrate pathological behaviour” vs. “All of the models struggle to accurately capture the behavior”); please harmonize the two versions.","section":"Abstract"},{"comment":"The sentence “…with divergent behaviour in the ¯r → 0 being dictated by w5,I[83]” has an ambiguous citation placement; clarify whether [83] is the source of this behaviour or whether the bracket is a typographical artifact.","section":"IVC3"}],"recommendation":"major_revision","confidential_remarks":"To the editor: this is a solid and useful paper that fits the journal well, but its revision hinges on the numerical-relativity validation. The referee's major comment 1 is the one I would weigh most heavily: the paper's new empirical payload is the Γ4–Γ7 data, and the community standard for NR work is to report resolution-convergence and extraction-consistency results. Given that the same pipeline has been validated at lower energies against [81,83], I would not treat this as fatal — the qualitative high-energy conclusions are robust to the plausible size of systematic errors — but the claims of sub-degree precision cannot stand without supporting evidence. I also recommend that the authors recalibrate the abstract's “higher-order information improves the agreement” sentence so that in-sample calibration is not presented as predictive improvement. If these two points are addressed, I would expect the paper to be publishable; if the authors can also supply an out-of-sample test for one of the pseudo-5PM schemes, the paper would be considerably stronger."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe paper is worth reading if you work on PM-resummed scattering or EOB modeling. It adds four new high-energy NR scattering simulations (Γ4-Γ7) and uses them to compare the L-resummed, weob, and SEOB-PM models across the full energy range. The data tables are complete, the extraction procedure follows established references, and the authors are candid about the limits of what they can conclude.\n\nThe strongest part is the qualitative result: all three resummation schemes degrade as the energy increases, with PM hierarchical shifts and divergences. This is supported by large residuals, tens to hundreds of degrees, so it does not depend on delicate error bars. The paper also demonstrates that NR-calibrated pseudo-5PM coefficients and a Padé-resummed EOB potential can improve agreement, but flags correctly that these are effective parameters, not a derivation of the true 5PM term.\n\nThe soft spots are real but manageable. The new NR runs have no resolution or extraction-window convergence tests; the error bars come only from the polynomial-order variation in the 1/r extrapolation. That is a genuine gap for a benchmark paper, and the near-plunge points carry asymmetric errors up to -16°. The rule for excluding some points from the quantitative analysis is not specified. For the higher-order-improvement claim, the only 5PM input is the conservative 1GSF term, and the authors themselves note that missing radiative and 2GSF contributions could change the conclusions. The pseudo-5PM fits absorb non-perturbative physics, so they are a proof of principle, not a measurement.\n\nI would send this to peer review. The central comparison is transparent, the data are useful, and the conclusions are framed more carefully than the abstract suggests. The referees should ask for convergence tests and code/data release, and should push on the exclusion rule. But the core message is solid: none of the current resummations work at high energy, and better resummations, or more complete 5PM information, are needed.\n\nBring it to the reading group and keep it on your citation list if you do PM-NR comparisons.","headline":"New NR scattering data at Γ4-Γ7 make this a useful benchmark paper; the central qualitative finding is solid, but missing convergence tests and reliance on fitted pseudo-5PM terms need attention.","tokens_in":44266,"tokens_out":2500,"would_cite":true,"duration_ms":22877,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["04.25.dg","04.30.-w","04.70.-s"],"model":"deepseek-v4-flash","headline":"This paper shows that L-resummed, weob, and SEOB-PM scattering models all break down at high energies, and that NR-calibrated pseudo-5PM terms with Padé-resummed EOB potentials restore partial agreement.","keywords":["post-Minkowskian expansion","resummation","scattering angle","numerical relativity","black hole scattering","effective one body","gravitational waves","high-energy limit"],"falsifier":"Compute the complete 5PM scattering-angle coefficient including radiation-reaction and second-self-force contributions, re-run the L-resummed, weob, and SEOB-PM predictions against the $\\Gamma_7$ NR data, and check whether the PM hierarchical shifts and divergences disappear; if they persist, the paper's remedy claim fails.","tokens_in":43199,"feed_emoji":"🕳️","tokens_out":7775,"duration_ms":65012,"temperature":0.7,"pith_summary":"The paper tests three resummation strategies that convert post-Minkowskian (PM) perturbative information into predictions for the scattering angle of two black holes, using new high-energy numerical-relativity simulations of equal-mass nonspinning binaries up to a Lorentz factor of 1.96. It finds that all three schemes—the L-resummed model, the weob model, and the SEOB-PM model—reproduce the numerical data well at low energies but fail at high energies, showing PM-order hierarchies that shift or diverge. The paper then uses the NR data to calibrate pseudo-5PM terms and Padé-resummed EOB radial potentials, arguing that including higher-order information improves agreement with NR, although the improvement depends on the resummation and on whether the critical angular momentum is estimated analytically or calibrated to NR.","feed_headline":"PM resummations break down in high-energy black hole scattering","feed_subtitle":"New numerical runs up to Lorentz factor 1.96 show divergences; NR-tuned 5PM terms and Padé potentials restore accuracy.","key_machinery":"The paper's central objects are the PM expansion of the scattering angle $\\theta(\\gamma,\\ell)=\\sum_i 2\\theta_i(\\gamma)/\\ell^i$ and three ways of resumming it. The L-resummation factors out the singular logarithmic behavior $L(x)=\\frac{1}{x}\\ln\\frac{1}{1-x}$ near the critical angular momentum $\\ell_0$, where scattering turns into plunge. The weob model rewrites the scattering integral in terms of a PM-expanded EOB radial potential $w(\\bar{r},\\gamma)$, and the SEOB-PM model instead feeds PM information into the EOB metric potential $A(r)$ in the post-Schwarzschild gauge. A Firsov-type inversion formula (Eq. 4.23) lets the authors extract the NR-informed potential $w_{\\mathrm{NR}}$ from their scattering angles, and Padé approximants of $w$ are tested as an additional resummation layer.","core_discovery":"The central discovery, on the paper's own terms, is that no current resummation of PM information is reliable in the strong-field, high-energy regime: the L-resummed angles develop divergences tied to the Cauchy estimate of $\\ell_0$, the weob potentials become over-attractive at 4PM and develop repulsive cores at low energy when partial 5PM information is added, and the SEOB-PM hierarchy inverts so that 2PM outperforms 4PM above $\\Gamma_5$. The paper demonstrates that NR-calibrating a pseudo-5PM coefficient improves the L-resummed model, and that a Padé resummation of the EOB radial potential before calibration gives a proof-of-principle cure for the weob model's high-energy failures.","pith_inferences":["The paper leaves open the possibility that full 5PM results, with radiation reaction and 2GSF terms, could change the hierarchy: the authors' own Sec. IVC3 caveat implies their pseudo-5PM fits may absorb effects beyond their nominal order.","A natural extension is to apply the same NR-calibration plus Padé procedure to spinning and unequal-mass binaries; nothing in the argument limits it to equal-mass nonspinning systems, but that is untested.","Because the pseudo-5PM coefficients are fit at each energy, using them in a waveform model would require a prescription for interpolating across energies or refitting the coefficient as a smooth function of $\\gamma$.","The universal logarithmic singularity motivating the L-resummation may need refinement at high energy: the self-force-enhanced singularity found in the scalar-field analogue could change $\\ell_0$ estimates and rescue the scheme, a direction the paper cites but does not pursue."],"forward_implications":["If the paper is right, waveform models that rely on these PM resummations cannot be trusted for high-energy or strong-field encounters until more PM orders or better resummations are added.","Accurate knowledge of the critical angular momentum $\\ell_0$ can matter as much as higher-order PM terms: the L-resummed model with NR-calibrated $\\ell_0$ stays within a few percent of NR, while the analytic Cauchy estimate degrades to roughly 30 percent residuals at the highest energy.","Partial 5PM information generally improves accuracy below $\\Gamma_4$ but can introduce repulsive cores and over-attractive behavior at other energies, so adding higher orders is not automatically safer.","Padé-resummed EOB potentials, calibrated to NR after resummation (the $w_{\\mathrm{5PM,III}}$ procedure), give the best high-energy scattering angles and avoid the unphysical turning points seen in the plain PM-expanded potential.","The new NR data at $\\Gamma_4$–$\\Gamma_7$ provide a benchmark against which future complete 5PM or 6PM predictions, once radiation-reaction and second-self-force contributions are known, can be tested."],"supporting_citations":[{"why":"Supplies the map from classical scattering states to the EOB Hamiltonian that the weob and EOB-PM constructions rely on.","marker":"[35]"},{"why":"Derives the 2PM Hamiltonian and the PM-expanded EOB potentials used to define the weob model.","marker":"[36]"},{"why":"Provides the only available 5PM input, the conservative first-self-force coefficient $\\theta^{1\\mathrm{GSF}}_{5,\\mathrm{cons}}$, inserted into all three resummations.","marker":"[61]"},{"why":"Provides the original low-energy NR scattering simulations and the numerical setup that this paper extends to higher energies.","marker":"[81]"},{"why":"Introduces the L-resummation and weob model and the Firsov-inversion potential extraction used throughout the comparison.","marker":"[82]"},{"why":"Supplies the NR data at $\\Gamma_1$–$\\Gamma_3$ and the scattering-angle extraction procedure adopted and refined here.","marker":"[83]"},{"why":"Introduces the SEOB-PM model, the third resummation scheme against which the new NR data are compared.","marker":"[84]"},{"why":"Provides the partie-finie procedure used to match PM coefficients to the EOB-PM radial potentials in the weob construction.","marker":"[122]"},{"why":"Supplies the Padé approximant technique applied to the EOB radial potentials as the proposed improvement strategy.","marker":"[127]"}],"fun_headline_variants":["PM resummations break in strong-field black hole scattering","High-energy scattering exposes pathological PM resummations","NR calibration rescues black hole scattering resummations","Padé resummation cures high-energy black hole model failures","Black hole scattering: NR data tames diverging PM hierarchies"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the lone 5PM input currently available—the conservative first-self-force term $\\theta^{1\\mathrm{GSF}}_{5,\\mathrm{cons}}$—stands in for the complete 5PM order; if the missing radiative and second-self-force pieces are large, all three resummations' high-energy behavior could change, exactly as the authors warn in Sec. IVC3.","fun_headline_variants_meta":{"raw":{"variants":["PM resummations break in strong-field black hole scattering","High-energy scattering exposes pathological PM resummations","NR calibration rescues black hole scattering resummations","Padé resummation cures high-energy black hole model failures","Black hole scattering: NR data tames diverging PM hierarchies"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000367,"raw_usage":{"total_tokens":1977,"prompt_tokens":954,"completion_tokens":1023,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":570,"completion_tokens_details":{"reasoning_tokens":942}},"tokens_in":570,"tokens_out":1023,"duration_ms":9502,"temperature":1.0,"reasoning_tokens":942,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T20:25:20.217948+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the complete 5PM scattering-angle coefficient including radiation-reaction and second-self-force contributions, re-run the L-resummed, weob, and SEOB-PM predictions against the $\\Gamma_7$ NR data, and check whether the PM hierarchical shifts and divergences disappear; if they persist, the paper's remedy claim fails.","supporting_citations":[],"review_version":1}