{"id":"1693572f-b72a-47d9-9aa1-6b880f55068e","arxiv_id":"2608.10168","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"For eccentric black-hole binaries observed by LISA, a parametrized deformation of the periastron-precession rate could be constrained to |δα|≲10^-4 at 90% credibility with the full-model template.","lead":"This paper introduces a new way to test Einstein's theory of gravity using the upcoming LISA space observatory, by looking for tiny changes in how the orbits of eccentric black hole binaries twist as they spiral together. It shows LISA could detect deviations as small as one part in ten thousand, but also warns that the answer depends strongly on how the test is written into the waveform model.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The headline |δα| bound is computed and recovered with the same secularly resummed full model; the GR limit of that model is not the published 1PN eccentric waveform, so a cross-model injection-recovery test is needed before the forecast can be treated as robust.","rationale":"The reader's weakest_assumption identifies the same load-bearing concern: the forecasts rely on same-model injection-recovery with a full model whose GR limit is a secularly resummed approximant rather than the published 1PN eccentric waveform. My reading of Secs. II.C, III, and IV confirms this. The paper is transparent about the limitation, and the framework itself is well posed; the placement-dependence warning is a genuine contribution. The concern is about physical faithfulness of the fiducial model, not about internal consistency or mathematical error. The proposed cross-model check directly tests whether the quoted |δα| bound survives when the injected GR signal is generated with the restricted model (or equivalently Ref. [81]) and recovered with the full model. This is the minimal test that would settle whether the headline forecast is a statement about LISA's ability to constrain a physical frequency-ratio deformation or an artifact of the chosen approximant. Until such a test is performed, CONDITIONAL remains the appropriate verdict. The mass-notation ambiguity noted by the reader is real but secondary; the cross-model robustness issue is the single most load-bearing concern because it directly undermines the numerical claim that is the paper's headline result.","tokens_in":32655,"tokens_out":4341,"duration_ms":50439,"concrete_test":"Compute the noise-weighted mismatch between H_F(δα = 0) and H_R(δα = 0) for the fiducial source (M = 3000 Msun, q = 1.2, e0 = 0.5, Tobs = 4 yr, tc = 4 yr) using the same LISA TDI inner product. If the mismatch exceeds 1/(2ρ^2) at ρ = 50, the two GR approximants are statistically distinguishable. Then inject zero-noise H_R signals with δα = 0 and recover them with H_F; measure the median δα bias and the coverage of the 90% credible interval. If the recovered posterior does not contain zero, or if it is narrower than the mutual mismatch of the two GR approximants allows, the headline bound is not robust to model uncertainty.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central forecast in Sec. III — |δα| ≲ 10^-4 for M = 3000 Msun, e0 = 0.5, SNR = 50 — is generated and recovered entirely with H_F. But H_F's δα = 0 limit is not the published 1PN eccentric waveform of Ref. [81]; Eq. (38) promotes the secular drift phase to Newtonian (p = ±2) and half-PN (p = ±1, ±3) carriers, introducing corrections at relative 1PN and 1.5PN order beyond the formal 1PN truncation. The paper is explicit that all injections are generated and recovered with the same waveform model (Sec. III) and lists this as a limitation (Sec. IV). The concern is not that the framework is internally inconsistent; it is that the claimed sensitivity is a property of this particular finite approximant. If a real GR eccentric signal is better represented by H_R/Ref. [81] or by a higher-PN model, the accumulated carrier phase assigned in H_F would differ, and the δα posterior could be biased or the quoted bound inflated. The factor-of-eight comparison between H_F and H_R therefore does not by itself demonstrate robustness of the physical test, because the two models differ not only in where δα is placed but also in their GR waveforms. The placement-dependence warning is valuable and well argued, but the flagship numerical constraint rests on an unvalidated modelling choice.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper constructs a frequency-domain eccentric inspiral waveform model for LISA in which a parameter δα deforms the conservative azimuthal-to-radial frequency ratio K = Ω_θ/Ω_r, with δα = 0 recovering GR. Two prescriptions are defined: a restricted model that applies the deformation only to explicit 1PN precession sidebands, and a full model that promotes the deformed secular phase to the dominant Newtonian and half-PN angular carriers before resummation. Using lisabeta and Bayesian inference with zero-noise injections, the authors find that the full model yields 90% bounds |δα| ≲ 10^-4 for a chirp mass 3000 M_sun, e0 = 0.5, SNR = 50 binary, and that increasing eccentricity sharpens the constraints. A central methodological conclusion is that the inferred sensitivity depends strongly on where the phenomenological deformation is placed in a finitely truncated waveform.","tokens_in":33014,"tokens_out":4886,"duration_ms":48629,"significance":"If the projected bounds are robust, this is a valuable new null test of conservative GR dynamics in eccentric inspirals, complementing quasicircular parametrized tests and pulsar-timing measurements. The paper is careful in several respects: the SPA construction and harmonic truncation are checked against mismatch criteria, the LISA response and TDI channels are included, and the placement-dependence warning is explicitly demonstrated with a controlled likelihood diagnostic. The main caveat, acknowledged in Sec. IV, is that all injections and recoveries use the same waveform model; the full model's GR limit is not the published 1PN eccentric waveform of Ref. [81]. The central numerical claim therefore needs a cross-model validation before it can be read as a physical forecast.","major_comments":[{"comment":"The headline bound is obtained through zero-noise injections generated and recovered with the full model H_F, whose GR limit is not the published 1PN eccentric waveform of Ref. [81] but a secularly resummed approximant. The difference is not negligible: Eq. (38) promotes the deformed secular phase to Newtonian and half-PN carriers, adding relative 1PN and 1.5PN phase corrections beyond the formal truncation of Ref. [81]. Because the same model is used on both sides of the likelihood, the reported |δα| ≲ 10^-4 could be an artefact of the finite approximant rather than a measure of the physical frequency-ratio deformation. I request a cross-model injection-recovery study (e.g., inject H_R or Ref. [81] waveforms and recover with H_F, or use a higher-PN eccentric model) to bound the systematic contribution to δα; if that is not possible, the paper should state explicitly that the bound is a property of H_F and not a physical forecast.","section":"Sec. II.C.2, Sec. III, Sec. IV"},{"comment":"The full model's deformation of the half-PN carriers introduces corrections at relative 1.5PN order, which are outside the formal 1PN completeness of the baseline waveform. Since δα is claimed to isolate the conservative frequency ratio with dissipative and radiative sectors held fixed, the entanglement of δα with uncontrolled higher-PN phase terms weakens this claim. The manuscript should either show, through a comparison with a 1.5PN or 2PN eccentric waveform, that the δα posterior is insensitive to these omitted terms, or restrict the interpretation of δα to the specific finite approximant used here.","section":"Sec. II.C.2, Eq. (41)"}],"minor_comments":[{"comment":"In the sentence describing the restricted model, 'these terms terms' should be 'these terms'.","section":"Sec. II.C.1"},{"comment":"The restricted and full posteriors are computed at different SNRs (ρ = 400 and ρ = 50); although the text notes this, the figure captions should state the SNRs explicitly so that the posterior widths are not compared directly.","section":"Sec. III, Figs. 6 and 8"},{"comment":"The characteristic-strain envelopes sum absolute squares without cross terms; the text should clarify that these curves are diagnostics and do not represent the true spectral density of the signal.","section":"Sec. II.F, Eqs. (64)-(66)"},{"comment":"The paper calls the full model the 'fiducial phenomenological prescription' without a criterion for choosing it over the restricted model; a brief justification (e.g., the likelihood diagnostic in Fig. 10) would help the reader understand why H_F is preferred for forecasts.","section":"Sec. IV"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within the journal's scope and the methodology is generally careful. I expect that a cross-model injection-recovery test, or a clearly softened interpretation of the headline bound, can be accommodated within a revision; without one of these, the central forecast should not be presented as a physical result."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know about this one. The real contribution is the placement-dependence argument: the same conservative deformation δα, inserted at different places in a finite waveform, changes the projected LISA sensitivity by orders of magnitude. That is a general caution for theory-agnostic tests, and the paper makes it carefully. The headline number, |δα| ≲ 10^-4 for a 3000 M_sun, e0 = 0.5, SNR = 50 binary, is internally consistent but it is a same-model forecast: injections are generated and recovered with the same secularly resummed full model, whose GR limit is not the published 1PN eccentric waveform of Ref. [81]. Treat that bound as a property of the approximant, not yet a robust LISA prediction.\n\nWhat is new: previous parametrized eccentric tests were small-eccentricity phasing [68] or eccentric precessing EOB [71]. This paper does finite-eccentricity 1PN frequency-domain phasing with full LISA response and two clear prescriptions, H_R and H_F. The SPA construction, harmonic truncation checks against 1/(2ρ^2), and the Bayesian treatment with cross-harmonic interference are solid. The paper is also honest: Sec. II.C.2 explicitly says H_F's δα=0 limit is not Ref. [81]'s waveform, and Sec. IV says the bounds are forecasts within a controlled model. That is good practice.\n\nWhere it is soft. The factor-of-eight sensitivity gain of H_F over H_R is not a controlled comparison of placement alone, because the two models also differ in their GR waveforms. A cross-model injection-recovery (inject Ref. [81] or a higher-PN waveform, recover with H_F) would settle this. Without it, the flagship bound is conditional on H_F faithfully representing a real GR eccentric signal. That is a load-bearing caveat, but the authors flag it themselves. The mass notation is genuinely confusing: M appears as total mass in Sec. II and as detector-frame chirp mass in Sec. III. Minor fix. The Double Pulsar comparison is reasonable and the paper is appropriately careful about credibility levels and definitions.\n\nWho this is for: anyone building LISA eccentric waveform models or parametrized GR tests. The framework is coherent, the numerics are checked, and the placement-dependence message is worth publishing. A serious referee should engage; the paper deserves review with a request for a cross-model injection test and cleanup of the mass notation. I'd accept for review.","headline":"A careful, honest parametrized-test framework for eccentric LISA inspirals whose headline forecast is same-model; the placement-dependence result is the real contribution.","tokens_in":33517,"tokens_out":3150,"would_cite":true,"duration_ms":28785,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"LISA observations of eccentric binaries can bound a deformation of periastron precession to about $10^{-4}$, provided the deformation is placed on the dominant waveform carriers.","keywords":["parametrized tests of general relativity","eccentric binaries","LISA","periastron precession","gravitational-wave astronomy","frequency-domain waveforms","Bayesian parameter estimation"],"falsifier":"Compute the mismatch between the full model's GR limit and the published 1PN eccentric waveform for identical source parameters; a mismatch exceeding $1/(2\\rho^2)$ at a forecast SNR $\\rho$ would show the quoted sensitivity depends on the approximant. Alternatively, inject waveforms from an independent higher-order eccentric model and recover them with the full model; a bias in $\\delta\\alpha$ away from zero beyond the 90% credible interval would falsify the robustness of the forecast.","tokens_in":32489,"feed_emoji":"🛰️","tokens_out":5295,"duration_ms":51276,"temperature":0.7,"pith_summary":"This paper proposes a null test of general relativity using LISA observations of eccentric black-hole binaries, based on a single parameter $\\delta\\alpha$ that rescales the conservative periastron-precession rate while leaving the dissipative inspiral fixed. It shows that where this deformation is placed in a finitely truncated waveform changes the projected sensitivity by orders of magnitude: assigning the deformed secular phase to the dominant angular carriers yields a 90% credible bound $|\\delta\\alpha|\\lesssim 10^{-4}$ for a $3000\\,M_\\odot$, $e_0=0.5$ binary at SNR 50, comparable to the best pulsar-timing tests. Higher eccentricity sharpens the constraint by adding harmonic structure that breaks parameter degeneracies. The paper's broader point is that a phenomenological deviation parameter is not fully specified until its projection onto the waveform basis is defined.","feed_headline":"LISA can test gravity's push on periastron to 1 part in 10^4","feed_subtitle":"A single precession-deformation parameter is bounded at 90% credibility, rivaling the Double Pulsar test.","key_machinery":"The central object is the parametrized azimuthal-to-radial frequency ratio $K_\\alpha = 1 + (1+\\delta\\alpha)k_{\\rm GR}$, where $k_{\\rm GR}=3(M\\Omega_r)^{2/3}/(1-e^2)$ is the leading 1PN periastron-advance per radial cycle. This ratio is inserted into a frequency-domain eccentric inspiral waveform built from generalized Hansen coefficients and the stationary-phase approximation, with the periastron phase accumulated as $\\gamma_\\alpha(e)=\\gamma_0+(1+\\delta\\alpha)[\\gamma_{\\rm GR}(e)-\\gamma_{\\rm GR}(e_0)]$. The two waveform prescriptions differ in which angular carriers inherit the secular phase: the restricted model keeps it on the explicit 1PN sidebands, while the full model promotes it to the dominant Newtonian and half-PN carriers, producing an enlarged sideband set and coherent phase accumulation across most of the signal power.","core_discovery":"The paper claims that LISA can measure the conservative azimuthal-to-radial frequency ratio of an eccentric inspiral with enough precision to constrain a fractional deformation of the leading-order periastron advance to $|\\delta\\alpha|\\sim 10^{-4}$ or better. The deformation is introduced through $K_\\alpha = 1 + (1+\\delta\\alpha)k_{\\rm GR}$, with $\\delta\\alpha=0$ recovering general relativity, while radiation reaction and waveform amplitudes remain at their GR values. Two implementations are compared: a restricted model that deforms only the explicit 1PN precession sidebands, and a full model that also assigns the deformed secular phase to the Newtonian and half-PN carriers. The full model achieves comparable constraints at an SNR about eight times lower, because the dominant waveform components coherently accumulate the modified phase. The paper presents this placement dependence as a central methodological result and frames the eccentric LISA signal as a new laboratory for conservative-dynamics tests of gravity.","pith_inferences":["The sideband-versus-carrier distinction likely applies to other parametrized tests of gravity: quasicircular ppE-style bounds can be strongly reweighted when the same deformation is carried by dominant harmonics in eccentric templates.","In the circular limit the deformation reduces to a 1PN correction to the azimuthal phase, so the $\\delta\\alpha$ constraint could be mapped onto modified binding-energy or periastron-advance coefficients and cross-checked against quasicircular LISA tests.","A testable extension would inject signals from an independent higher-order eccentric waveform model and recover them with the full model; a recovery bias in $\\delta\\alpha$ beyond the quoted credible interval would quantify how much of the sensitivity gain is an artifact of same-model recovery.","The factor-of-eight sensitivity gap between the two prescriptions suggests that future eccentric null-test waveforms should place deformations on dominant phase carriers, but it also means quoted bounds are approximant-dependent until higher-PN and spin effects are included."],"forward_implications":["LISA can place a 90% credible bound $|\\delta\\alpha|\\lesssim 10^{-4}$ on the conservative frequency-ratio deformation for a $3000\\,M_\\odot$, $e_0=0.5$ eccentric binary at SNR 50, tightening to $|\\delta\\alpha|\\lesssim 2.6\\times 10^{-5}$ at SNR 200.","The placement of a phenomenological deformation within a finitely truncated waveform can change the apparent sensitivity by orders of magnitude, so theory-agnostic tests must specify which waveform components carry the deviation.","Increasing initial eccentricity improves the marginalized constraint at fixed SNR by redistributing power across radial harmonics and reducing degeneracies with orbital and phase parameters.","For a fixed SNR, observing the final four years before coalescence yields constraints roughly fifteen times tighter than observing an earlier, less relativistic portion of the inspiral.","The forecast bounds are comparable to the fractional precision of the Double Pulsar periastron-advance test, while probing a much more relativistic velocity regime ($v\\sim0.03$–$0.08$)."],"supporting_citations":[{"why":"Supplies the base 1PN eccentric waveform and generalized Hansen coefficients that both parametrized models deform.","marker":"[81]"},{"why":"Provides the LISA time-delay-interferometry response and Bayesian parameter-estimation machinery used for the forecasts.","marker":"[80]"},{"why":"Defines the earlier small-eccentricity parametrized test that this paper extends to finite eccentricity.","marker":"[68]"},{"why":"Provides the conservative radial-harmonic truncation prescription adopted to retain roughly 99% of signal power.","marker":"[89]"},{"why":"Supplies the LISA noise power spectral density used in the likelihood.","marker":"[88]"},{"why":"Provides the Gaussian distinguishability criterion used to define the log-likelihood threshold in the controlled comparison.","marker":"[93]"},{"why":"Gives the Double Pulsar periastron-advance measurement used as the comparison benchmark for the forecast precision.","marker":"[94]"}],"fun_headline_variants":["LISA measures precession deviation to 1e-4 in eccentric inspirals","Eccentric binaries sharpen LISA's test of gravity to 10^-4","LISA bounds periastron precession deformation at 1e-4","Eccentric LISA signals constrain GR's precession to 10^-4","LISA's eccentric waveforms test general relativity at 1e-4 precision"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The forecasts assume the template waveform used for both injection and recovery accurately represents a real general-relativistic eccentric inspiral; if actual signals differ from this finite post-Newtonian approximant, the quoted bounds and the factor-of-eight sensitivity gain are not robust.","fun_headline_variants_meta":{"raw":{"variants":["LISA measures precession deviation to 1e-4 in eccentric inspirals","Eccentric binaries sharpen LISA's test of gravity to 10^-4","LISA bounds periastron precession deformation at 1e-4","Eccentric LISA signals constrain GR's precession to 10^-4","LISA's eccentric waveforms test general relativity at 1e-4 precision"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000335,"raw_usage":{"total_tokens":1929,"prompt_tokens":1090,"completion_tokens":839,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":706,"completion_tokens_details":{"reasoning_tokens":735}},"tokens_in":706,"tokens_out":839,"duration_ms":7996,"temperature":1.0,"reasoning_tokens":735,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T04:11:23.359120+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the mismatch between the full model's GR limit and the published 1PN eccentric waveform for identical source parameters; a mismatch exceeding $1/(2\\rho^2)$ at a forecast SNR $\\rho$ would show the quoted sensitivity depends on the approximant. Alternatively, inject waveforms from an independent higher-order eccentric model and recover them with the full model; a bias in $\\delta\\alpha$ away from zero beyond the 90% credible interval would falsify the robustness of the forecast.","supporting_citations":[],"review_version":1}