{"id":"f04c4b44-cdfd-4890-aefa-03c791e33112","arxiv_id":"2505.08409","paper_version":2,"verdict":"REJECT","confidence":"HIGH","novelty_score":4.0,"correctness_risk":"high","formal_verification":"none","parameter_count":12,"one_line_summary":"Using quasi-periodic oscillation data under the relativistic precession model, the paper reports 68% constraints on nine modified Kerr spacetimes, with only Kerr-Newman keeping a Kerr-compatible parameter range.","lead":"The authors fit X-ray oscillation data from three black hole binaries to ten competing descriptions of a spinning black hole and report that eight of the nine modified models prefer a non-Kerr geometry at 68% confidence. The paper matters as a check on whether astrophysical black holes are truly described by Einstein's Kerr solution, though the paper's own model comparison finds Kerr still optimal.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claimed 68% exclusion of Kerr is internally inconsistent: for every modified model, χ²(Δ*=0) − χ²_min < 1, so a profile-likelihood 68% interval should contain zero.","rationale":"The reader's verdict is REJECT with high confidence, and I agree that the central claim fails. However, the reader's weakest_assumption focuses on the relativistic precession mapping of QPO frequencies, while the decisive problem is internal to the paper's own χ² statistics. The reader's rationale notes that Eq. (27) has as many fitted parameters as data points, which is correct, but that concern is about goodness-of-fit and model comparison. The sharper issue is the nested-model inconsistency: since each modified metric contains Kerr at Δ*=0, the profiled χ² at Δ*=0 is exactly the reported Kerr χ²_min = 0.976. The differences between this value and each modified model's χ²_min are all below the 68% threshold of Δχ² = 1. Therefore, under standard profile-likelihood confidence regions, zero must be included in the 68% interval for every model, directly contradicting Table II's positive-definite intervals and the abstract's claim that the Kerr solution is statistically deviant. This is not a disagreement with the external consensus; it is an internal inconsistency between the paper's χ² table and its headline inference. The Bayes-factor and AIC sections independently show no significant preference for the modified models, with Kerr remaining the AIC-best model. Even if one were to reinterpret the reported intervals as Bayesian credible intervals, the Bayes factors with |ln R| < 0.11 would not support the language of 'demonstrating statistical deviation.' For these reasons, the appropriate verdict remains REJECT: the load-bearing conclusion is unsupported, and the paper's own numbers provide a direct test that would settle the issue.","tokens_in":32528,"tokens_out":6158,"duration_ms":68143,"concrete_test":"For each of the nine modified models, set Δ*=0 and minimize Eq. (27) with respect to all spins a*_p, masses M*_p, and the four orbital radii, i.e. profile over every nuisance parameter. Record the resulting χ²_0 and compare Δχ² = χ²_0 − χ²_min(model) with the threshold Δχ² = 1 (68% CL for one parameter). Because each model reduces to Kerr at Δ*=0, χ²_0 should equal the Kerr χ²_min = 0.976 if the implementation is consistent. If Δχ² < 1 for all models, then Table II's positive-definite 68% intervals are invalid and the conclusion of Kerr exclusion collapses. This check uses only the published frequency expressions in Appendix A and Table II, and requires no new observational data.","verdict_should_be":"REJECT","load_bearing_attack":"The central claim is that eight of nine modified spacetimes exclude the Kerr limit at 68% CL because their fitted modification parameters are positive-definite (Table II and Section VI). This claim fails even granting the relativistic precession model and the saturated 11-parameter fit of Eq. (27). Each modified metric reduces to Kerr at Δ*=0, so the minimized χ² for the modified model with Δ* fixed to zero is exactly the Kerr model's χ²_min = 0.976 reported in Table II. For a single parameter of interest, the 68% profile-likelihood confidence interval is the set of Δ* with χ²(Δ*) − χ²_min ≤ 1 (or ≤ 2.30 for a 2D region). Using the paper's own Table II values, the differences χ²_min(Kerr) − χ²_min(model) are: Bardeen 0.404, ABG 0.548, Hayward 0.534, KN 0.547, KTN 0.552, BK 0.479, Kerr-MOG 0.496, Kerr-Sen 0.277, PFDM 0.264. Every one of these is below 1, so Δ*=0 should lie inside the 68% confidence region for every model. The positive-definite intervals reported in Table II are therefore not standard profile-likelihood confidence intervals; they appear to have been obtained by slicing parameter space at the modified best-fit values of the other parameters, rather than profiling over all nuisance parameters, or by a Bayesian credible interval that is then in direct tension with the paper's own Bayes factors (all |ln R| < 0.11) and AIC analysis, which keeps Kerr as the best model. The headline inference that QPO data 'demonstrate statistical deviation of the Kerr solution' is thus unsupported by the paper's own statistics.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper derives test-particle orbital and epicyclic frequencies in Kerr and nine single-parameter modified Kerr spacetimes (Bardeen, ABG, Hayward, KN, KTN, BK, Kerr-MOG, Kerr-Sen, PFDM), then uses the relativistic precession model and QPO measurements from three microquasars to fit each spacetime through the chi-squared function in Eq. (27). It reports 68% confidence intervals for the modification parameter of each model, the spins and masses of the three sources, and the four orbital radii. The central conclusion, repeated in the abstract and Section VI, is that all modified spacetimes except Kerr-Newman require a positive-definite modification parameter at 68% CL, 'demonstrating statistical deviation of the Kerr solution.' A Bayes-factor and AIC comparison is also performed; the AIC actually identifies Kerr as the best model, and all Bayes factors are inconclusive.","tokens_in":33001,"tokens_out":5280,"duration_ms":51868,"significance":"If the 68% exclusion claim were correct, it would be a significant observational challenge to the Kerr paradigm and a useful demonstration that QPO data can discriminate among black hole spacetimes. The paper also provides a convenient compilation of epicyclic-frequency formulas in Appendix A and a transparent statement of the data and fitting function, which are useful for follow-up work. However, the central inference is not reliable: the fit has zero degrees of freedom, and the claimed 68% intervals are inconsistent with the profile-likelihood values implied by the paper's own Table II and with its own model-selection results. The significance of the paper therefore rests on a statistical artifact rather than on evidence.","major_comments":[{"comment":"The chi-squared function in Eq. (27) contains 11 measured frequencies in Table I but also 11 fitted parameters: the modification parameter Delta*, the three spins a*_p, the three masses M*_p, and the four radii r1, r1', r2, r3. The fit therefore has zero degrees of freedom, and the reported chi2_min values between 0.4 and 1.0 are a measure of the model's flexibility, not of its predictive success. In this regime the 68% intervals in Table II and the associated statements of 'stringent constraints' in Section IV are not statistically meaningful; a saturated fit cannot yield valid confidence intervals through the conventional asymptotic arguments used here.","section":"IV, Eq. (27), Table I"},{"comment":"For every modified metric, Delta*=0 reduces the model exactly to Kerr. Hence the minimum of chi2(Delta*) over the remaining parameters at Delta*=0 is the Kerr value chi2_min = 0.976 in Table II. Using the paper's own numbers, the improvement over Kerr is 0.404 (Bardeen), 0.548 (ABG), 0.534 (Hayward), 0.547 (KN), 0.552 (KTN), 0.479 (BK), 0.496 (Kerr-MOG), 0.277 (Kerr-Sen), and 0.264 (PFDM). For a single parameter of interest, the 68% profile-likelihood interval is defined by Delta-chi2 <= 1, so Delta*=0 lies inside the 68% interval for every model. The positive-definite intervals reported in Table II therefore cannot be profile-likelihood intervals; they appear to be conditional slices taken at the best-fit values of the nuisance parameters. The conclusion in the abstract and Section VI that eight spacetimes 'mandate positive-definite parameters at 68% CL' is not supported and is in direct tension with the Bayes factors (all |ln R| < 0.11 in Table IV) and the AIC analysis in Table V, which keeps Kerr as the best model.","section":"Table II and Section VI"},{"comment":"The analysis assumes the relativistic precession mapping nu_u = nu_phi, nu_l = nu_phi + nu_r, and nu_c = nu_phi - nu_theta exactly, while the four radii are free nuisance parameters. Under this assumption the fitted Delta* values are only interpretable as spacetime parameters; if the QPO peaks have a different physical origin, or if the frequencies do not correspond to equatorial circular test-particle orbits, every reported modification-parameter constraint loses its geometric meaning. Given the ongoing debate on QPO mechanisms, the paper should at least demonstrate robustness to alternative QPO models or explicitly frame all constraints as conditional on the RP model. This limitation is load-bearing for the abstract's physical claim, but the paper does not address it.","section":"Section IV before Eq. (27)"}],"minor_comments":[{"comment":"The heading 'Model camparison' should be 'Model comparison'.","section":"Section V title"},{"comment":"The heading 'Rotating BHs modiffed by dark matter' should be 'Rotating BHs modified by dark matter'.","section":"Section II.D title"},{"comment":"The label 'Badeen' in the first panel should be 'Bardeen'.","section":"Figure 1"},{"comment":"The text 'GR0 J1655-40' should be 'GRO J1655-40'.","section":"Section IV, discussion near [143]"},{"comment":"The best-fit radii are listed without uncertainties, so the reader cannot assess whether the four radii are pinned by the data or merely absorbed by the saturated fit.","section":"Table III"},{"comment":"The 'marginal likelihood' values in Table II are not reproducible because the priors used in Eq. (29) are not specified; the Bayes factors in Table IV depend on those priors.","section":"Section V.A, Eq. (29)"}],"recommendation":"reject","confidential_remarks":"The central statistical claim is internally contradicted by the paper's own chi-square values and model-selection tables; this is not a matter of presentation or style. The zero-degree-of-freedom fit and the non-profiled confidence intervals would need to be replaced, and doing so would almost certainly remove the 'statistical deviation of Kerr' headline. I see no way to preserve the main conclusion within the current data set and fitting setup."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a carefully built catalog of epicyclic frequencies for Kerr and nine single-parameter generalizations, fitted to four QPO datasets, and the comparison tables are honest. But the headline claim—that eight of nine models exclude Kerr at 68% CL—is contradicted by the paper's own chi-square values. That claim should not survive review.\n\nWhat the paper does well: the Appendix works out explicit omega_phi, omega_r, omega_theta expressions for all ten spacetimes; that is tedious and useful. The simultaneous fit of ten models to the same QPO data with a consistent relativistic-precession mapping is a reasonable way to compare models. The spin and mass results for Kerr line up with earlier QPO-based estimates, and the cited comparisons to Bardeen, KN, KTN, BK, MOG, and PFDM constraints are real. The AIC and Bayes tables are reported straightforwardly, and the AIC column actually keeps Kerr as the best model.\n\nThe soft spot is load-bearing. Equation (27) has 11 fitted parameters for 11 data points, so the reported chi-square-minimum values are not evidence of fit quality. More decisively, every modified model reduces to Kerr at the modification parameter Delta*=0. The paper's own Table II gives chi2_min(Kerr)=0.976 and chi2_min(model) between 0.424 and 0.712. For a single parameter of interest, a 68% profile-likelihood interval is the set of Delta* with chi2(Delta*)-chi2_min <= 1. Since chi2(Delta*=0)=0.976, the difference for every model is below 1, so zero lies inside the 68% region for every model. The positive-definite intervals in Table II are not standard profile-likelihood intervals; they look like slices with other parameters fixed at the modified best fit. The abstract's \"statistical deviation of the Kerr solution\" is unsupported by the paper's own numbers and is also in tension with its own Bayesian and AIC results, all of which are inconclusive.\n\nThe RP-model assumption is another limitation, though the authors are upfront about it. If the QPO peaks come from a different mechanism or from non-circular orbits, the fitted parameters lose their geometric meaning.\n\nBottom line: the audience is QPO phenomenology and modified-gravity constraints, and the frequency catalog will be useful. But the paper deserves a serious referee, not publication in this form. I would send it to peer review rather than desk reject, with the expectation of major revision: replace the interval construction with a proper profile likelihood, state the degree-of-freedom situation honestly, and soften the conclusion to \"no significant preference.\" I would not cite the Kerr-exclusion claim.","headline":"A useful catalog of epicyclic frequencies and an honest model-comparison table, but the headline Kerr-exclusion claim is contradicted by the paper's own chi-square values.","tokens_in":33599,"tokens_out":4276,"would_cite":false,"duration_ms":43928,"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":"QPO data from three microquasars exclude the Kerr limit at 68% confidence for eight of the nine single-parameter modified Kerr spacetimes tested.","keywords":["quasi-periodic oscillations","microquasars","black holes","modified Kerr spacetimes","relativistic precession model","Kerr-Newman","epicyclic frequencies"],"falsifier":"Inject a known Kerr spacetime with the same masses, spins, and noise levels as the real data, generate four QPO frequency sets under the RP model, and run the paper's $\\chi^2$ pipeline. If the best-fit modification parameters for the eight non-KN geometries routinely land away from zero at 68% confidence when the input geometry is exactly Kerr, then the reported exclusion is an artifact of the fitting procedure; if they center on zero, the observed exclusions carry real geometric information.","tokens_in":32292,"feed_emoji":"🕳️","tokens_out":10307,"duration_ms":95981,"temperature":0.7,"pith_summary":"The paper asks whether the spacetime around stellar-mass black holes is exactly Kerr or one of nine single-parameter modifications, using the quasi-periodic oscillation (QPO) frequencies of three microquasars as the observable. Under the relativistic precession model, each observed peak is mapped to a combination of the orbital and epicyclic frequencies of a test particle, and a $\\chi^2$ fit constrains the modification parameter, spin, and mass of each geometry. Its central finding is that only the Kerr-Newman charge interval contains zero, while the fitted intervals for the other eight modified spacetimes are positive at 68% confidence, which the authors read as a statistical deviation of the Kerr solution. The conclusion is qualified by model selection: Bayes factors mildly prefer five modified geometries, while the Akaike criterion keeps Kerr as the best model.","feed_headline":"Eight modified Kerr spacetimes exclude the Kerr limit at 68% CL","feed_subtitle":"Microquasar QPO fits keep only Kerr-Newman compatible with zero at 68% confidence.","key_machinery":"The load-bearing object is the relativistic precession model of QPOs, supplied by reference [71] and encoded in Eq. (27). For each of the ten metrics, the paper computes the orbital frequency $\\omega_\\phi$ and the radial and latitudinal epicyclic frequencies $\\omega_r$, $\\omega_\\theta$ from linearized geodesic deviations on equatorial circular orbits; under the RP mapping these become functions of the black hole mass, spin, modification parameter, and the orbital radius of the emitting ring. A single global $\\chi^2$ compares four observed QPO frequency sets with these model predictions, leaving the four orbital radii as free nuisance parameters. The modification parameters enter through the metric functions, such as the $\\Delta$ factors and mass functions that differ from Kerr, so the fit simultaneously constrains the geometry and the microquasar properties.","core_discovery":"On the paper's own terms, the discovery is that QPO data, interpreted through the relativistic precession mapping $\\nu_u = \\nu_\\phi$, $\\nu_l = \\nu_\\phi + \\nu_r$, and $\\nu_c = \\nu_\\phi - \\nu_\\theta$, exclude the Kerr limit for eight of the nine tested single-parameter deviations. The fitted 68% values are $b^* = 0.229^{+0.045}_{-0.034}$ for Bardeen, $Q_1^* = 0.360^{+0.012}_{-0.011}$ for ABG, $l^* = 0.295 \\pm 0.019$ for Hayward, $n^* = 0.257^{+0.054}_{-0.025}$ for Kerr-Taub-NUT, $q^* = 0.152 \\pm 0.019$ for the braneworld Kerr, $\\alpha^* = 0.795 \\pm 0.011$ for Kerr-MOG, $Q_3^* = 0.361^{+0.028}_{-0.024}$ for Kerr-Sen, and $k^* = 0.019^{+0.002}_{-0.003}$ for the perfect-fluid dark matter geometry, all strictly positive. Only the Kerr-Newman charge, $Q_2^* = -0.004 \\pm 0.078$, spans negative and positive values and therefore contains the Kerr case at zero. The same fits yield masses for the three microquasars that are mostly consistent with optical/near-infrared dynamical measurements, while the inferred spins are systematically lower than iron-line/continuum-fitting values, a discrepancy the paper leaves open.","pith_inferences":["Beyond the paper: the shared positive offset across eight unrelated geometries suggests the RP model's radial-frequency prediction may carry a systematic bias, so the geometric conclusion should be tested by injecting synthetic Kerr data into the same pipeline before treating it as physical.","Beyond the paper: the paper quotes all exclusions at the 68% level; the 95% contours already plotted in Figures 3 and 4 would show how many of the eight exclusions survive a stricter threshold.","Beyond the paper: because the four orbital radii are free nuisance parameters, the model has built-in flexibility; adding a prior on the radii or fitting more QPO sets per source could either sharpen or dissolve the positive-parameter pattern."],"forward_implications":["If the RP interpretation is right, microquasar QPO data currently favor a non-Kerr geometry in eight of nine tested one-parameter extensions, with Kerr-Newman as the sole exception that keeps the Kerr limit inside its 68% interval.","The fitted masses for the three microquasars agree with independent dynamical measurements in most models, which supports the internal consistency of the radii and frequency assignments used in the fit.","The systematic gap between QPO-inferred spins and iron-line/continuum-fitting spins indicates that at least one of these spin diagnostics is model-dependent; the paper notes this could mean the continuum method overestimates spin or the QPO model is incomplete.","Bayes factors and AIC give opposite rankings, with a slight preference for five modified models in the Bayes analysis but Kerr as the best model in the AIC analysis, so current QPO data cannot decisively choose among the ten geometries."],"supporting_citations":[{"why":"Defines the relativistic precession mapping that translates observed QPO peaks into the orbital and epicyclic frequencies used in the $\\chi^2$ fit.","marker":"[71]"},{"why":"Source of the four QPO frequency sets from the three microquasars in Table I; these are the data being fitted.","marker":"[127–134]"},{"why":"Cited to justify that the RP model is the only one currently able to explain both HFQPO and LFQPO simultaneously, so the model choice is not arbitrary.","marker":"[60]"},{"why":"Earlier QPO/RP constraint on the Bardeen parameter, used as a consistency check that the fitted value $b^*$ is plausible.","marker":"[83]"},{"why":"Prior gravitational-wave-based constraint on the Kerr-Newman charge, used as a consistency check for the only modification parameter that remains compatible with zero.","marker":"[142]"},{"why":"Prior QPO/RP analysis of GRO J1655-40 in Kerr spacetime, used to benchmark the spin-mass constraint and to frame the discrepancy with continuum-fitting spins.","marker":"[147]"}],"fun_headline_variants":["Eight modified Kerr models fail to include zero parameter at 68%","Only Kerr-Newman accommodates zero modification in QPO data","QPO observations rule out Kerr limit for eight spacetime variants","Eight of nine Kerr tweaks contradict pure Kerr at 68% confidence","Microquasar QPO fits favor modified Kerr except for Kerr-Newman"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing assumption is that each observed QPO peak is produced by the relativistic precession of a test particle on an equatorial circular orbit, with the specific identifications $\\nu_u=\\nu_\\phi$, $\\nu_l=\\nu_\\phi+\\nu_r$, and $\\nu_c=\\nu_\\phi-\\nu_\\theta$; if the peaks have a different physical origin, or the orbits are not equatorial circular test-particle orbits, the fitted modification parameters lose their geometric meaning.","fun_headline_variants_meta":{"raw":{"variants":["Eight modified Kerr models fail to include zero parameter at 68%","Only Kerr-Newman accommodates zero modification in QPO data","QPO observations rule out Kerr limit for eight spacetime variants","Eight of nine Kerr tweaks contradict pure Kerr at 68% confidence","Microquasar QPO fits favor modified Kerr except for Kerr-Newman"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000405,"raw_usage":{"total_tokens":2264,"prompt_tokens":1261,"completion_tokens":1003,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":877,"completion_tokens_details":{"reasoning_tokens":911}},"tokens_in":877,"tokens_out":1003,"duration_ms":9076,"temperature":1.0,"reasoning_tokens":911,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T21:56:24.456376+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Inject a known Kerr spacetime with the same masses, spins, and noise levels as the real data, generate four QPO frequency sets under the RP model, and run the paper's $\\chi^2$ pipeline. If the best-fit modification parameters for the eight non-KN geometries routinely land away from zero at 68% confidence when the input geometry is exactly Kerr, then the reported exclusion is an artifact of the fitting procedure; if they center on zero, the observed exclusions carry real geometric information.","supporting_citations":[{"cited_title":"B Chen, Z","cited_arxiv_id":null,"evidence_quote":"Prior gravitational-wave-based constraint on the Kerr-Newman charge, used as a consistency check for the only modification parameter that remains compatible with zero."},{"cited_title":"Anjum, M","cited_arxiv_id":null,"evidence_quote":"Prior QPO/RP analysis of GRO J1655-40 in Kerr spacetime, used to benchmark the spin-mass constraint and to frame the discrepancy with continuum-fitting spins."}],"review_version":1}