REVIEW 1 major objections 5 minor 2 cited by
This paper tests general relativity with 168 binary black-hole, neutron-star, and mixed mergers, and finds no evidence for deviations from the theory.
Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →
T0 review · deepseek-v4-flash
2026-08-01 12:52 UTC pith:3MYYCF3L
load-bearing objection The new LVK GR-test catalog is the reference to cite for O4 results: the null result is mostly overdetermined, but the uncalibrated positive tensor–vector polarization Bayes factor is a real loose end the authors wave away. the 1 major comments →
GWTC-5.0: Tests of General Relativity
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
Core claim
The central claim is that, across 168 confident two-or-more-detector events, no test reveals physics beyond general relativity. After subtracting the maximum-likelihood GR waveform, residual coherent power is consistent with detector noise for every event; model-comparison Bayes factors disfavor scalar, vector, and vector–scalar polarization hypotheses; parameterized post-Newtonian deviation parameters are all consistent with GR at the 90% credible level, with combined bounds improved 1.2–2.6-fold over the previous catalog; and the remnant of each merger is consistent with a Kerr black hole in both time-domain and frequency-domain ringdown analyses. In the frequency-domain ringdown combinati
What carries the argument
The load-bearing objects are: (1) the likelihood that models detector data as stationary Gaussian noise plus a GR waveform; (2) the null-stream projector built from detector beam-pattern functions, which geometrically removes tensor-signal content and exposes any non-tensor residue; (3) the post-Newtonian phase coefficients of the inspiral waveform, into which fractional deviations are injected and then constrained; and (4) the quasi-normal-mode spectrum of the remnant black hole, described as damped sinusoids whose frequencies and damping times are fixed by the remnant mass and spin. The workhorse is the Bayesian evidence ratio between GR and modified hypotheses, with events combined both b
Load-bearing premise
The conclusion would collapse if detector noise is not well modeled by stationary Gaussian statistics or if the event-selection procedure preferentially hides deviations, because the tests would then mistake systematics for consistency.
What would settle it
Reanalyze the combined ringdown statistic after correcting for selection effects and glitch-based noise; if the GR value then falls outside the 90% credible region for the damping-time deviation, or if any single future high-SNR event yields a 90% credible interval on a post-Newtonian or ringdown deviation that excludes zero and survives injection tests, the 'no evidence for physics beyond GR' claim is falsified.
If this is right
- If general relativity is correct, the combined constraints on post-Newtonian deviations and ringdown parameters should tighten as the catalog grows beyond 168 events, with the loudest events dominating.
- The polarization results imply that any non-tensor gravitational-wave polarization, if present in nature, must be weaker than the levels excluded here; scalar and vector hypotheses are strongly disfavored.
- The ringdown results imply that black holes formed in these mergers have no measurable deviation from the predicted Kerr spectrum, and that resolving overtones in louder events could turn the present consistency check into a multi-mode black-hole spectroscopy test.
- The improved inspiral-coefficient bounds translate directly into agnostic limits on alternative theories that predict such deviations, even though the paper presents its mappings to specific theories as illustrative only.
- The mild damping-time deviation in the combined ringdown analysis predicts that either selection-corrected combination or a few more loud events will confirm it as a statistical fluctuation or reveal it as the first sign of deviation.
Where Pith is reading between the lines
- The most consequential unstated implication is that the current 'no evidence' verdict is not yet a measurement of GR's correctness but a statement that deviations, if any, are below sensitivity; a single loud merger like GW250114 can move the combined ringdown statistic by tens of percentage points in quantile, so the result depends on the luck of the loudest events.
- The treatment of GW240621 suggests a testable prediction: if the apparent high-frequency ringdown content is a non-Gaussian noise artifact, the same artifact should appear in detector noise around other loud events; a systematic glitch-morphology study could identify its cause and remove this class of spurious modes.
- Because the polarization analysis uses a single effective basis mode for many events, extending the null-stream test to a two-mode basis for the loudest three-detector events would either confirm the positive tensor–vector Bayes factor as a real preference or expose it as an approximation artifact.
- The paper's illustrative mappings to modified-gravity theories could be turned into real constraints by reweighting the agnostic posteriors with full theory-specific waveforms, which would settle whether the improved post-Newtonian bounds translate into new limits on specific alternatives.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper presents the results of seven tests of general relativity applied to the 168 confident binary signals in GWTC-5.0 (72 new O4b events, 5 newly qualified O4a events, plus the cumulative pre-O4b catalog). The tests are: residual analysis after best-fit GR waveform subtraction (Sec. 4), polarization Bayes factors for scalar/vector/tensor hypotheses (Sec. 5), parameterized post-Newtonian inspiral deviations with TIGER and FTI (Sec. 6), and three ringdown analyses (time-domain RD, frequency-domain pSEOBNR, and QNMRF) (Sec. 7). The central claim, stated in the Abstract and Sec. 8, is that no evidence for physics beyond GR is found, with improved 90% bounds on PN deviations by factors of 1.2-2.6 and improved ringdown constraints. The paper also reports two noteworthy sub-results: a positive combined tensor-vector polarization Bayes factor (log10 B_TV^T = 0.91, 90% CI excluding zero, Table 6) that the authors label inconclusive pending background studies, and a frequency-domain ringdown combined GR quantile Q2D_GR = 98.1% that places the GR prediction in the tails, attributed to finite catalog size and possible selection effects.
Significance. If the results hold, this constitutes the most comprehensive strong-field null test of GR to date, using the largest event sample yet compiled. The paper has notable strengths: seven complementary analyses with internal cross-checks; public data and reproducible analysis configuration (LVK-DCC/Zenodo); explicit injection studies that validate suspected biases in FTI (GW241011 233834) and ringdown analyses (GW240621 195059); and honest reporting of limitations, including non-stationary noise, selection effects, and the effective polarization approximation. The residual-test p-values consistent with uniformity and the improved PN bounds are particularly credible. However, the uncalibrated positive tensor-vector Bayes factor is a caveat that should be more prominently disclosed in the abstract and conclusions, as discussed in the major comments.
major comments (1)
- [Sec. 5, Table 6, Eq. (10)] The combined tensor-vector Bayes factor log10 B_TV^T = 0.91^{+0.71}_{-0.71} has a 90% credible interval that excludes zero. The paper explicitly states that 'Verifying these interpretations would require dedicated background studies' and that the result is 'inconclusive,' yet the Abstract's blanket statement 'Overall, we find no evidence for physics beyond GR' and the Conclusions' 'no significant evidence is found for additional polarization content' do not mention this uncalibrated positive value. This is a load-bearing point: the no-evidence claim for polarizations rests on treating a moderate Bayes factor as noise without a null-hypothesis calibration. The authors should either (i) perform a background/injection study to calibrate the distribution of the combined Bayes factor under Gaussian noise, or (ii) explicitly state in the Abstract and Conclusions that the polarization test yiel
minor comments (5)
- [Sec. 7.2, par. 7] The pSEOBNR hierarchical analysis does not account for selection effects, and the paper notes these would shift δτ̂220 toward lower values, reducing the apparent inconsistency with GR. This is acknowledged, but the quoted combined GR quantile (Q2D_GR = 98.1%) is not accompanied by a discussion of how the selection-effect correction would change it. A brief quantitative estimate or a more explicit statement that the quoted quantile is a selection-biased upper bound would improve clarity.
- [Sec. 2, Tables 2-3] The event selection tables would benefit from specifying which events have data-quality issues or glitches that required special handling (e.g., GW241127 061008 with the Livingston glitch). This information appears in later sections but is not collated in the main event tables.
- [Sec. 8, par. 4] Typo: 'GW250207 11564' should be 'GW250207 115645'.
- [Sec. 4, Fig. 1] The caption states 'Solid (empty) circles represent O4b (pre-O4b) events', but the text in the same paragraph says 'pre-O4b events' are from earlier runs. Clarify the distinction between O4b and pre-O4b for consistency.
- [Sec. 6, Table 7] The illustrative theory mappings in Table 7 are useful, but the table caption could more prominently state that the dCS entry is omitted because the small-coupling condition fails; currently this is only in the text. A footnote in the table would help.
Circularity Check
No significant circularity: deviation parameters are inferred objects and GR is an external benchmark; stated limitations are honest, not definitional.
full rationale
The paper's central claims are null tests whose 'predictions' are external benchmarks, not fitted values in disguise. For the residual test, the GR template is subtracted and the residual consistency with noise is assessed against an independent noise-background distribution (Sec. 4, Eq. 3-6); the conclusion is a statistical consistency claim, not a quantity derived from the fitted waveform. For TIGER/FTI, the parameters δφ̂_i are the objects of Bayesian inference and the GR value (δ=0, with GR PN coefficients) is fixed from outside the fit (Sec. 6, Eq. 12). The hierarchical combination treats these as measurements of common or population-level deviation parameters; nothing is 'predicted' that was used to define the model. For the ringdown tests, δf̂_220 and δτ̂_220 are inferred deviations, and the GR prediction (Kerr QNM frequency/damping time) is computed from remnant mass and spin via standard fitting formulas and NR-informed mappings (Sec. 7.2, Eqs. 17-18) — an external benchmark, not the fitted values. The quoted GR quantiles (Q2D_GR = 98.1%) measure where this external GR value falls in the inferred posterior; this is a consistency statistic, not a circular redefinition. The theory mappings in Tables 7 and 10 are explicitly labeled 'illustrative translations' and 'should not be interpreted as statistical constraints,' so no circularity burden is taken there. The positive tensor-vector Bayes factor (log10 B_TV^T = 0.91) is honestly reported as inconclusive, with the paper explicitly noting that 'verifying these interpretations would require dedicated background studies' and that non-Gaussian noise may bias model comparison; this is a stated limitation, and the paper declines to elevate it to evidence — a correctness or calibration concern, not a circularity. Self-citations (GWTC-4.0 papers, catalog papers, waveform-model papers) are normal references to prior data products and software; the load-bearing derivations here do not reduce to those citations. No step in the derivation chain is equivalent, by construction, to its own input.
Axiom & Free-Parameter Ledger
free parameters (6)
- PN deviation parameters δφi (FTI/TIGER) =
90% upper bounds |δφk| ≤ 6.0e-4–0.9 (FTI), 4.4e-4–0.8 (TIGER); all consistent with 0
- Ringdown fractional QNM deviations δf220, δτ220 (pSEOBNR) =
δf220 = 0.01 (+0.03/−0.03); δτ220 = 0.09 (+0.13/−0.10), hierarchical (Eq. 22)
- Hierarchical hyperparameters (μ, σ, ρ) =
μδf = 0.01(+0.02/−0.02), σδf < 0.03, μδτ = 0.09(+0.07/−0.06), σδτ < 0.10, ρ = −0.07(+0.76/−0.69)
- Polarization effective scaling factors C =
amplitudes uniform [0,1], phases uniform [0,2π]
- QNM mode amplitudes A_lmn (ringdown) =
A221 = 4.68(+7.84/−3.63)e-21 at t>=2 tMf for GW240621_195059; A440 = 1.37(+6.97/−1.21)e-21
- Inspiral/ringdown selection thresholds =
FTI inspiral SNR ≥ 20; TIGER SNR ≥ 15; pSEOBNR SNR ≥ 8 in both parts; ringdown ρpp ≥ 10; polarization Δf = 16 Hz
axioms (6)
- domain assumption Detector noise is stationary and Gaussian; d(t) = h(t) + n(t)
- domain assumption GR waveform models (IMRPhenomXPHM SpinTaylor, SEOBNRv5HM ROM, NRSur7dq4) faithfully represent GR signals
- domain assumption The ringdown is a superposition of Kerr quasinormal modes with constant amplitudes from a suitable start time
- ad hoc to paper Non-tensor polarization modes are complex-proportional to the basis mode (effective polarization approximation)
- domain assumption ΛCDM cosmology and catalog PE priors (Abac et al. 2026h,j)
- standard math Degeneracies fix which PN coefficients are testable: 2.5PN non-log term degenerate with reference phase; −1PN and 0.5PN terms vanish in GR
read the original abstract
The signals from the LIGO-Virgo-KAGRA network of gravitational-wave (GW) detectors allow us to perform sensitive tests of general relativity (GR) in the dynamical and strong-field regime of gravity. We present the results of seven tests of GR using the observed binary signals in the fifth GW Transient Catalog (GWTC-5.0), i.e., up to and including the second part of the fourth observing run (O4b). We restrict our analysis to the confident signals, henceforth called events, observed by at least two detectors that have estimated false alarm rates $\le 10^{-3} \ \rm{yr}^{-1}$. These include 72 events from O4b and five events from the first part of the fourth observing run that are now analyzed due to their increased significance from updated search results, bringing the total number of events for tests of GR in the cumulative GWTC to 168. After subtracting the best-fit waveforms, we find the residuals are consistent with detector noise for all events considered. We also find no strong evidence for additional polarizations beyond those predicted by GR. We perform tests of GW generation, improving the constraints on deviations from the GR post-Newtonian coefficients by factors of 1.2-2.6. Finally, we find overall consistency of the remnants with GR using both time- and frequency-domain methods. For GW240621_195059, postmerger data are consistent with the dominant quadrupolar ($\ell=|m|=2$) mode of a Kerr black hole and its first overtone, with spurious high-frequency content preventing a spectroscopic constraint of GR. In the frequency-domain ringdown analysis, the GR prediction lies in the tails of the combined results, possibly due to the limited catalog size. However, the combined results indicate improved consistency with GR over GWTC-4.0, owing to the contribution of GW250114 with a network matched-filter signal-to-noise ratio of 76.9. Overall, we find no evidence for physics beyond GR.
Figures
Forward citations
Cited by 2 Pith papers
-
Towards long and accurate numerical relativity waveforms of binary black holes beyond general relativity
Spectral methods plus comoving fixing-the-equations drivers yield 40+ cycle equal-mass sGB binary waveforms with phase error ≲1 rad, distinguishable from GR and merging earlier.
-
Constraining deviations from the Teukolsky equation with GW250114
GW250114's fundamental ringdown mode bounds theory-agnostic deviations from the Teukolsky equation to be consistent with zero at characteristic scales of 60-100 km.
Reference graph
Works this paper leans on
-
[1]
Aasi, J., Abbott, B. P., Abbott, R., et al. 2015, CQGra, 32, 074001, doi: 10.1088/0264-9381/32/7/074001 Abac, A. G., Abbott, R., Abouelfettouh, I., et al. 2024, ApJL, 970, L34, doi: 10.3847/2041-8213/ad5beb Abac, A. G., Abouelfettouh, I., Acernese, F., et al. 2025a, ApJL, 995, L18, doi: 10.3847/2041-8213/ae0c06 —. 2025b, ApJL, 993, L21, doi: 10.3847/2041-...
Pith/arXiv arXiv 2015
-
[4]
https://arxiv.org/abs/2511.19593 Chung, A. K.-W., Lam, K. K.-H., & Yunes, N. 2025, PhRvD, 111, 124052, doi: 10.1103/g83n-rrlj Chung, A. K.-W., & Yunes, N. 2024, PhRvL, 133, 181401, doi: 10.1103/PhysRevLett.133.181401 —
arXiv 2025
-
[5]
https://arxiv.org/abs/2506.14695 Colleoni, M., Vidal, F. A. R., Garc´ıa-Quir´os, C., Akc ¸ay, S., & Bera, S. 2025, PhRvD, 111, 104019, doi: 10.1103/PhysRevD.111.104019 Cornish, N., Sampson, L., Yunes, N., & Pretorius, F. 2011, PhRvD, 84, 062003, doi: 10.1103/PhysRevD.84.062003 Cornish, N. J., & Littenberg, T. B. 2015, CQGra, 32, 135012, doi: 10.1088/0264-...
Pith/arXiv arXiv 2025
-
[7]
1989, PhRvD, 40, 3884, doi: 10.1103/PhysRevD.40.3884 Haegel, L., O’Neal-Ault, K., Bailey, Q
https://arxiv.org/abs/2511.11886 G¨ursel, Y ., & Tinto, M. 1989, PhRvD, 40, 3884, doi: 10.1103/PhysRevD.40.3884 Haegel, L., O’Neal-Ault, K., Bailey, Q. G., et al. 2023, PhRvD, 107, 064031, doi: 10.1103/PhysRevD.107.064031 Hamilton, E., London, L., & Hannam, M. 2023, PhRvD, 107, 104035, doi: 10.1103/PhysRevD.107.104035 Hanna, C., Caudill, S., Messick, C., ...
Pith/arXiv arXiv 1989
-
[9]
https://arxiv.org/abs/2605.15271 32 Kidder, L. E., Will, C. M., & Wiseman, A. G. 1993, PhRvD, D47, 4183, doi: 10.1103/PhysRevD.47.R4183 Klimenko, S., Vedovato, G., Drago, M., et al. 2016, PhRvD, 93, 042004, doi: 10.1103/PhysRevD.93.042004 Klimenko, S., Yakushin, I., Rakhmanov, M., & Mitselmakher, G. 2004, CQGra, 21, S1685, doi: 10.1088/0264-9381/21/20/011...
Pith/arXiv arXiv 1993
-
[10]
2026, QNMRF software, https://github.com/Sizheng-Ma/qnm filter, GitHub Ma, S., Sun, L., & Chen, Y
https://arxiv.org/abs/2604.08680 Ma, S., Lu, N., Finch, E., et al. 2026, QNMRF software, https://github.com/Sizheng-Ma/qnm filter, GitHub Ma, S., Sun, L., & Chen, Y . 2023a, PhRvD, 107, 084010, doi: 10.1103/PhysRevD.107.084010 —. 2023b, PhRvL, 130, 141401, doi: 10.1103/PhysRevLett.130.141401 Ma, S., & Yang, H. 2024, PhRvD, 109, 104070, doi: 10.1103/PhysRe...
Pith/arXiv arXiv 2026
-
[12]
https://arxiv.org/abs/2105.09485 Yagi, K., Stein, L. C., Yunes, N., & Tanaka, T. 2012a, PhRvD, 85, 064022, doi: 10.1103/PhysRevD.85.064022 34 Yagi, K., Yunes, N., & Tanaka, T. 2012b, PhRvL, 109, 251105, doi: 10.1103/PhysRevLett.116.169902 Yunes, N., & Hughes, S. A. 2010, PhRvD, 82, 082002, doi: 10.1103/PhysRevD.82.082002 Yunes, N., & Pretorius, F. 2009, P...
Pith/arXiv arXiv 2010
-
[2014]
https://arxiv.org/abs/1409.4431 Brito, R., Buonanno, A., & Raymond, V . 2018, PhRvD, 98, 084038, doi: 10.1103/PhysRevD.98.084038 Bucciotti, B., Juliano, L., Kuntz, A., & Trincherini, E. 2024, PhRvD, 110, 104048, doi: 10.1103/PhysRevD.110.104048 Bucciotti, B., Kuntz, A., Serra, F., & Trincherini, E. 2023, JHEP, 12, 048, doi: 10.1007/JHEP12(2023)048 Buonann...
Pith/arXiv arXiv 2018
-
[2019]
https://arxiv.org/abs/1901.08580 Saleem, M., Datta, S., Arun, K. G., & Sathyaprakash, B. S. 2022, PhRvD, 105, 084062, doi: 10.1103/PhysRevD.105.084062 S¨anger, E. M., Roy, S., Agathos, M., et al. 2026, PhRvD, 113, 084070, doi: 10.1103/r43k-51yq Sberna, L., Bosch, P., East, W. E., Green, S. R., & Lehner, L. 2022, PhRvD, 105, 064046, doi: 10.1103/PhysRevD.1...
Pith/arXiv arXiv 1901
-
[2021]
2023, PhRvL, 131, 169001, doi: 10.1103/PhysRevLett.131.169001 —
https://arxiv.org/abs/2107.05609 —. 2023, PhRvL, 131, 169001, doi: 10.1103/PhysRevLett.131.169001 —. 2024,ringdownpackage, 1.0.0, ringdown.readthedocs.io, Zenodo, doi: 10.5281/zenodo.5094067 Isi, M., Giesler, M., Farr, W. M., Scheel, M. A., & Teukolsky, S. A. 2019b, PhRvL, 123, 111102, doi: 10.1103/PhysRevLett.123.111102 Isi, M., & Weinstein, A. J. 2017, ...
Pith/arXiv arXiv 2023
-
[2025]
https://arxiv.org/abs/2512.04593 Chandrasekhar, S. 1983, The mathematical theory of black holes (Oxford, UK: Oxford University Press) Chavda, A., Lagos, M., & Hui, L. 2025, JCAP, 07, 084, doi: 10.1088/1475-7516/2025/07/084 Cheung, M. H.-Y ., Baibhav, V ., Berti, E., et al. 2023, PhRvL, 130, 081401, doi: 10.1103/PhysRevLett.130.081401 Chiaramello, D., Cibr...
arXiv 1983
-
[2026]
2026, PhRvD, 113, 024048, doi: 10.1103/bgjv-lvyd Detweiler, S
https://arxiv.org/abs/2605.16183 De Amicis, M., Cannizzaro, E., Carullo, G., & Sberna, L. 2026, PhRvD, 113, 024048, doi: 10.1103/bgjv-lvyd Detweiler, S. L. 1980, ApJ, 239, 292, doi: 10.1086/158109 Dhani, A., V¨olkel, S. H., Buonanno, A., et al. 2025, PhRvX, 15, 031036, doi: 10.1103/5pks-qz6b Di Renzo, F., Fidecaro, F., Razzano, M., & Sorrentino, N. 2024, ...
Pith/arXiv arXiv 2026
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.