{"id":"7edaa924-85b7-4702-95c2-72f740e9c089","arxiv_id":"2411.10066","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":16,"one_line_summary":"The timing anomaly of PSR J0337+1715 is achromatic and is equally well fit by a roughly Moon-mass circum-ternary planet or by exceptionally strong red noise, giving model-dependent limits on strong-equivalence-principle violation.","lead":"A strange four-microsecond wobble in the pulses of a pulsar orbiting two white dwarfs may be caused by a tiny planet, about half the mass of the Moon, circling all three stars, or by an unusually strong red noise inside the pulsar. The two explanations fit today's data equally well, and future observations could distinguish them, refining tests of a fundamental law of gravity.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Planet-specific claims depend on the mutual-interaction term at the outer-binary period (Eq. B.22), whose Δχ² ≈ 1.6 over Kepler for two extra parameters is not a significant detection; with PΠ = 3310 d exceeding the 2988 d span, the Planet model is degenerate with PL3 red noise.","rationale":"The paper is honest and technically careful: it rules out chromatic DM variations, releases data and code, performs MCMC with convergence diagnostics, and repeatedly acknowledges the marginality of the planet model. Credit is also due to the stability analysis and the kick estimate, which are presented conditionally on the planet hypothesis. The central model-independent findings — that the low-frequency signal is achromatic, that PL3 red noise and a Keplerian planet fit comparably, and that the SEP limits are model-dependent — are supported. The fragile part is precisely the planet-specific interpretation. The only discriminator is the mutual-interaction term of Eq. (B.22), and the reported statistics do not rise to a detection: Δχ² ≈ 1.6 for two extra parameters, no quoted p-value or Bayes factor, and explicit statements in Sections 4.2 and Appendix C that the relevant frequency shows no clear signature and that likelihood differences of a few units are within fit systematics. This is compounded by the fact that PΠ = 3310 d exceeds the 2988 d data span, making the direct Keplerian signal a partial sinusoid that is degenerate with the PL3 model. Therefore the planet mass, inclination, and Kozai-resonance coincidence are not adequately supported by the current data. A likelihood-ratio test or noise-simulation control would settle the question. If the test confirms non-detection, the paper should be revised to remove the planet-specific claims from the title and abstract and to present the planet as a speculative but viable model. This is exactly the reader's CONDITIONAL verdict; no adjustment is needed.","tokens_in":36681,"tokens_out":9546,"duration_ms":100443,"concrete_test":"Using the released ToAs and Nutimo (Zenodo doi:10.5281/zenodo.13899771), compute a formal likelihood-ratio test between the Kepler and Planet models: Δχ² = χ²_Kepler − χ²_Planet with 2 degrees of freedom. With the reported Δχ² ≈ 1.6 this yields p ≈ 0.45, i.e., no evidence for mutual interactions. As a control for the multimodal landscape documented in Appendix C, repeat the Minuit refit from the mean and maximum-posterior starting points of both models and from the PL3 and PL5 solutions; if the spread in Δχ² across these starts exceeds ~2, the improvement is indistinguishable from fit systematics. Only a positive result of this test (p < 0.01 or Δχ² > 9.2) would justify retaining the planet mass, inclination, and Kozai claims; otherwise they should be removed or explicitly labeled as unconstrained.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing step for the planet interpretation is the claimed detection of mutual interactions, Eq. (B.22): δΔ_R = α t (γ_c cos n_O t + γ_s sin n_O t). This term is the only observable that distinguishes the full Planet model from a Keplerian sinusoid, and it enters at the outer-binary frequency. The current data do not support a detection. The fitted planet period is PΠ = 3310 d (Table 1), longer than the 2988 d observing span (MJD 56492–59480), so the Keplerian Rømer signal covers less than one cycle and is degenerate with the PL3 red-noise model. Table 2 shows the Planet model improves χ² over Kepler by only Δχ² ≈ 1.6 for two extra parameters, with no quoted significance. Section 4.2 states that the outer-binary frequency 'does not allow for a clear signature with the current data as it is already very well fitted due to the outer binary contribution,' and Appendix C warns that 'a difference of a few units in log-likelihood may not be safely considered as significant given the systematic uncertainty of the fitting procedure,' with local-minimum scatter of tens in χ². Consequently the inferred mass (1.23e-8 M_sun), inclination (119°), and Kozai-resonance coincidence are not robustly constrained. If the mutual-interaction term is absent or mis-modeled, the Planet model collapses to a Keplerian sinusoid statistically equivalent to PL3, and the title's 'explanation' is unsupported. The achromaticity result and the model-dependent SEP limits survive, but the planet-specific claims need either a significance statement or removal.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper analyzes 8 years of Nançay timing data of PSR J0337+1715 to model a ~4.4 microsecond low-frequency residual. Three model families are compared: chromatic (DM) red noise, achromatic power-law red noise (PLn), and a circum-ternary planet in a hierarchical orbit. The authors report that DM variations are ruled out, that both an achromatic red-noise (PL3) model and a planet model fit the data equally well, and that the two hypotheses cannot be statistically separated with current data. Under the planet assumption, a marginal mutual-interaction signal is used to constrain the planet mass to ~1.2e-8 solar masses (0.4 Moon masses) and an inclination near the Kozai resonance. Updated strong-equivalence-principle limits are derived: |Delta| < 1.5e-6 (planet) and |Delta| < 2.3e-6 (red noise) at 95% confidence. The paper also discusses a formation scenario and a low supernova kick velocity for the triple system. The manuscript is transparent about the model-selection ambiguity, but the title and abstract present the planet interpretation as an 'explanation' despite the authors' own caveats that the statistical differences are within systematic uncertainties.","tokens_in":37136,"tokens_out":4272,"duration_ms":43498,"significance":"If the planet interpretation were robust, this would be a remarkable discovery: the lowest-mass exoplanet around a pulsar and the first circum-ternary planet, with an orbit coincident with a von Zeipel-Lidov-Kozai resonance. However, the data do not establish the planet's existence. The robust contributions are the exclusion of DM variations, the identification of an exceptionally strong achromatic low-frequency signal, the model-dependent SEP limits, and the detailed perturbative treatment of the planetary Rømer delay in Appendix B. The paper also makes its dataset, code, and MCMC results publicly available (Zenodo), which is a strength for reproducibility. Nevertheless, the central planet-specific claims—mass, inclination, and resonance coincidence—depend on a marginal interaction term whose detection is not statistically significant, and the planet model is degenerate with red noise over the current timing span. The paper's scientific value is real but its headline conclusion is speculative; the manuscript needs substantial revision to align claims with evidence.","major_comments":[{"comment":"The Planet model improves χ² over the Kepler model by only Δχ² ≈ 1.6 for two additional parameters, and Appendix C explicitly states that 'a difference of a few units in log-likelihood may not be safely considered as significant given the systematic uncertainty of the fitting procedure.' Therefore the 'marginal detection of mutual interactions' (Eq. B.22) claimed in §4.2 and §5.2 is not statistically supported. The derived planet mass, inclination, and orbital parameters in Table 1 are thus not robustly constrained. Please reframe these quantities as conditional constraints under an assumed Planet model, not as detections, and remove or qualify the word 'detection' throughout the abstract and conclusions.","section":"§4.1, Table 2, Appendix C"},{"comment":"The best-fit planet period PΠ = 3310 d is longer than the 2988 d observing span (MJD 56492–59480), so the Keplerian Rømer signal covers less than one full cycle. As the paper notes in §4.2, the outer-binary frequency component does not provide a clear signature because it is already well fitted by the triple-system model. Consequently, the planet model is degenerate with the PL3 red-noise model, and the only distinguishing feature is the mutual-interaction term, which is not significantly detected (see comment above). The title 'Explanation of the exceptionally strong timing noise ... by a circum-ternary planet' overstates the evidence; the title and abstract should explicitly acknowledge that the red-noise and planet hypotheses are both viable and currently indistinguishable.","section":"§3.2.2, Table 1, §4.2"},{"comment":"The claimed coincidence of the planet's inclination with the exterior von Zeipel-Lidov-Kozai resonance at 116.6° is based on a very broad posterior, δiΠ = 119+16 −42°, which spans roughly 77°–135°. Given that the inclination constraint itself relies on the undetected mutual-interaction term, the phrase 'intriguingly coincident' (also used in the abstract) is not statistically meaningful. Please provide a quantitative estimate of the chance probability of such an alignment, or clearly label this as an unquantified speculation that does not support the planet interpretation.","section":"§5.2, Fig. 8, Table 1"}],"minor_comments":[{"comment":"The note to Table C.1 states Ndof = 13534 − Npar, whereas the text in §2 and the note to Table 2 give Ndof = 12474 − Npar for the same dataset. This inconsistency should be corrected.","section":"Appendix C, Table C.1"},{"comment":"The perturbed Rømer delay is written in the main text as δΔ_R = α t (γ_c cos n_O t + γ_s sin n_O t), but in Appendix B the same quantity is denoted δΔ_1 and uses E_1 = n_1(t − T_1). Please unify the notation to avoid confusion between the outer-binary mean motion and the planet's mean anomaly.","section":"Appendix B.2, Eq. (B.22)"},{"comment":"The abstract states that mutual interactions 'allow us to constrain its mass to ∼ 0.5 M_Moon as well as its inclination.' Given that the detection is marginal (Δχ² ≈ 1.6) and the paper itself cautions about systematic uncertainties, this phrasing should be softened to reflect that these constraints are obtained only under the assumption that the Planet model is correct.","section":"Abstract and §5.2"}],"recommendation":"major_revision","confidential_remarks":"This is a solid observational paper with a clear and transparent presentation, and the data release is commendable. The main issue is the gap between the paper's title/abstract, which present the circum-ternary planet as an explanation, and the paper's own statistical results, which show that the red-noise and planet models are indistinguishable. The planet-specific claims rest on a Δχ² of about 1.6, which is not significant. This is fixable by reframing the conclusions and title, but the revision is substantive rather than cosmetic because it affects the paper's central claim. I believe the paper can be made acceptable after these changes."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know before you read it. The solid result: the ~4.4 μs low-frequency residual in PSR J0337+1715 is achromatic, so dispersion-measure variations are ruled out, and the eight-year Nançay dataset yields new SEP limits that are honestly presented as model-dependent (|Δ| < 1.5×10⁻⁶ under the planet model, < 2.3×10⁻⁶ under red noise). The softer result: the planet. On the paper's own statistics, a circum-ternary planet and a steep three-harmonic red-noise process (PL3) fit about equally well, and the one observable that distinguishes them is not significantly detected.\n\nWhat is genuinely good. The analytic first-order perturbation theory in Appendix B — the mutual-interaction term δΔ_R = αt(γ_c cos n_O t + γ_s sin n_O t) — is a real derivation from Laplace-Lagrange theory, validated against the numerical integration, and it tells future observers exactly what signature to look for. The model comparison is unusually honest: Appendix C documents local-minimum scatter of tens in χ² and states flatly that differences of a few log-likelihood units are not significant. Data and code are on Zenodo. The point that PL2 and a low-eccentricity Keplerian planet are equivalent at leading order is a useful, clean observation.\n\nWhere it is soft. The planet period is 3310 d against a 2988 d data span, so the Keplerian Rømer signal covers less than one cycle and is degenerate with red noise by construction. The mutual-interaction term improves χ² over the Keplerian planet by only ~1.6 for two extra parameters, at a frequency the outer binary already fits well; Section 4.2 concedes there is no clear signature, and no detection significance is quoted. Everything quantitative about the planet — mass 1.23×10⁻⁸ M_sun, inclination 119°, the Kozai-resonance coincidence — rides on that term, and the inclination's error bars are tens of degrees wide, so a resonance a few degrees wide is at best a weak hint. The paper concedes this in the abstract's \"marginal detection\"; the title's \"explanation by a circum-ternary planet\" does not.\n\nBottom line: this deserves a serious referee. The achromaticity result, the perturbation model, and the model-dependent SEP bounds are citable and will stand regardless of the planet's fate. The referee's main job is to make the authors either quantify the significance of the mutual-interaction term or soften the title and the planet framing. Good reading-group material on how easily a sub-cycle Keplerian signal masquerades as red noise.","headline":"Careful, honest timing analysis: achromaticity and model-dependent SEP limits are solid, but the circum-ternary planet is an unproven explanation and the title overstates it.","tokens_in":37791,"tokens_out":4764,"would_cite":true,"duration_ms":46671,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["97.60.Gb","04.80.Cc"],"model":"deepseek-v4-flash","headline":"For PSR J0337+1715, the ~4.4-microsecond low-frequency timing residual is achromatic and is either exceptionally strong red noise or a Moon-mass planet around the triple system, with the gravity-test limit depending on which model is right.","keywords":["PSR J0337+1715","pulsar timing","strong equivalence principle","circum-ternary planet","timing noise","red noise","von Zeipel-Lidov-Kozai resonance","supernova kick"],"falsifier":"Continue timing PSR J0337+1715 for roughly another five to eight years and measure the residual Fourier amplitude at the outer-binary period: the planet model predicts this amplitude grows linearly in time (the $\\delta\\Delta_R$ term), whereas any red-noise or Keplerian model does not; if the amplitude does not grow above the white-noise floor, the planet model is falsified, and if it does, the red-noise model is.","tokens_in":36471,"feed_emoji":"🪐","tokens_out":8419,"duration_ms":72539,"temperature":0.7,"pith_summary":"PSR J0337+1715 is a millisecond pulsar in a compact triple system with two white dwarfs, the best known laboratory for testing the strong equivalence principle with a strongly self-gravitating object. This paper tackles a ~4.4-microsecond, ~8-year-period residual that had been limiting the test, and shows it cannot be caused by dispersion-measure variations: the signal is achromatic. Two models fit equally well: an exceptionally strong power-law red-noise process, or a small planet in a ~3310-day orbit around the triple system. If the planet is real, its mutual interactions with the triple system are marginally detected, giving a mass of about 0.4 Moon masses and an inclination close to the exterior von Zeipel-Lidov-Kozai resonance. The resulting strong-equivalence-principle bound is $|\\Delta| < 1.5\\times10^{-6}$ (planet) or $|\\Delta| < 2.3\\times10^{-6}$ (red noise) at 95% confidence, so which model is correct matters for the gravity test.","feed_headline":"A Moon-mass planet may explain this pulsar's extreme timing noise","feed_subtitle":"The signal is achromatic, and the gravity-test limit depends on which model is right.","key_machinery":"The distinguishing mechanism is the first-order mutual-interaction Rømer delay of a circum-ternary planet, $$\\delta\\Delta_R = \\$\\alpha$ t\\,(\\gamma_c\\cos n_O t + \\gamma_s\\sin n_O t),$$ a sinusoidal oscillation at the outer-binary frequency $n_O$ whose amplitude grows linearly in time (Eq. B.22). It is derived with Laplace-Lagrange perturbation theory in Jacobi coordinates and, because $\\gamma_c$ and $\\gamma_s$ depend independently on inclination and node, it lifts the Keplerian degeneracy between mass and inclination. The competing red-noise branch uses a truncated Fourier series $F(t_a)=\\sum_{k=1}^n A k^{-\\gamma}\\sin(2\\pi k\\nu t_a+\\phi_k)$ as a deterministic stand-in for a stochastic Gaussian-process red noise with a power-law spectrum.","core_discovery":"With eight years of Nançay timing data, the paper establishes that the low-frequency timing residual of PSR J0337+1715 has grown to ~4.4 microseconds over ~3000 days, exceeds the ~2-microsecond single-ToA uncertainty, and is achromatic: a model with ten dispersion-measure bins finds no significant DM variation. The signal is therefore not a propagation effect. The paper then shows that an achromatic red-noise model with three Fourier components and a power-law spectrum (PL3) and a planet model both reduce the residuals to white noise, with differences too small to choose between them by information criteria given fitting systematics. In the planet interpretation, a hierarchical four-body numerical integration yields a $1.23^{+1.1}_{-0.66}\\times10^{-8}\\,M_\\odot$ companion (about 0.4 Moon masses) in a ~3310-day, mildly eccentric orbit inclined near 119 degrees with respect to the triple-system plane, with a marginal detection of mutual interactions that is necessary to fix mass and inclination. If the signal is instead red noise, its amplitude is five to ten times larger than empirical scaling laws predict for a typical millisecond pulsar. The SEP test gives $|\\Delta|<1.46\\times10^{-6}$ under the planet model and $|\\Delta|<2.29\\times10^{-6}$ under PL3, a model dependence that motivates continued timing.","pith_inferences":["If the planet interpretation survives further data, the Kozai-resonance coincidence would be evidence that the planet is the sole survivor of a larger population of small bodies ejected during common-envelope evolution, implying that circum-ternary debris may be common around compact triple systems.","If the red-noise interpretation is instead correct, J0337's amplitude being five to ten times above empirical scaling laws would challenge the assumption that timing red noise tracks spin-down power, making J0337 a useful outlier for models of magnetospheric or superfluid-core noise.","The linearly growing mutual-interaction signal could in principle be searched for in timing data of other hierarchical triple pulsars, offering a generic test for circum-ternary planets rather than only this system.","The $\\pm180^\\circ$ degeneracy in the longitude of ascending node resolved here by annual-orbital parallax could be lifted independently by the planet's mutual-interaction term if the planet model is correct, providing a cross-check that is testable with future data."],"forward_implications":["Dispersion-measure variations are excluded as the source of the ~4.4 microsecond signal, so any explanation must be achromatic.","If the planet is real, it would be among the lightest exoplanets known, and its inclination near the exterior von Zeipel-Lidov-Kozai resonance suggests a resonance-stabilized survivor of the system's violent formation.","The strong-equivalence-principle limit depends on the noise model: $|\\Delta|<1.46\\times10^{-6}$ (Planet) versus $|\\Delta|<2.29\\times10^{-6}$ (PL3) at 95% confidence, a 30% improvement in the planet case and a 10% worsening in the red-noise case relative to the previous bound.","A longer observation span will produce a clear signature distinguishing the planet from red noise, because the mutual-interaction term grows linearly in time.","The inferred low supernova kick ($\\sim110$–$125\\,\\mathrm{km/s}$) supports the idea that small kicks are necessary for the survival of pulsar triple systems."],"supporting_citations":[{"why":"Supplies the numerical timing model, the Jacobi-coordinate parametrisation, and the previous $|\\Delta|<2.1\\times10^{-6}$ SEP limit that this work extends.","marker":"V oisin et al. (2020b)"},{"why":"Establishes J0337 as a triple system and motivates using it to test the strong equivalence principle.","marker":"Ransom et al. (2014)"},{"why":"Independent analysis that set the prior best SEP bound ($|\\Delta|<2.6\\times10^{-6}$) and provides the comparison baseline.","marker":"Archibald et al. (2018)"},{"why":"Formation scenario for the triple system on which the planet-survival and supernova-kick estimates are built.","marker":"Tauris & van den Heuvel (2014)"},{"why":"Formula for the probability that a planet remains bound after supernova mass-loss and kick.","marker":"Hills (1983)"},{"why":"Empirical red-noise scaling law that makes J0337's red-noise amplitude an outlier by a factor of five to ten.","marker":"Shannon & Cordes (2010)"},{"why":"Provides the masses of the PSR B1257+12 planets, the benchmark against which the candidate planet's ~0.4 Moon mass is compared.","marker":"Konacki & Wolszczan (2003)"},{"why":"One of the pulsar-timing-array studies whose millisecond-pulsar red-noise amplitudes are used to classify J0337's noise as exceptionally strong.","marker":"Chalumeau et al. (2022)"}],"fun_headline_variants":["Moon-mass planet may explain pulsar's extreme timing noise","Pulsar's timing noise hints at a Moon-mass companion","Tiny planet could be behind pulsar's odd timing","Gravity test results hinge on pulsar's mystery signal","A Moon-sized planet might cause pulsar's jitter"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The full planet interpretation — its mass, inclination, and the claimed Kozai-resonance coincidence — rests on the marginal detection of the mutual-interaction term at the outer-binary frequency, which the current data do not show as a clear signature; without that term the planet reduces to a Keplerian sinusoid that is degenerate with the PL3 red-noise model.","fun_headline_variants_meta":{"raw":{"variants":["Moon-mass planet may explain pulsar's extreme timing noise","Pulsar's timing noise hints at a Moon-mass companion","Tiny planet could be behind pulsar's odd timing","Gravity test results hinge on pulsar's mystery signal","A Moon-sized planet might cause pulsar's jitter"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000694,"raw_usage":{"total_tokens":3269,"prompt_tokens":1207,"completion_tokens":2062,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":823,"completion_tokens_details":{"reasoning_tokens":1980}},"tokens_in":823,"tokens_out":2062,"duration_ms":15259,"temperature":1.0,"reasoning_tokens":1980,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T20:01:52.300221+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Continue timing PSR J0337+1715 for roughly another five to eight years and measure the residual Fourier amplitude at the outer-binary period: the planet model predicts this amplitude grows linearly in time (the $\\delta\\Delta_R$ term), whereas any red-noise or Keplerian model does not; if the amplitude does not grow above the white-noise floor, the planet model is falsified, and if it does, the red-noise model is.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Formula for the probability that a planet remains bound after supernova mass-loss and kick."},{"cited_title":"M., Stairs, I","cited_arxiv_id":null,"evidence_quote":"Establishes J0337 as a triple system and motivates using it to test the strong equivalence principle."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Formation scenario for the triple system on which the planet-survival and supernova-kick estimates are built."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Empirical red-noise scaling law that makes J0337's red-noise amplitude an outlier by a factor of five to ten."},{"cited_title":"& Wolszczan, A","cited_arxiv_id":null,"evidence_quote":"Provides the masses of the PSR B1257+12 planets, the benchmark against which the candidate planet's ~0.4 Moon mass is compared."}],"review_version":1}