{"id":"fb3fd81a-2d11-45cd-a175-521faed8c93d","arxiv_id":"2501.04082","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"A catalog of 584 TESS eclipsing binaries shows a small population rotating at 7/8 of the orbital period, independently confirming a puzzling subsynchronous group first seen in Kepler data.","lead":"This paper measures rotation periods for 584 starspot-modulated eclipsing binaries from TESS and compares them with their orbital periods. It finds a small population rotating at about 7/8 of the orbital period, matching a puzzling population first seen in Kepler data.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The statistical case for the 7:8 population is built on a CDF pseudo-likelihood (Eq. 4) whose adjacent points are strongly correlated; the reported ln Bayes factors (~10^3) are uncalibrated, and the TESS-only excess is weak (about 6%).","rationale":"The paper delivers real value: a hand-vetted catalog of 584 starspot-modulated TESS EBs, orbital and rotation periods from three methods, and a genuine cross-instrument check of the L17 subsynchronous population. The qualitative statement that the TESS and Kepler spin-orbit distributions are consistent is supported by the reported Cramer-von Mises p = 0.99, and the authors are explicit about their period-uncertainty limitation. The load-bearing defect is that the only quantitative significance claim is built on the Equation (4) pseudo-likelihood, which is not a valid likelihood: the empirical CDF values are correlated to an extreme degree, the sigma = 0.025 is a fiducial fudge factor calibrated to the width of the 1:1 peak, and the resulting ln Bayes factors in the thousands are not interpretable as calibrated evidence. An honest reading is that the existence of some subsynchronous excess in TESS receives moderate support from the histogram and the distributional consistency with Kepler, but the specific assertion that the 7:8 rotators 'cannot be attributed to noise' outruns the statistics. I therefore keep the reader's CONDITIONAL verdict rather than strengthening it: the requested checks (a properly calibrated likelihood, per-source or at least realistic uncertainties, and a TESS-only statement) are exactly what would decide the matter. I also flag the odd pairing of p = 0.99 for the two-sample CvM test with a claimed 6% versus 15% subsynchronous fraction as an internal tension worth resolving.","tokens_in":19750,"tokens_out":17296,"duration_ms":166274,"concrete_test":"Recompute the Section 3.5 model comparison with a proper likelihood: multinomial/Poisson counts of Porb/Prot in ~20 bins over [0.7, 1.3] (same six models, Table 2 priors, dynesty), reporting ln BF(double vs single Gaussian) for TESS-only, Kepler-only, and combined. If TESS-only ln BF falls below ~10 (or changes sign), the decisive TESS significance claim fails. As a calibration cross-check, draw N=584 synthetic samples from the best-fit single Gaussian and run them through the paper's own Eq-4 CDF pipeline; if a substantial fraction of unimodal realizations yield ln BF > 1330, the published Bayes factors are uncalibrated regardless of the TESS-only result.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The headline claim that the 7:8 rotators 'cannot be attributed to noise' rests on the Bayes factor analysis of Sections 3.5 and 4.5, which uses the improper likelihood of Equation (4): N residuals between the model CDF Y(x_i) and the empirical CDF C(x_i) at the observed spin-orbit ratios, each treated as an independent Gaussian error with a common sigma = 0.025. The empirical CDF at adjacent order statistics is not independent; C(x_i) and C(x_{i+1}) share all but one data point of information, so the effective number of independent constraints is of order tens, not N = 584 (TESS) or ~1400 (combined). The resulting ln Bayes factors (1330, 2607, 3838 for TESS, Kepler, combined) therefore have no calibrated statistical meaning, and the Figure 19 robustness test varies sigma yet retains the same correlated residual structure, so it cannot rescue the test. The TESS-only evidence is weak on its own (about 6% of TESS sources in the 0.82-0.92 window versus L17's 15%, described as 'a weak subsynchronous peak'), so confirmation of significance largely rides on the combined sample and on the two-sample Cramer-von Mises p = 0.99 claim, which is difficult to reconcile with a 6%-versus-15% gap in the same window. The targeted recomputation of periods for targets already in the 0.82-0.92 range (Section 4.2) additionally means the TESS sample fed into this test is not uniformly measured. The conclusion itself concedes that the lack of per-period uncertainties 'limits statistical methods of confirming the presence of the 7:8 rotators.'","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a new catalog of rotation periods for 584 high-confidence starspot-modulated eclipsing binaries from TESS, together with orbital periods, eccentricities, and light-curve classifications. Using Lomb-Scargle, ACF, and PDM period measurements with visual vetting, the authors compare the distribution of the spin-orbit ratio Porb/Prot against the Kepler sample of Lurie et al. (2017). They report a weak subsynchronous excess at Porb/Prot ~ 0.89 in TESS (about 6% of the sample in the range 0.82-0.92, versus 15% in Kepler), and use Bayesian model comparison with Gaussian and Lorentzian CDF fits to claim that a double-peaked model is decisively preferred over a single-peaked model, with natural log Bayes factors of order 10^3. The paper concludes that roughly 10% of short-period EBs rotate near the 7:8 spin-orbit ratio and that this population is statistically significant and not an instrument-specific artifact.","tokens_in":20149,"tokens_out":3604,"duration_ms":35220,"significance":"If the central claim holds, this is a valuable independent confirmation of an unusual subsynchronous population in eclipsing binaries, obtained with a different instrument and different analysis choices than the original Kepler detection. The catalog itself is a useful community resource: it provides inspected rotation periods, alias flags, eclipse parameters, and cross-method comparisons for 584 TESS EBs, and the methodology (random-forest classification plus human vetting, three period-finding methods, and eclipse masking) is carefully documented. The paper also honestly states its main limitation: individual per-period uncertainties are not measured, and the authors acknowledge that this limits statistical confirmation of the 7:8 population. However, the headline significance claim rests on a pseudo-likelihood whose statistical meaning is not established, and the TESS-only evidence is explicitly weak. The paper's value as a catalog and as a qualitative confirmation is real, but the quantitative claim of statistical significance needs substantially more careful treatment before it can be accepted.","major_comments":[{"comment":"The likelihood in Eq. (4) treats the empirical CDF values C(x_i) at the observed spin-orbit ratios as N independent data points with a common Gaussian error sigma = 0.025. This is not a valid statistical model: adjacent order statistics of a CDF are strongly correlated, because C(x_i) and C(x_{i+1}) share all but one data point of information. The effective number of independent constraints is far smaller than N = 584 (TESS) or ~1400 (combined). Consequently, the natural log Bayes factors reported in Table 6 (1330, 2607, 3838 for TESS, Kepler, and combined) are not calibrated, and the enormous values do not have their usual evidential meaning. The robustness test in Figure 19 varies sigma but retains the same correlated residual structure, so it cannot rescue the test. The authors should replace this with a likelihood that respects the dependence structure of the data (e.g., a quantile or order-statistic likelihood, a bootstrap over light curves, or a direct comparison of fitted parameter posteriors), or explicitly present the Bayes factors as exploratory rather than as decisive evidence.","section":"Section 3.5, Eq. (4)"},{"comment":"The authors recomputed rotational and orbital periods using full-sector data only for targets already falling in the subsynchronous range 0.82 < Porb/Prot < 0.92 (Section 4.2). This differential remeasurement can sharpen precisely the feature being tested, because the targets in the candidate peak are the ones that receive the higher-quality full-sector treatment. The TESS sample fed into the model comparison in Section 4.5 is therefore not uniformly measured. The authors should either apply the full-sector treatment to the entire sample or demonstrate, with a controlled experiment, that the targeted remeasurement does not preferentially move sources into the subsynchronous window.","section":"Section 4.2 and Section 4.5"},{"comment":"The TESS-only evidence for the subsynchronous population is weak: the paper states that 6% of TESS sources fall in the 0.82-0.92 window, compared with 15% in L17, and describes the TESS feature as 'a weak subsynchronous peak.' This is difficult to reconcile with the two-sample Cramer-von Mises p-value of 0.99 reported in Section 3.5, which is also used to justify pooling the samples. A p-value of 0.99 for a 6%-versus-15% difference in the same window is suspicious and suggests either a lack of power in the test or a problem with how the CDF comparison was performed. The significance claim should be reported for TESS alone with an honest, calibrated test; if TESS alone does not reach significance, the conclusion that the 7:8 population is 'confirmed' in TESS should be softened accordingly.","section":"Section 4.2 and Section 3.5"},{"comment":"The authors acknowledge in the conclusion that the lack of robust per-period measurement uncertainties 'limits statistical methods of confirming the presence of the 7:8 rotators.' This is a load-bearing limitation, not just a caveat for future work. With a uniform, hand-assigned sigma = 0.025 for all spin-orbit ratios, the model comparison cannot distinguish measurement scatter from an intrinsic physical population width. In particular, the reported amplitude of the secondary peak and the claim that the subsynchronous rotators are 'distinct from the synchronous rotators' are conditional on this unverified error model. The analysis would be substantially strengthened by deriving per-target period uncertainties (e.g., from Gaussian process or bootstrap fits) and propagating them into the distribution comparison.","section":"Section 5 and Section 3.5"}],"minor_comments":[{"comment":"The expression for ln L omits the constant term -(N/2) ln(2 pi), which cancels in model comparison but should be included for completeness; as written, the formula is not exactly a Gaussian log-likelihood.","section":"Equation (4)"},{"comment":"In the Kepler Median row for the skewed Gaussian model, sigma1 is listed as 0.0426, identical to the TESS value, while the Kepler Mean value for the same parameter is 0.0399; this looks like a copy-paste error and should be checked.","section":"Table 5"},{"comment":"The concluding sentence says systems 'rotate very closely to 7/8ths of their orbital period,' but the analysis is framed throughout in terms of Porb/Prot ~ 0.875, which corresponds to Prot being about 8/7 of Porb (i.e., slower rotation). The wording '7/8ths of their orbital period' is mathematically the opposite and should be corrected to avoid confusion.","section":"Section 5"},{"comment":"The Ricker et al. 2014 reference appears twice in the reference list with identical bibliographic details; one duplicate should be removed.","section":"References"},{"comment":"The caption for the two panels in Figure 12 would be clearer if it explicitly stated that the left panel is 20% cross-method confidence and the right panel is 10%, matching the order in the text.","section":"Figure 12"}],"recommendation":"major_revision","confidential_remarks":"This is a useful catalog paper with a genuinely interesting qualitative result, but the statistical evidence for the headline claim is not yet sound. The CDF pseudo-likelihood issue is the main obstacle; it is fixable with a more appropriate statistical comparison, but the current Bayes factors should not be cited as decisive. The paper would benefit from a more conservative framing of the TESS-only confirmation, given the weak 6% excess and the admitted lack of per-period uncertainties."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my take: the catalog is the real contribution, and the confirmation of the 7:8 population is plausible but not proven by the statistics as written.\n\nWhat's genuinely new: this is the largest TESS EB rotation-period catalog to date – 584 high-confidence starspot EBs with periods from Lomb-Scargle, ACF, and PDM, all visually vetted by at least two people. That's a reusable resource. The comparison with L17's Kepler sample is a legitimate independent check of the subsynchronous population with a different instrument and different analysis choices. The paper is also honest: it admits the TESS-only excess is weak (6% vs 15%) and that lack of per-period uncertainties limits the statistical confirmation.\n\nThe soft spot is Section 3.5. The likelihood in Eq. 4 treats the empirical CDF values at each data point as independent measurements with a common sigma = 0.025. They are not independent – adjacent CDF points share nearly all of the data. So the ln Bayes factors in the thousands do not have their usual meaning. The robustness test in Figure 19 varies sigma but keeps the same correlated residual structure, so it can't calibrate the test. The two-sample Cramer-von Mises p=0.99 also sits oddly next to the 6%-vs-15% difference in the same window; it's possible but it should be explained. And the decision to recompute periods only for targets already in the 0.82–0.92 range (Sec 4.2) makes the sample non-uniform; it could sharpen the very peak being tested.\n\nThat said, I don't think the paper is wrong in its broad conclusion – the TESS histogram does show a bump near 0.88, and the L17 population is recovered in a different survey. The problem is that the significance claim is overbuilt. A cleaner approach would be a histogram-based likelihood or, better, per-source period uncertainties from a GP or similar, plus a proper two-sample test that doesn't assume independent CDF points. This is fixable in revision.\n\nBottom line: useful catalog, plausible result, overstated significance. Worth sending to a good referee – the field needs this sample, and the statistical fix is tractable.","headline":"A genuinely useful TESS EB rotation catalog and a plausible independent recovery of the 7:8 population, but the significance claims rest on a pseudo-likelihood that overstates certainty.","tokens_in":20670,"tokens_out":2990,"would_cite":true,"duration_ms":28796,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["97.80.-d","97.10.Kc"],"model":"deepseek-v4-flash","headline":"About 10% of short-period eclipsing binaries rotate at 7/8 of their orbital period instead of synchronously, a population this paper finds independently in TESS and Kepler data.","keywords":["tidal synchronization","eclipsing binaries","starspot rotation","spin-orbit resonance","TESS","Bayesian model comparison","subsynchronous rotation","stellar rotation periods"],"falsifier":"Re-derive rotation periods for all 584 binaries with a blind full-posterior method (for example a Gaussian process) without first targeting objects in the 0.82 to 0.92 ratio range; the 7:8 peak is real only if it re-emerges and the fraction of systems in that bin does not depend on which objects were re-measured. A complementary test is to inject synthetic starspot light curves with known 1:1 synchronization into the same pipeline: a spurious 7:8 peak would indicate that the eclipse-masking or period-finding procedure creates the artifact.","tokens_in":19561,"feed_emoji":"🔭","tokens_out":9680,"duration_ms":81024,"temperature":0.7,"pith_summary":"This paper asks whether the odd subsynchronous population of eclipsing binaries first seen in Kepler data is real or a survey artifact, and answers by measuring rotation periods for 584 starspot-modulated short-period binaries from the TESS mission. It finds the same two-peaked structure in the ratio of orbital to rotation period: a dominant synchronous peak at 1:1 and a secondary peak near 7/8, comprising roughly 10% of the sample. Bayesian model comparison strongly favors a double-peaked distribution over a single-peaked one for the TESS, Kepler, and combined samples, which the authors take as evidence that the 7:8 rotators cannot be attributed to noise. Because the peak appears with a different telescope, detectors, period-finding codes, and human vetting, the paper concludes it is a genuine astrophysical population that tidal evolution theory will have to explain.","feed_headline":"TESS confirms: 1 in 10 close binaries spins at 7/8 orbital speed","feed_subtitle":"The Kepler oddity reappears in an independent survey, pointing to a real tidal spin state rather than an instrument artifact.","key_machinery":"The central object is the spin-orbit ratio $P_{\\rm orb}/P_{\\rm rot}$, summarized as an empirical cumulative distribution function and fitted by single- and double-peaked Gaussian and Lorentzian models through nested sampling; the evidence for a second population is quantified by Bayes factors (marginal likelihood ratios) between model pairs. The value $7/8$ marks the subsynchronous peak. The measurement pipeline is also load-bearing: a random-forest classifier trained on Kepler labels and then human-vetted, followed by three period-finding methods (Lomb–Scargle, autocorrelation, phase dispersion minimization) applied after masking eclipses, with the adopted period chosen by visual inspection of phase-folded light curves. The Bayes factor test is the mechanism that converts the appearance of a second peak into the claim that the 7:8 rotators are distinct from noise.","core_discovery":"The central claim is that the subsynchronous population of eclipsing binaries discovered in Kepler data is present in TESS data as well, and that the 7:8 overdensity is statistically distinct from the synchronous population. Working from the TESS Eclipsing Binary Catalog, the authors classify 4584 light curves, retain 584 high-confidence starspot-modulated systems with orbital periods under 10 days, and measure rotation periods with Lomb–Scargle, autocorrelation, phase dispersion minimization, and visual inspection. About 6% of the TESS sample falls in the spin-orbit range $0.82 < P_{\\rm orb}/P_{\\rm rot} < 0.92$, and this secondary peak is decisively favored over a unimodal model in Bayes factor tests on the cumulative distribution. An independent Cramér–von Mises test finds the Kepler and TESS distributions consistent with a common origin (p = 0.99). The paper therefore concludes that roughly 10% of short-period eclipsing binaries rotate very close to 7/8 of their orbital period instead of synchronously, and that existing tidal-plus-magnetic-braking models, which produce a broad subsynchronous spread, cannot account for the tightness of the 7:8 ratio.","pith_inferences":["Because the Bayes factor calculation treats correlated CDF points as independent, the reported log Bayes factors above 1000 are not trustworthy at face value; a bootstrap or a likelihood that models the CDF's correlation structure would give a more honest significance level, and the paper's sigma-sensitivity test cannot fix this because it only varies the assumed error magnitude, not the independe","A clean test of the artifact hypothesis would be to search for the same 7:8 peak in K2 data or in synthetic light curves drawn from a different window function, where the sampling and systematics differ from both missions; the paper's claim predicts the peak should appear regardless of window function.","If the peak is real, the near-exact rational ratio 7:8 suggests a dynamical resonance or a preferred spin state, possibly a spin-orbit commensurability set by tidal torque or a triaxial shape, rather than a continuously varying pseudosynchronous state; this predicts the ratio should be independent of orbital period and eccentricity within the population.","The paper re-measured periods only for objects already inside the 0.82–0.92 window, which risks sharpening the very peak being tested; a blinded re-analysis of the entire sample would settle whether the narrowness of the peak is a selection effect."],"forward_implications":["The 7:8 subsynchronous rotators are a real population appearing in two independent surveys, so future theories of tidal evolution must explain a tight ratio near 7/8 rather than a broad subsynchronous spread.","Existing simulations that couple tidal dissipation and magnetic braking can produce subsynchronous rotation for orbital periods longer than about 4 days but not the tight 7:8 overdensity, ruling out current implementations of those processes as a complete explanation.","The published catalog of 584 rotation periods, orbital periods, and eccentricities provides a new resource for studying tidal synchronization, differential rotation, and circumbinary exoplanets.","Among the three period-finding methods, phase dispersion minimization matches the visually inspected period in 96% of cases and has 98% ten-percent accuracy, making it the recommended method for starspot-modulated EB light curves; Lomb–Scargle tends to overestimate rotation periods, often by a factor of two."],"supporting_citations":[{"why":"The Kepler-based study that first identified the synchronous and 7:8 subsynchronous populations; the TESS results are compared directly against it as the benchmark discovery.","marker":"L17"},{"why":"The tidal-plus-magnetic-braking simulations that reproduce broad subsynchronous rotation but fail to produce the tight 7:8 overdensity, defining the theoretical gap the new observations sharpen.","marker":"F19"},{"why":"Source of the TESS Eclipsing Binary Catalog with the initial 4584 targets and their orbital periods.","marker":"Prsa et al. 2021"},{"why":"Provides the Bayes factor significance scale (ln B > 5 as decisive) used to judge that the double-peaked model wins.","marker":"Kass & Raftery 1995"},{"why":"The TESS mission paper establishing the survey as an independent data set with different instrumental systematics from Kepler.","marker":"Ricker et al. 2014"},{"why":"Supports the choice of the autocorrelation function as the more reliable period method for starspot-modulated light curves, informing the adopted rotation periods.","marker":"Gordon et al. 2021"}],"fun_headline_variants":["TESS data confirms odd 7:8 spin-orbit binary population","TESS sees same odd 7/8 spin state in 10% of close binaries","Two surveys agree: 10% of tight binaries spin at 7/8 orbital rate","TESS validates mysterious 7/8 spin-orbit ratio in eclipsing binaries","TESS finds same subsynchronous spin peak as Kepler in close binaries"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The significance of the 7:8 population rests on treating the points of the empirical cumulative distribution as independent measurements sharing one error of $\\sigma = 0.025$ in the likelihood; adjacent CDF points are strongly correlated, so the very large Bayes factors almost certainly overstate the evidence.","fun_headline_variants_meta":{"raw":{"variants":["TESS data confirms odd 7:8 spin-orbit binary population","TESS sees same odd 7/8 spin state in 10% of close binaries","Two surveys agree: 10% of tight binaries spin at 7/8 orbital rate","TESS validates mysterious 7/8 spin-orbit ratio in eclipsing binaries","TESS finds same subsynchronous spin peak as Kepler in close binaries"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000268,"raw_usage":{"total_tokens":1712,"prompt_tokens":1130,"completion_tokens":582,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":746,"completion_tokens_details":{"reasoning_tokens":476}},"tokens_in":746,"tokens_out":582,"duration_ms":5527,"temperature":1.0,"reasoning_tokens":476,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T21:41:15.060412+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Re-derive rotation periods for all 584 binaries with a blind full-posterior method (for example a Gaussian process) without first targeting objects in the 0.82 to 0.92 ratio range; the 7:8 peak is real only if it re-emerges and the fraction of systems in that bin does not depend on which objects were re-measured. A complementary test is to inject synthetic starspot light curves with known 1:1 synchronization into the same pipeline: a spurious 7:8 peak would indicate that the eclipse-masking or period-finding procedure creates the artifact.","supporting_citations":[],"review_version":1}