{"id":"04029b97-3685-4bdf-b1cf-262a1fbe0896","arxiv_id":"2608.13527","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Using 2,052 single and 188 binary TESS planet hosts, the authors show that detection sensitivity drops sharply for planets transiting the fainter component of unresolved binaries, and they provide downloadable correction grids.","lead":"The authors measured how often TESS would detect transiting planets when the host star has a close stellar companion that only high-resolution imaging can reveal. They publish grids showing how detection sensitivity drops for small planets in unresolved binaries, especially around the fainter star.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Sensitivity grids are built from detected TOIs only, so the quantitative detection fractions are conditional on a planet already having been found; an injection-recovery test is needed before the lookup tables are used for occurrence-rate corrections.","rationale":"The central claim is quantitative, not merely qualitative: the paper gives specific detection fractions and publishes grids for use in other studies. The load-bearing assumption is that scaling the catalog SNR of already-detected TOIs via Eq. 2 yields an unbiased detection probability. That assumption is insecure because the sample is conditioned on detection: a quiet star that happens to host a detectable planet contributes a sensitivity curve, while an equally quiet star that hosts no detectable planet does not. For binaries the selection is stronger, since dilution lowers the chance of entering the TOI catalog at all. This is exactly the reader's weakest assumption, and the paper does not validate it with injection-recovery or with an alternative denominator. The concern does not overturn the qualitative result: dilution alone guarantees that secondary-host transits are harder to detect, and the bias from using favorable detected systems most likely makes the secondary-host sensitivity estimates too optimistic rather than too pessimistic. But the lookup tables are positioned as a community product for occurrence-rate corrections, so a systematic offset in the absolute detection fractions would matter for downstream analyses. The reader's CONDITIONAL verdict remains appropriate; no verdict change is needed, but the revision should either run an injection-recovery test or state explicitly that the grids are conditional on a TOI having been detected and should not be used as absolute completeness corrections.","tokens_in":12333,"tokens_out":13490,"duration_ms":155104,"concrete_test":"Select a representative sample of TESS targets in the same spectral-type and T-mag range that have WIYN/Gemini/SOAR speckle observations, irrespective of whether they are TOIs (e.g., a random subsample of the TIC with speckle coverage). Inject synthetic transits of planets with radii 1-20 R_Earth and periods 1-20 d into their light curves, computing the diluted depth and transit duration correctly for the primary and, for known unresolved binaries, for the secondary; then run the SPOC and QLP detection statistics (or a calibrated proxy) on the injected light curves. Compare the recovered fraction per radius/period bin with the grids in Figs. 7 and 9; a systematic offset would confirm that conditioning on detected TOIs biases the quantitative sensitivity estimates.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The grids in §4.1 and §4.2 are averages of scaled catalog SNRs over systems that already contain a detected, vetted TOI (Eq. 2 in §3.3). This estimates P(SNR_hyp > threshold | TOI present), not P(planet detectable | star observed by TESS). Stars with no detected transit never enter the sample, and because unresolved binaries are harder to detect in, the 188 binary TOIs constitute a stronger-selected subsample than the 2052 single-star TOIs. The consequence is that the quantitative detection fractions -- for example the ~90% for a 5 R_Earth, 10 d planet around an M dwarf in §4.1 and the ~10% for Neptune-sized secondary hosts at Delta_mag >= 4 in §4.2 -- are not unbiased sensitivity estimates and are unsafe as occurrence-rate corrections without a target-star denominator. The qualitative conclusion that secondary-host transits are strongly suppressed is physically expected and robust; the problem is the calibration of the lookup tables, which the paper offers as a community resource.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper quantifies how unresolved stellar companions affect TESS's ability to detect transiting planets by constructing radius-period sensitivity grids from 2052 single-star TOIs and 188 binary TOIs with speckle observations. For each TOI, the catalog transit SNR is scaled to hypothetical planet radii and periods via Eq. (2), compared against pipeline detection thresholds (7.1 for SPOC, 9 for QLP), and, for binaries, the planet radius is corrected for dilution using Eq. (3). The resulting detection fractions are presented as functions of spectral type for single stars and of magnitude difference for binaries, with the key finding that planets transiting the fainter secondary are strongly suppressed at large Delta-mag. The grids are released as machine-readable lookup tables.","tokens_in":12502,"tokens_out":7193,"duration_ms":70637,"significance":"If the calibration is valid, the paper provides a valuable community resource: it extends earlier injection-recovery work (Ziegler et al. 2021) by explicitly treating secondary-host transits and by delivering interpolable sensitivity grids. The qualitative conclusion—that TESS is substantially less sensitive to small planets in unresolved binaries, especially around the secondary—is physically expected and robust. The paper also draws a useful connection to the observed dearth of secondary-host systems in Lester et al. (2022). However, the absolute detection fractions are conditioned on the sample of already-detected TOIs, which limits their use as occurrence-rate completeness corrections without a target-star denominator.","major_comments":[{"comment":"The detection fractions are conditional on a TOI already having been detected and vetted in that system. Since all 2052 single and 188 binary systems were selected because they contain a detected transiting planet candidate, the fraction of grids exceeding threshold estimates P(SNR_hyp > threshold | TOI present), not P(planet detectable | star observed by TESS). Stars that were observed but produced no TOI are absent from the denominator, and because unresolved binaries are harder to detect in, the binary subsample is a stronger-selected set than the single-star subsample. Consequently, headline numbers such as ~90% for a 5 R_Earth, 10-day planet around an M dwarf (Fig. 7) and ~10% for Neptune-sized secondary hosts at Delta-mag >= 4 (Fig. 9) are not unbiased sensitivity estimates and are unsafe as occurrence-rate corrections without a target-star denominator. The qualitative conclusion that secondary-host transits are strongly suppressed is physically expected and likely robust, but the calibration of the lookup tables needs an injection-recovery test or a denominator built from the full TESS target list before the grids are released as a community resource. The caveat in §4.1 that the calculations do not account for orbital geometry does not address this conditionality.","section":"§3.3, Eq. (2); §4.1"},{"comment":"The application in Fig. 8 divides observed radius-period histograms by the sensitivity grids to produce 'sensitivity-corrected' period-radius distributions. This usage treats the grids as standard survey completeness corrections. Given the conditional-on-detection issue above, the corrected distributions inherit the selection bias of the TOI catalog and of the speckle-observed subset (targets selected for follow-up), so they are not a faithful estimate of the underlying period-radius distribution. The authors should either remove this application or clearly reframe it as a relative comparison within the detected TOI population.","section":"§4.1, Fig. 8"},{"comment":"The quantitative detection fractions in Fig. 9 are quoted without uncertainties (e.g., 'only about half' and 'about ten percent' at Delta-mag >= 4, based on N=43 systems, and N=19 for Delta-mag < 1). Since these numbers are central to the paper's quantitative claims and are intended for external use, the grids should be accompanied by uncertainties, for example binomial confidence intervals on the detection fractions or bootstrap resampling over systems. Without such uncertainties, the significance of differences between spectral-type and Delta-mag bins cannot be assessed.","section":"§4.2, Fig. 9"}],"minor_comments":[{"comment":"The criterion for a star to be considered 'vetted' with the speckle imagers is not defined (number of observations, detection limits, filters); please specify it so that the sample can be reproduced.","section":"§2"},{"comment":"The DBSCAN clustering parameters (epsilon and min_samples) are not reported; for reproducibility, state them and the distance metric used.","section":"§2"},{"comment":"The conversion from speckle Delta-mag to secondary stellar radius through the Pecaut & Mamajek (2013) sequence should state which photometric band is used and whether the sequence is applied iteratively; the current description is ambiguous.","section":"§3.2"},{"comment":"Eq. (2) is applied over the full 1-20 R_Earth radius interval even when the observed radius is near one end of this range; extreme extrapolations could be clipped or flagged to avoid giving undue weight to very large scaling factors.","section":"§3.1, Eq. (2)"},{"comment":"The footnote mentions 14 TOIs with SNR<7.1; please state whether any of these are in the sample and how the detection threshold is applied to them.","section":"§3.1, footnote 4"},{"comment":"The two panels have different x-axis scales, which makes the comparison of the primary- and secondary-host X_r distributions less direct; a shared axis or an inset would help the reader.","section":"Fig. 4"}],"recommendation":"major_revision","confidential_remarks":"The conditional-on-detection bias is the main substantive concern. If the authors can add an injection-recovery validation or clearly restrict the scope to relative sensitivity within the detected TOI population, the paper would be suitable for publication. The qualitative conclusions are sound, but the absolute lookup-table calibration needs to be addressed before the grids are used by the community."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colin, quick take on arXiv:2608.13527.\n\nThe paper does what the title says: it converts a known bias into downloadable quantitative sensitivity grids. That's genuinely useful. It improves on Ziegler et al. (2021) by explicitly modeling planets transiting the secondary star, which is the case where TESS really loses sensitivity. The result that secondary-host planets are heavily suppressed at large Delta mag is physically sound and well demonstrated. The single-star grids by spectral type are also a nice community resource.\n\nThe method is standard SNR scaling from Ciardi et al. (2013), applied at scale. The math is straightforward; the presentation is clear; and they are honest about the impact-parameter assumption in Eq. 2. I also appreciate that they make the tables available.\n\nNow the soft spot. As the stress-test note says, the grids are built from TOIs that already have a detected, vetted transit. So each grid estimates P(SNR_hyp > threshold | a transit was detected in this system), not P(planet detectable | TESS observed this star). Stars with no transit never enter the sample. For occurrence-rate corrections you need the full target-star denominator. The paper does not flag this, and it matters because the tables are offered as a community resource. The qualitative conclusion -- secondary-host planets are hard to detect -- survives this critique, since that deficit comes from the dilution physics, not the sample selection. But the absolute detection fractions in Figures 7 and 9 are likely optimistic, because detected TOIs skew toward high-SNR systems. I'd like to see an injection-recovery test or at least a prominent caveat stating that the grids are conditional and should not be used as absolute completeness without a target-star denominator.\n\nMinor point: there are no uncertainties on the grids. X_r depends on estimated secondary radii from Pecaut-Mamajek, and the SNR scaling has its own scatter. A statement of how that propagates would help.\n\nWho is this for? Exoplanet occurrence-rate people and anyone interpreting planets in binary TOIs. It deserves a serious referee, conditional on the authors addressing the conditional-detection issue. I'd send it out.","headline":"A useful, well-executed quantification of TESS's reduced sensitivity to planets in unresolved binaries, with a real caveat: the lookup tables are conditional on a transit having been detected and need an injection-recovery calibration before use as occurrence-rate corrections.","tokens_in":13054,"tokens_out":2450,"would_cite":true,"duration_ms":25118,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"TESS is far less sensitive to transiting planets in unresolved stellar binaries than around single stars, with the worst sensitivity for planets orbiting the fainter star when the magnitude difference is large.","keywords":["TESS","transiting exoplanets","stellar binaries","speckle interferometry","detection sensitivity","transit signal-to-noise","planet occurrence rates","dilution"],"falsifier":"Run injection-recovery simulations on the actual TESS light curves of the binary TOIs, planting synthetic transits of known depth, period, and duration on the secondary star, and compare the recovered detection fraction per Δmag bin with the grids' predictions; a systematic mismatch of more than about ten percent in any bin would invalidate the SNR-scaling method.","tokens_in":12120,"feed_emoji":"🪐","tokens_out":5713,"duration_ms":51112,"temperature":0.7,"pith_summary":"TESS's ability to find transiting exoplanets is substantially degraded when the host star has an unresolved stellar companion, and the loss of sensitivity is most severe for planets that orbit the fainter star. The paper quantifies this by scaling the observed transit signal-to-noise ratios of 2052 single-star and 188 binary TOIs to hypothetical planets of any radius and period. For binaries in which the stars differ by four or more magnitudes, only about half of close-in gas giants and about ten percent of Neptune-sized planets transiting the secondary would pass the detection thresholds of the TESS pipelines. The resulting sensitivity grids give future occurrence-rate studies a way to correct for the bias against small planets in binaries.","feed_headline":"TESS blind to most small planets in faint-star binaries","feed_subtitle":"For binaries 4+ magnitudes apart, only ~10% of Neptune-sized planets around the fainter star would be detected.","key_machinery":"The central tool is the SNR scaling relation $SNR_{\\rm hyp} = SNR_{\\rm obs}\\, (R_{\\rm hyp}/R_{\\rm obs})^2\\, (P_{\\rm hyp}/P_{\\rm obs})^{-1/3}$ (Eq. 2), which maps each detected planet's catalog signal-to-noise ratio to hypothetical planets of other radii and periods, combined with a dilution-correction factor $X_r$ (Eq. 3) that rescales the planet radius in binaries according to which star is transited. Applied to thousands of TOIs, these relations convert a list of detected planets into grids of detection fractions: the fraction of systems in which a planet of a given radius and period would cross the pipeline threshold (SNR ≥ 7.1 for SPOC, ≥ 9 for QLP).","core_discovery":"Using the transit SNR, radius, and period of every TOI in their speckle-vetted sample, the authors calculate the detection fraction of hypothetical planets as a function of radius and orbital period. They find that TESS sensitivity to transiting planets in unresolved binaries depends sharply on which star hosts the planet: for magnitude differences Δmag ≥ 4, roughly half of close-in gas giants and only about ten percent of Neptune-sized planets would be detected if they transit the fainter star, whereas sensitivity to planets transiting the brighter star is nearly identical to that for single-star hosts. For single stars, sensitivity improves with later spectral type, from about 50 percent for a 5 R⊕ planet in a 10-day orbit around an F dwarf to about 90 percent around an M dwarf.","pith_inferences":["The same dilution physics applies to any transit survey with unresolved stellar companions, so the approach and grids could be adapted to Kepler and future missions like PLATO.","The sample's lack of near-equal binaries at 0.4–1.2 arcsec separation points to a selection effect in speckle follow-up; a completeness correction for that gap would refine the Δmag<1 grid.","The SNR scaling assumes a fixed transit chord, but inclined orbits change transit duration with period; full light-curve injection tests on binary TOIs would test how much this approximation matters."],"forward_implications":["Occurrence-rate studies that use TESS planets in binaries must divide observed counts by these sensitivity grids; otherwise they will systematically underestimate the population of small planets around secondary stars.","The missing radius valley in binary-host planets reported previously may be partly a detection artifact rather than a purely astrophysical difference.","For binaries with Δmag ≥ 4, the practical assumption that the planet transits the primary is justified, but it leaves the planetary content of high-Δmag binaries essentially unknown.","The grids can be interpolated to estimate TESS sensitivity for any TOI of known spectral type, radius, and period, not just the sample stars."],"supporting_citations":[{"why":"Supplies the SNR scaling relation (Eq. 2) used to project detectability of hypothetical planets.","marker":"Ciardi et al. 2013"},{"why":"Establishes the dilution-bias framework and the radius-correction factor used for binary TOIs.","marker":"Ciardi et al. 2015"},{"why":"Provides the stellar sequence used to estimate the secondary star's radius from speckle Δmag.","marker":"Pecaut & Mamajek 2013"},{"why":"One of the three speckle-imaging programs that vetted the single and binary TOIs.","marker":"Howell et al. 2021"},{"why":"Gemini speckle observations that identified companions and showed small planets are preferentially found around single stars.","marker":"Lester et al. 2021"},{"why":"SOAR speckle survey and injection tests that motivated the binary-sensitivity question; also provided many binary TOIs.","marker":"Ziegler et al. 2021"},{"why":"Defines the QLP detection threshold (SNR ≥ 9) used for QLP-detected TOIs.","marker":"Kunimoto et al. 2023"},{"why":"Supplies the geometric argument that secondary stars have lower transit probability, compounding the detection bias.","marker":"Bouma et al. 2018"}],"fun_headline_variants":["TESS finds few small planets orbiting faint binary stars","TESS misses 90% of Neptune-sized planets in unresolved binaries","TESS sensitivity drops for planets around fainter stars in binaries","Bias: TESS detects few small planets on fainter binary members"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The analysis assumes that the sample of already-detected TOIs with speckle observations is representative enough that scaling each detected system's catalog transit SNR to hypothetical planets yields the true detection probability of TESS in that system.","fun_headline_variants_meta":{"raw":{"variants":["TESS finds few small planets orbiting faint binary stars","TESS misses 90% of Neptune-sized planets in unresolved binaries","TESS sensitivity drops for planets around fainter stars in binaries","Bias: TESS detects few small planets on fainter binary members"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00069,"raw_usage":{"total_tokens":3141,"prompt_tokens":974,"completion_tokens":2167,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":590,"completion_tokens_details":{"reasoning_tokens":2093}},"tokens_in":590,"tokens_out":2167,"duration_ms":17612,"temperature":1.0,"reasoning_tokens":2093,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T04:56:52.747708+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run injection-recovery simulations on the actual TESS light curves of the binary TOIs, planting synthetic transits of known depth, period, and duration on the secondary star, and compare the recovered detection fraction per Δmag bin with the grids' predictions; a systematic mismatch of more than about ten percent in any bin would invalidate the SNR-scaling method.","supporting_citations":[{"cited_title":"R., Fabrycky, D","cited_arxiv_id":null,"evidence_quote":"Supplies the SNR scaling relation (Eq. 2) used to project detectability of hypothetical planets."}],"review_version":1}