{"id":"93df890f-dcf3-42d0-a428-6b2602731cb9","arxiv_id":"2508.04673","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":8,"one_line_summary":"A TESS primary mission search finds five new exocomet-like transit candidates and estimates a 2.64e-4 per star per year occurrence rate, concluding such transits are rare at 0.1% to 1% depths.","lead":"Using two years of TESS data, this survey found six exocomet-like transit signals, including five new candidates around stars ranging from A/F to G-type main sequence stars and two giants. It estimates that such transits are very rare at depths of 0.1% to 1%, and that higher precision photometry will be needed to find more.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Injection-recovery uses the same skew-normal shape as the detection pipeline; real exocomet shapes could bias the headline 2.64e-4 rate.","rationale":"The reader's weakest assumption is exactly the injection-recovery shape dependence; it is the largest systematic in the occurrence-rate estimate. However, the paper's central qualitative conclusion ('very rare') would survive even a factor of a few change, and the detection of β Pic and careful vetting are genuine strengths. The CONDITIONAL verdict remains appropriate; no further adjustment needed.","tokens_in":30637,"tokens_out":10760,"duration_ms":123940,"concrete_test":"Recompute f_inj_rec injecting a grid of shapes (varying skewness 0.5–8, egress timescale ratios 2–10, and a physical dust-tail profile) at the depths/magnitudes of the five candidates, then re-derive Eq. 11. If the summed rate moves by more than the Poisson uncertainty (factor ~2), the headline rate is shape-dependent.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The occurrence rate in Eq. 11 relies on f_inj_rec measured in §3.5 by injecting a single skew-normal template (Eq. 8) based on β Pic. The detection cuts (e.g., 0<skewness<8, α>1.02) are matched to that model family, so f_inj_rec measures recovery of the exact assumed shape. Real transits around faint stars and giants may have different tail structures, asymmetries, or durations, and the median filter or common-time cut (§4.1) may suppress them. The correction factors reach ~90, so even a ~30% overestimate of f_inj_rec shifts the rate by a large factor. The qualitative conclusion that transits are rare is probably safe, but the quoted 2.64e-4 star^-1 yr^-1 and the factor-38 contrast with Kepler are not quantitatively robust without testing the shape sensitivity.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents an automated search for single, asymmetric exocomet-like transits in TESS primary-mission light curves (Sectors 1–26, 15.5 million light curves). The pipeline uses a moving-average test statistic, fits symmetric and skewed Gaussian profiles, and applies a sequence of cuts on SNR, duration, depth, skewness, and asymmetry ratio, followed by extensive manual vetting. The authors recover the known β Pic exocomet transit and report five new candidates, including three main-sequence stars (one F-type, two probable G-types) and two giant/supergiant stars. They use injection–recovery tests to estimate detection efficiency and derive a total occurrence rate of 2.64 × 10⁻⁴ star⁻¹ yr⁻¹ excluding β Pic, nearly 40 times higher than the earlier Kepler rate, and conclude that photometrically transiting exocomets are very rare at 0.1–1% depths.","tokens_in":30901,"tokens_out":6197,"duration_ms":72381,"significance":"If the quantitative occurrence rate is robust, this would be the first large-scale TESS census of photometric exocomet transits, with new candidate host types (G-type main-sequence stars and evolved stars) that challenge the A/F-star bias suggested by Kepler. The paper is carefully structured and the vetting is unusually detailed: the authors test a complementary TESS-SPOC sample, perform a time-reversed search, and provide injection-recovery maps over magnitude and depth. The code is publicly available, and the recovery of the known β Pic event is a useful end-to-end validation. The main value lies in the candidate list and the demonstration that a systematic TESS search is feasible; the specific occurrence rate, however, depends on several assumptions that are not yet tested.","major_comments":[{"comment":"The injection–recovery efficiency f_inj_rec is measured by injecting synthetic transits drawn from the same skew-normal family (Eq. 8) that defines the shape-selection cuts (0 < skewness < 8, α > 1.02). This measures recovery of the assumed model, not of real exocomet transits. Since correction factors reach ~90 (§6.3), even a modest misestimate of f_inj_rec changes the headline rate by a large factor. Please test sensitivity to shape assumptions by injecting alternative tail profiles (e.g., power-law egress, two-component coma+tail, different asymmetry/duration ratios) and recompute the efficiency map and final rate. At minimum, quantify how the occurrence rate shifts if the injection template is varied within plausible bounds.","section":"§3.5, Eq. (8)–(11)"},{"comment":"The thresholds (SNR>5, duration>0.4 d, depth<1%, 0<skewness<8, α>1.02) were developed after experimenting on Sector 6, which contains the known β Pic transit. This tuning means the completeness measured by injection–recovery is not a blind-search completeness; part of the efficiency is fitted to a known positive detection. The paper states that Sectors 9 and 17 were checked, but the thresholds are still informed by β Pic. Please quantify how the candidate list and f_inj_rec change if thresholds are set using only the comparison catalogues (Fig. 4) and not Sector 6, or state explicitly how much of the efficiency correction is attributable to this tuning.","section":"§4, Table 1"},{"comment":"No uncertainties are propagated into the occurrence rate. With five detections (excluding β Pic), the Poisson 95% interval spans roughly a factor of several; the injection-recovery map itself has binomial uncertainty (20,000 injections per bin), but no error bars are shown on the efficiency values. The factor-38 contrast with the Kepler rate (6.7 × 10⁻⁶) is therefore not statistically significant without confidence intervals. Please provide Poisson upper/lower limits on N_det and propagate the f_inj_rec uncertainties, or explicitly state that the quoted rates are order-of-magnitude estimates. Also clarify the counting in the text: it says 'summing the three bins' but there are five candidates excluding β Pic; specify exactly how many candidates and bins are included in the 1.95 × 10⁻⁵ star⁻¹ sector⁻¹ value.","section":"§6.3, Eq. (11)"},{"comment":"The occurrence rate treats all five new candidates as real exocomet transits, but the paper itself notes that some are tentative: TIC 110969638 (G supergiant) could be intrinsic stellar variability (§6.2), TIC 280832588 is described as tentative and fails the asymmetry cut in TESS-SPOC (§5.1.2), and the G-type classifications for two candidates lack spectroscopic confirmation. If one or more are false positives, the rate is overestimated. Please provide a conservative rate obtained by excluding the most uncertain candidate(s), or at least state how the conclusion changes under such an assumption.","section":"§5, §6.2"}],"minor_comments":[{"comment":"The abstract states the rate is 'much higher' than Kepler's; given the small number statistics and the unpropagated uncertainties, this should be softened to 'higher, though with large uncertainties'.","section":"Abstract / §6.3"},{"comment":"Reference to 'Fig 6.3' should be 'Fig. 7'.","section":"§6.3"},{"comment":"Typo: 'TIC 2297970952' should be 'TIC 229790952'.","section":"§5.2 heading"},{"comment":"The citation 'Iglesias et al. (in review)' lacks a full bibliographic entry; please provide an arXiv identifier or update the reference list.","section":"§1"},{"comment":"The axes are labelled in a way that is easy to misread (e.g., '10 3 10 2' appears without superscripts or units). Please add clear axis labels such as 'Transit depth (%)' and 'Magnitude (T mag)' and ensure the colour-bar scale is legible.","section":"Figure 7"}],"recommendation":"major_revision","confidential_remarks":"The paper is a useful and careful search, but the headline occurrence rate is not yet quantitatively supported. The circularity between the injection shape and the selection cuts, the tuning on Sector 6, and the lack of propagated uncertainties are all addressable with additional experiments and analysis. I would be comfortable with publication after these points are handled; the qualitative conclusion that transiting exocomets are rare is likely safe, but the specific rate and the factor-38 contrast with Kepler should be presented as provisional."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is the first systematic exocomet transit search over the full TESS primary mission, and it is a genuinely useful piece of work. It recovers the known beta Pic event, adds five new candidates (including two around giants), and gives a completeness-corrected occurrence rate. The pipeline description is clear, the vetting is thorough (centroids, background, pixel-level checks, a reverse-time control), and the code and data are public. The authors are honest about small-number statistics and about the tentative nature of the G-type and giant-star classifications.\n\nThe main soft spot is the injection-recovery correction. The injected transits are drawn from the same skew-normal family used by the detector, so the efficiency map partly measures how well the pipeline finds its own template. The correction factors reach ~90, and without a sensitivity test against different shape families, the quoted 2.64e-4 star^-1 yr^-1 and the factor-38 contrast with Kepler are not quantitatively robust. A 30% bias in the correction shifts the rate by a lot. The qualitative conclusion—that 0.1–1% exocomet transits are rare—is probably safe, because it rests on only a handful of detections in ~15 million light curves regardless of the exact correction.\n\nThe threshold tuning on Sector 6 (which contains beta Pic) is a minor issue; the authors checked Sectors 9 and 17 as well, so the cuts are not obviously circular. The absence of propagated uncertainties is a real omission for a headline rate, but it is also consistent with how the field normally handles a few detections. The G-type hosts and the giant-star candidates are the most exciting bits, and they are also the least certain; the paper says so itself, but the conclusions in Section 6.1 lean on spectral types that are only SED-based.\n\nWho is this for? Anyone working on exocomet occurrence or planning PLATO transit searches. It is the current TESS benchmark and will be cited. It deserves a serious referee: the main request should be shape-sensitivity tests for the completeness correction and bootstrap or Poisson confidence intervals on the rate, not a desk reject.","headline":"A careful first TESS census with a handful of promising candidates; the headline rate is model-dependent but the rarity conclusion is solid.","tokens_in":31353,"tokens_out":1448,"would_cite":true,"duration_ms":20058,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Searching 15.5 million TESS light curves, this paper finds five new exocomet-like transit candidates plus the known Beta Pic event, and derives an efficiency-corrected occurrence rate of 2.64 x 10^-4 star^-1 year^-1, concluding that photome","keywords":["exocomets","TESS","transit photometry","occurrence rate","Beta Pictoris","skewed-Gaussian transit model","G-type exocomet hosts","giant-star exocomet candidates"],"falsifier":"Spectroscopic follow-up of the two probable G-type candidates (TIC 73149665 and TIC 143152957) would settle the host-star claim, and higher-precision photometry of TIC 229790952, already observed in 42 TESS sectors, could confirm whether additional shallow dips recur and distinguish a genuine exocomet from stellar variability. More directly, re-running the injection-recovery analysis with injected transits drawn from a physically motivated comet model rather than a skewed Gaussian would test the assumption on which the occurrence rate rests.","tokens_in":30564,"feed_emoji":"☄️","tokens_out":7303,"duration_ms":74538,"temperature":0.7,"pith_summary":"This paper reports a systematic search for transiting exocomets in 15.5 million TESS light curves from the primary mission. The authors recover the known exocomet transit around Beta Pic and identify five new candidate systems, concluding that photometrically transiting exocomets are extremely rare at the 0.1%-1% transit-depth level. The search extends the candidate population beyond the young A/F-type stars favored by earlier work: two candidates are consistent with G-type main-sequence hosts and two with evolved giants, the first photometric exocomet candidates around such stars. The efficiency-corrected occurrence rate is 2.64 x $10^{-4}$ $star^{-1}$ $year^{-1}$ for transits roughly 0.25% or deeper, about forty times the Kepler-based rate, a difference the authors attribute to their detection-efficiency correction and to small-number statistics.","feed_headline":"Five new exocomet candidates found in 15 million TESS stars","feed_subtitle":"Survey shows comet-like transits are rare at 0.1-1% depths, with first candidates around Sun-like and giant stars.","key_machinery":"The skewed-Gaussian transit model and its companion asymmetry ratio $\\alpha$. Each candidate event is fitted twice, once with a symmetric Gaussian and once with a skew-normal profile built from the normal density phi and its cumulative Phi with skewness $\\beta$ (Eq. 8), and the ratio $\\alpha$ of the two fit residuals (Eq. 9) decides whether the dip is asymmetric; the sign of $\\beta$ fixes the direction, with positive $\\beta$ giving the sharp-ingress, gradual-egress shape expected of a comet's trailing tail. The search uses a moving-average test statistic over window widths 0.1-2.5 days with a 2.5-day median-filter detrend, and the occurrence rate (Eq. 11) divides the detection count by the injection-recove","core_discovery":"The central claim is that single, asymmetric photometric dips, the signature of an exocomet occulting its host star with a trailing dust tail, are very rare in TESS data at the depths this survey can reach. Screening 15,477,322 light curves with an automated single-transit search and an asymmetry measure based on skewed-Gaussian fitting, the authors recover the known Beta Pic event and report five new candidates: three around main-sequence stars (including two probable G-type hosts near 5500-5700 K), one around a K-type giant, and one around a G-type supergiant. Counting these detections with injection-recovery corrections gives a combined occurrence rate of 2.64 x $10^{-4}$ $star^{-1}$ $year^{-1}$ for","pith_inferences":["If the two G-type candidates are confirmed by spectroscopy, the claimed host-star bias of exocomet detections toward young A/F stars weakens, and targeted searches around Sun-like stars become a viable strategy to test.","The occurrence-rate correction assumes real transits share the skewed-Gaussian shape family of the Beta Pic template; re-running the injection-recovery tests with physically motivated coma-and-tail profiles (rather than skew-normal curves) would directly test whether the large correction factors, up to about 90 in the faintest bins, are unbiased.","The common-time cut removed entire clusters of candidate events flagged as pipeline artifacts; if any of those were real, the true rate is higher than 2.64 x 10^-4 star^-1 year^-1, making the headline number a lower bound for all asymmetric shapes, not just the favored one.","Giant-star hosts may represent a distinct physical channel: because the sublimation radius expands as a star evolves, equal-depth transits around a larger star imply larger or more actively outgassing bodies, a scenario that multi-sector monitoring of TIC 229790952 (42 sectors available) could test by catching repeat events."],"forward_implications":["Photometrically transiting exocomets are rare: roughly six events in 15.5 million light curves, giving an efficiency-corrected occurrence rate of 2.64 x 10^-4 star^-1 year^-1 for transits about 0.25% or deeper.","Exocomet-like transits are not confined to young A/F stars: two candidates are consistent with G-type main-sequence hosts and two with evolved giants, the first photometric exocomet candidates around such stars.","Beta Pic is an outlier: its per-star occurrence rate is at least roughly 1000 times higher than every other candidate, so the system should not be treated as representative of typical stellar hosts.","The TESS rate exceeds the Kepler rate by about forty times, but the difference likely reflects small-number statistics and the efficiency correction the Kepler search lacked; a larger detection sample is needed to draw firm conclusions.","The reported rate is likely a lower bound, because the pipeline keeps only the most significant dip per light curve and would miss shallower companion transits such as the two Beta Pic events recovered only with dedicated modeling."],"supporting_citations":[{"why":"Supplies the automated single-transit search framework the TESS search is built on, plus the Kepler occurrence rate (6.7 x 10^-6 star^-1 yr^-1) used for comparison.","marker":"Kennedy et al. (2019)"},{"why":"The TESS detection of the Beta Pic exocomet that this survey recovers; its transit shape and depth set the injection-template parameters.","marker":"Zieba et al. (2019)"},{"why":"The two Kepler exocomet transit discoveries whose 0.1%-depth, multi-day transit shapes set the duration and depth priors for the search.","marker":"Rappaport et al. (2018)"},{"why":"Theoretical comet-transit shape prediction (sharp ingress, gradual tail egress) that motivates the asymmetry-based selection.","marker":"Lecavelier Des Etangs et al. (1999)"},{"why":"Describes the eleanor light-curve pipeline that produced the 30-minute data; its noise properties justify the T < 13 magnitude limit.","marker":"Feinstein et al. (2019)"},{"why":"TESS eclipsing-binary catalogue used with other catalogues to justify the 0.4-day duration cut that removes binaries and TOIs.","marker":"Kruse et al. (2021)"},{"why":"Beta Pic size distribution showing shallower transits are more common, supporting the rarity conclusion and the outlook that high-precision photometry is needed.","marker":"Lecavelier des Etangs et al. (2022)"}],"fun_headline_variants":["TESS finds five exocomet candidates, including Sun-like hosts","Exocomet transits rare at 0.1% depth, TESS search shows","First exocomet candidates around Sun-like stars found in TESS","New TESS survey spots five exocomet candidates"],"cache_read_input_tokens":2816,"weakest_assumption_plain":"The headline rate depends on measuring how many events the pipeline misses by injecting synthetic transits drawn from the same skewed-Gaussian shape model used to select candidates, with the Beta Pic event as template; if real comet transits, especially around faint stars and giants, were shaped differently or were discarded by the median filter or common-time cut, the 2.64 x $10^{-4}$ $star^{-1}$ $year^{-1}$ rate would be biased.","fun_headline_variants_meta":{"raw":{"variants":["TESS finds five exocomet candidates, including Sun-like hosts","Exocomet transits rare at 0.1% depth, TESS search shows","First exocomet candidates around Sun-like stars found in TESS","New TESS survey spots five exocomet candidates"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00087,"raw_usage":{"total_tokens":3686,"prompt_tokens":904,"completion_tokens":2782,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":648,"completion_tokens_details":{"reasoning_tokens":2714}},"tokens_in":648,"tokens_out":2782,"duration_ms":26080,"temperature":1.0,"reasoning_tokens":2714,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T23:48:50.963385+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Spectroscopic follow-up of the two probable G-type candidates (TIC 73149665 and TIC 143152957) would settle the host-star claim, and higher-precision photometry of TIC 229790952, already observed in 42 TESS sectors, could confirm whether additional shallow dips recur and distinguish a genuine exocomet from stellar variability. More directly, re-running the injection-recovery analysis with injected transits drawn from a physically motivated comet model rather than a skewed Gaussian would test the assumption on which the occurrence rate rests.","supporting_citations":[{"cited_title":"M., Hope G., Hodgkin S","cited_arxiv_id":null,"evidence_quote":"Supplies the automated single-transit search framework the TESS search is built on, plus the Kepler occurrence rate (6.7 x 10^-6 star^-1 yr^-1) used for comparison."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The two Kepler exocomet transit discoveries whose 0.1%-depth, multi-day transit shapes set the duration and depth priors for the search."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Theoretical comet-transit shape prediction (sharp ingress, gradual tail egress) that motivates the asymmetry-based selection."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"TESS eclipsing-binary catalogue used with other catalogues to justify the 0.4-day duration cut that removes binaries and TOIs."}],"review_version":1}