{"id":"00285aaf-f192-43fe-be9a-c6d7a6a01e05","arxiv_id":"1908.06192","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":14,"one_line_summary":"Estimates of the frequency of Earth-size planets in the habitable zone of Sun-like stars drop fourfold to eightfold when short-period small planets are excluded from the extrapolation.","lead":"This paper argues that the usual estimate for how often Sun-like stars host Earth-size planets in the habitable zone is too high, because many small, close-in planets are actually stripped cores of larger planets. Excluding those short-period small planets from the extrapolation lowers the habitable-zone planet frequency from about 40 percent to about 5 to 10 percent.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Low eta hinges on separable f(P)f(R) model that the paper's own period-dependent radius valley contradicts; a non-separable refit is needed.","rationale":"The paper is a transparent, reproducible analysis with public code and data, and the empirical drop in 1-1.8 R⊕ planets at longer periods is worth taking seriously. The reader's CONDITIONAL verdict is appropriate. My independent reading identifies the same load-bearing assumption: the separable power-law model (Eq. A1) is used to extrapolate from the fitted period range into the HZ, yet the paper's own physical motivation, the period-dependent radius valley, is a direct violation of separability. This is not merely a philosophical concern about functional form; it is a concrete internal tension. If the radius distribution changes with period, the fitted radius slope over 25-400 days need not apply at 330-800 days. Whether this bias raises or lowers eta depends on the direction of the valley movement, but the paper does not test it. The proposed non-separable refit, or a period-binned comparison of radius distributions, would settle whether the 5-10% value is robust. I therefore keep the verdict CONDITIONAL (UNCHANGED) with the same medium confidence, and do not escalate to rejection because the paper's qualitative conclusion, that short-period planets inflate eta, is supported by the data and by the plausibility of stripped cores.","tokens_in":9738,"tokens_out":6952,"duration_ms":69203,"concrete_test":"Refit the DR25+Gaia data with epos using a non-separable model in which the radius break follows the observed period dependence, R_br(P) = R_br,0 (P/P0)^-0.09 (Van Eylen et al. 2018), while keeping the same completeness and vetting treatment. Compute the posterior for eta in the 0.9-2.2 P_earth, 0.7-1.5 R_earth HZ. If the resulting eta is consistent with the 5-10% range, the separable model is not the driver; if eta rises toward 20-40%, the central claim is an artifact of the separability assumption.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative result, the fourfold-to-eightfold drop in eta to 5-10%, depends on extrapolating the occurrence surface from P < 400 days to the habitable zone (0.9-2.2 yr) using the separable broken power law dN/dlogP dlogR = A f(P) f(R) (Eq. A1). The paper itself invokes a period-dependent radius valley (Section 1, citing Van Eylen et al. 2018, R_valley ∝ P^-0.09) as evidence for stripped cores. A separable model cannot represent a radius distribution that shifts with period, so the radius slope aR fitted over 25-400 days may not describe the 330-800 day HZ. If the valley moves to smaller radii at longer periods, the 1-1.8 R⊕ population is depleted by a different mechanism than assumed, and the underlying rocky-planet population below the valley could be more numerous than the power-law extrapolation implies. The apparent drop in 1-1.8 R⊕ occurrence between 10 and 30 days (Figure 1) is precisely the kind of period-radius correlation that Eq. A1 is structurally unable to capture. Thus the quoted eta may be an artifact of imposing separability on a correlated occurrence surface, rather than a robust measurement.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper estimates the frequency of Earth-size planets in the habitable zone of Sun-like stars using the Kepler DR25 catalog with Gaia DR2 stellar parameters and the epos forward-modeling code. It fits a separable broken power-law occurrence model in period and radius over 2-400 days, then integrates the posterior over a conservative habitable zone (0.9-2.2 P⊕, 0.7-1.5 R⊕). The central comparison is between a fit that includes all short-period planets (Model#1, η⊕ ≈ 41%) and fits restricted to P > 12 or 25 days (Models#3-7, median η⊕ ≈ 5-12%). The authors interpret the drop as evidence that short-period small planets, many of which may be stripped sub-Neptune cores, bias η⊕ upward, and they propose young-cluster observations to quantify this contamination.","tokens_in":10058,"tokens_out":6803,"duration_ms":67512,"significance":"If the result holds, it materially lowers the expected yield of Earth analogues for future direct-imaging missions and sharpens the debate on the origin of the radius valley. The paper is methodologically transparent: the code is public, the MCMC setup is standard, several comparison models with different period and radius cuts are presented, and the binned inverse-efficiency occurrence rates in Figure 1 independently show that the small-planet population drops beyond ~10 days. However, the quantitative claim is sensitive to the assumed parametric form of the occurrence surface and to extrapolations beyond the fitted period range, and these dependencies are not fully quantified.","major_comments":[{"comment":"The central result depends on the separable broken power-law model dN/dlogP dlogR = A f(P) f(R). This form forces the same period dependence on all planet radii. The manuscript itself cites a period-dependent radius valley (Section 1, R_valley ∝ P^-0.09) and Figure 1 shows that the occurrence of 1-1.8 R⊕ planets drops by roughly a factor of two from the ~10-day bin to the ~30-day bin while the 1.8-3.2 R⊕ occurrence rises by ~50%. Those trends cannot be represented simultaneously by a separable model; the fitted radius slope aR is an average over the fitted period range. Since the HZ lies at 0.9-2.2 P⊕ (well beyond most of the fitted data), applying that averaged slope to Earth-size planets at long periods is not justified. Please refit with a non-separable occurrence surface (e.g., a period-dependent radius break or slope), or otherwise demonstrate that the fourfold-to-eightfold drop survives when the radius distribution is measured in period bins that overlap the HZ.","section":"Section 2.1, Eq. A1"},{"comment":"The claim of a 'fourfold to eightfold drop' is based on the median values of the posterior distributions. The 1σ uncertainties on the long-period models are large: Model#5 gives η⊕ = 5.9 +6.0 -3.5% and Model#6 gives 5.4 +7.0 -3.7%. At the upper 1σ boundary these values are ~12%, which is only a factor of ~3.3 below the Model#1 median of 40.6%; Model#4's upper bound gives a factor of ~2.5. The abstract's '~5-10%' similarly understates the posterior width. Please report the drop as a posterior distribution or with explicit uncertainty propagation, and adjust the abstract and text accordingly.","section":"Table 2"},{"comment":"All fits are restricted to orbital periods P < 400 days, but the η⊕ integral is defined over 0.9-2.2 P⊕, i.e., up to ~800 days. The period power law (bP ≈ 0.14) is therefore extrapolated by a factor of two in period, and the radius distribution is assumed to remain fixed. No test is provided for this extrapolation. I request a sensitivity check: refit the long-period models using only P > 100 days or P > 200 days and recompute η⊕, or otherwise bound the systematic error from the period extrapolation.","section":"Section 2.1 and Appendix A"}],"minor_comments":[{"comment":"The sentence in Section 2.1 describing Figure 1 ('Note that the small planets' ~30 days bin has an even higher survey completeness that the bin at 120 days...') is grammatically awkward and should be rephrased for clarity.","section":"Figure 1"},{"comment":"The discussion of 'core-powered mass loss correlates with the bolometric luminosity of the star' would benefit from a more precise statement of whether the relevant quantity is stellar luminosity or a more specific function of stellar properties.","section":"Section 3"},{"comment":"The statement that 'the typical uncertainty in planet radius is included in these Monte Carlo simulations but it is not propagated in the detection efficiency or vetting' leaves it unclear how much the reported uncertainties are affected; a brief justification or a sensitivity test would improve the presentation.","section":"Appendix A"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper does what it says: it re-computes eta_Earth from Kepler DR25+Gaia with updated completeness, and shows that excluding short-period small planets drops the value by a factor of 4-8. The binned inverse-efficiency rates in Figure 1 are the most convincing part. The 1-1.8 R⊕ occurrence rate falls between 10 and 30 days while the 1.8-3.2 R⊕ rate rises, so the short-period population genuinely looks different. That is a data statement, not a model artifact.\n\nWhat is new here is the systematic quantification. Lopez & Rice (2018) raised the concern qualitatively; this paper runs multiple epos models, different period cuts, and public code, and gives error bars. That is a useful contribution, and the citation pattern is proper.\n\nThe soft spot is the separable f(P)f(R) model in Eq. A1. The paper itself cites the period-dependent radius valley (R_valley ∝ P^-0.09), which a separable model cannot represent. The stress-test worry is fair: if the radius distribution shifts with period, then a power-law fit to 12-400 days may not describe the HZ at 330-800 days. But I do not think this kills the result. The drop survives in the binned data and in a single-power-law fit to long-period small planets. The absolute 5-10% eta should be treated with caution until a non-separable model is tried, but the direction and rough magnitude are probably right.\n\nA smaller caveat: the \"stripped cores\" framing is an interpretation. The paper shows that short-period small planets are overabundant and that this is consistent with photoevaporation or core-powered mass loss, but it does not directly identify them as stripped cores. That is okay for a letter, but readers should separate the occurrence measurement from the physical story.\n\nOverall, this is a solid, honest paper. It deserves a serious referee. I would send it to peer review and ask for a short robustness test with a non-separable occurrence model, or at least an explicit discussion of how the radius valley's period dependence would shift the extrapolated eta.\n\nWho is this for? The occurrence-rate community and anyone estimating yields for LUVOIR or HabEx. It changes the expected target counts by a lot if true, so it matters.","headline":"A transparent re-analysis that makes a strong case that short-period small planets inflate eta_Earth; the four-to-eightfold drop is real, though the absolute 5-10% value depends on a separable model that the paper itself partly undermines.","tokens_in":10584,"tokens_out":4429,"would_cite":true,"duration_ms":45295,"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":"Earth-size planets in the habitable zone of Sun-like stars may be four to eight times rarer than previously estimated, because short-period planets used in extrapolations are often stripped cores.","keywords":["eta Earth","habitable zone","Kepler","occurrence rates","stripped cores","radius valley","sub-Neptunes","exoplanet demographics"],"falsifier":"Measure the occurrence of $1-1.8\\,R_\\oplus$ planets around Sun-like stars at orbital periods of 100-400 days with independent confirmation, for example via radial-velocity follow-up or long-period transit detections with TESS or PLATO. If the occurrence per log-period bin at those periods matches the extrapolation from periods shorter than 25 days, $\\eta_\\oplus$ would be near 40%; if it matches the authors' longer-period fit, $\\eta_\\oplus$ would be near 10%.","tokens_in":9531,"feed_emoji":"🪐","tokens_out":4928,"duration_ms":43715,"temperature":0.7,"pith_summary":"The paper argues that the commonly quoted frequency of Earth-size planets in the habitable zone of Sun-like stars, $\\eta_\\oplus$, is inflated because the extrapolations lean on small short-period planets, many of which are probably sub-Neptune cores stripped by atmospheric loss. Re-fitting the Kepler DR25+Gaia sample with the same forward model but excluding planets with periods shorter than 12 or 25 days lowers $\\eta_\\oplus$ by a factor of roughly four to eight, to about 5-10%. The paper also shows that the occurrence of $1-1.8\\,R_\\oplus$ planets drops by about a factor of two between 10-day and 30-day orbits in the high-completeness regime, so the short-period population is not representative of longer-period rocky planets. If correct, this changes the expected yield of Earth analogues for future direct-imaging missions and redirects attention to quantifying stripped cores in young clusters.","feed_headline":"Stripped cores inflate Earth-twin counts fourfold","feed_subtitle":"Kepler short-period planets mislead habitable-zone estimates; true eta may be only 5-10 percent.","key_machinery":"The load-bearing object is the separable broken power-law occurrence model $dN/d\\log P\\,d\\log R = A f(P) f(R)$ (Eq. A1), implemented in the forward-modeling code epos, together with detection and vetting efficiency curves for the Kepler DR25+Gaia sample. The model assumes the period and radius distributions factor, with a broken power law in period (break near 10-12 days) and, for the full radius range, a broken power law in radius (break near 3.3 $R_\\oplus$). Fitting this model to subsets with different minimum periods isolates the influence of short-period planets; the radius-valley evidence from photoevaporation models supplies the physical reason why those planets should be excluded.","core_discovery":"The central claim is that estimates of $\\eta_\\oplus$ are biased high by the inclusion of short-period small planets. Using a separable broken power-law model fit to the Kepler DR25 catalogue with Gaia DR2 stellar radii, the authors reproduce the standard result $\\eta_\\oplus \\sim 41\\%$ when fitting periods $2-400$ days. Restricting the same fits to periods beyond 12 or 25 days, where photoevaporation should be minimal, yields $\\eta_\\oplus \\sim 4-11\\%$, a fourfold to eightfold drop. The drop is driven by the slope of the radius distribution: at short periods the inferred occurrence of planets rises steeply toward Earth radii because that population is contaminated by stripped sub-Neptune cores, whereas at longer periods the small-planet occurrence is lower and flatter.","pith_inferences":["Beyond the paper: a low $\\eta_\\oplus$ strengthens the case that atmospheric loss sculpts the radius distribution, making the period-radius correlation a central observable for demographic models.","Beyond the paper: the same stripped-core bias may affect $\\eta_\\oplus$ estimates for M and K dwarfs, where short-period small planets are also used to anchor extrapolations.","Beyond the paper: if future surveys find that Earth-size planets at 0.9-2.2 year periods are as common as the short-period extrapolation predicts, the low $\\eta_\\oplus$ claim would be ruled out; this is directly testable with TESS or PLATO long-period detections."],"forward_implications":["If $\\eta_\\oplus$ is 5-10 percent, the expected number of detectable Earth analogues around nearby Sun-like stars is several times smaller than the 20-40 percent baseline used in mission planning.","Future occurrence studies should fit periods starting beyond roughly 12-25 days, or explicitly model the stripped-core population, rather than extrapolating the full short-period sample.","Kepler's apparent lack of reliable habitable-zone candidates is consistent with a low true $\\eta_\\oplus$, not just incompleteness.","Observations of young clusters can quantify how many short-period sub-Neptunes lose their envelopes, providing a direct correction to the Kepler small-planet population."],"supporting_citations":[{"why":"Supplies the forward-modeling epos approach and the separable broken power-law fitting method that the paper modifies.","marker":"M18 (Mulders et al. 2018)"},{"why":"Provides the Kepler DR25 catalogue, the planet candidate sample the analysis fits.","marker":"Thompson et al. 2018"},{"why":"Supplies Gaia DR2 revised stellar radii used to recompute detection efficiencies and remove giants.","marker":"Berger et al. 2018"},{"why":"Defines the baseline high $\\eta_\\oplus$ estimate that the paper's low values are compared against.","marker":"Burke et al. 2015"},{"why":"Establishes the radius valley, the observed deficit near $1.8\\,R_\\oplus$ that motivates the stripped-core interpretation.","marker":"Fulton et al. 2017"},{"why":"Provides the photoevaporation theory predicting that short-period small planets include stripped sub-Neptune cores.","marker":"Owen & Wu 2013"},{"why":"Confirms the period dependence of the radius valley, supporting the claim that short-period planets are not representative.","marker":"Van Eylen et al. 2018"},{"why":"Predicted that separable power-law fits likely overestimate $\\eta_\\oplus$ because of stripped cores.","marker":"Lopez & Rice 2018"},{"why":"Provides the conservative habitable-zone boundaries used to define the integration range for $\\eta_\\oplus$.","marker":"Kopparapu et al. 2013"},{"why":"Identifies the orbital-period break near 10 days that motivates the 12-day minimum period used in the paper's main model.","marker":"Howard et al. 2012"}],"fun_headline_variants":["Stripped cores quadruple Earth-twin estimates","Habitable-zone Earth overcounts traced to stripped cores","Photoevaporation inflates Kepler's Earth census","True Earth analog rate: 5-10%, not 41%","Short-period planets skew Earth-size planet counts"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The separable broken power-law model assumes the planet radius distribution seen at short periods, or at large radii, continues unchanged at the longer periods of the habitable zone; if the radius distribution shifts with orbital period (a period-radius correlation), the extrapolation could be biased.","fun_headline_variants_meta":{"raw":{"variants":["Stripped cores quadruple Earth-twin estimates","Habitable-zone Earth overcounts traced to stripped cores","Photoevaporation inflates Kepler's Earth census","True Earth analog rate: 5-10%, not 41%","Short-period planets skew Earth-size planet counts"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000355,"raw_usage":{"total_tokens":1919,"prompt_tokens":929,"completion_tokens":990,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":545,"completion_tokens_details":{"reasoning_tokens":910}},"tokens_in":545,"tokens_out":990,"duration_ms":9264,"temperature":1.0,"reasoning_tokens":910,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:53:33.411886+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the occurrence of $1-1.8\\,R_\\oplus$ planets around Sun-like stars at orbital periods of 100-400 days with independent confirmation, for example via radial-velocity follow-up or long-period transit detections with TESS or PLATO. If the occurrence per log-period bin at those periods matches the extrapolation from periods shorter than 25 days, $\\eta_\\oplus$ would be near 40%; if it matches the authors' longer-period fit, $\\eta_\\oplus$ would be near 10%.","supporting_citations":[{"cited_title":"E., Coughlin, J","cited_arxiv_id":null,"evidence_quote":"Provides the Kepler DR25 catalogue, the planet candidate sample the analysis fits."},{"cited_title":"A., Huber, D., Gaidos, E., van Saders, J","cited_arxiv_id":null,"evidence_quote":"Supplies Gaia DR2 revised stellar radii used to recompute detection efficiencies and remove giants."},{"cited_title":"J., Christiansen, J","cited_arxiv_id":null,"evidence_quote":"Defines the baseline high $\\eta_\\oplus$ estimate that the paper's low values are compared against."},{"cited_title":"J., Petigura, E","cited_arxiv_id":null,"evidence_quote":"Establishes the radius valley, the observed deficit near $1.8\\,R_\\oplus$ that motivates the stripped-core interpretation."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the photoevaporation theory predicting that short-period small planets include stripped sub-Neptune cores."},{"cited_title":"K., Ramirez, R., Kasting, J","cited_arxiv_id":null,"evidence_quote":"Provides the conservative habitable-zone boundaries used to define the integration range for $\\eta_\\oplus$."},{"cited_title":"W., Marcy, G","cited_arxiv_id":null,"evidence_quote":"Identifies the orbital-period break near 10 days that motivates the 12-day minimum period used in the paper's main model."}],"review_version":1}