{"id":"33a90a9c-94d3-4730-a61f-2a0336119f26","arxiv_id":"1908.10011","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"KMT-2016-BLG-1836Lb is a newly discovered microlensing planet with host-star mass ratio q about 0.004, with Bayesian estimates of about 2.2 Jupiter masses at 3.5 AU from an M or K dwarf.","lead":"Using high-cadence KMTNet survey data, astronomers found a giant planet orbiting a small, distant star in the Galactic bulge via gravitational microlensing. The planet, roughly twice Jupiter's mass, sits beyond the snowline of its M or K dwarf host, and the discovery adds a data point to the microlensing planet census.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"2L1S vs 1L2S discrimination uses only I-band data; existing V-band light curves are not jointly fit, so the planet-vs-binary-source question is less settled than claimed.","rationale":"The paper is a competent microlensing analysis and the reader's conditional verdict is appropriate, but the single most load-bearing assumption is not the Bayesian prior choice. The central claim is that KMT-2016-BLG-1836 contains a planet, and that claim turns on excluding the binary-source model. The authors do exclude 1L2S by Delta-chi^2 ~ 38, but only in the I band. The V-band data were collected specifically for color information, yet they are not included in the 2L1S vs 1L2S comparison. A binary-source with a color difference would predict a different V-band bump than an achromatic planetary caustic, so a joint I+V fit is the direct, inexpensive test that would settle the main alternate interpretation. If that test passes, the planet discovery is robust; if it fails, the discovery is not established. The prior-dependence of the physical parameters is a real but secondary concern, already acknowledged in the text, and it only affects the derived masses and distance, not whether the anomaly is planetary. The reader identified the prior issue as the weakest assumption; my analysis points to the missing chromatic test, hence disagreement on the weakest assumption while agreeing with the conditional verdict.","tokens_in":15352,"tokens_out":16076,"duration_ms":179363,"concrete_test":"Jointly fit the 2L1S (Wide and Close) and 1L2S models to the I- and V-band light curves from all three KMTNet sites, allowing per-filter source fluxes and a free qf,V = f2,V/f1,V for the 1L2S model. Compare Delta-chi^2 and inspect V-band residuals in the anomaly window HJD' ~ 7493. If the V-band points follow the achromatic 2L1S bump, the 1L2S exclusion is strengthened; if they are better matched by a chromatic 1L2S bump, the planet interpretation is not established. Report how many V-band points fall in the anomaly window and the best-fit qf,V.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 3.2 excludes the binary-source (1L2S) model by Delta-chi^2 ~ 38 using the I-band light curve, but Section 2 states that about 10% of KMTC and 5% of KMTS images were taken in V-band. Those V-band data are used only for the CMD in Section 4.1, not in the model comparison. In the 1L2S model (Eqs. 4-5), the anomaly is produced by a second source with flux ratio qf,lambda, which is color-dependent: if the second source has a different V-I color, the predicted bump in V differs from the achromatic 2L1S prediction. Without fitting V, the Delta-chi^2 = 38 exclusion is a single-band result and does not test the chromatic signature of the degeneracy. The 1L2S best fit also has u0,2 ~ 0.002, i.e., the second source passes almost exactly behind the lens; a joint I+V fit with free qf,V is the standard check of whether this is physically plausible. Since the core of the paper's claim is that the anomaly is a planet rather than a second source, this missing color test is the most load-bearing unresolved issue. The Bayesian prior dependence, by contrast, is explicitly acknowledged and only affects the derived masses and distances, not the existence of the planetary interpretation.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper reports the discovery and characterization of the microlensing planet KMT-2016-BLG-1836Lb, with a planet-host mass ratio q ≈ 0.004. The analysis uses KMTNet high-cadence observations, mostly in I band, with a small fraction of V-band images used for color measurement. A 2L1S grid search yields four minima; the best 'Wide' solution (s ≈ 1.3) is preferred over the 'Close' solution by Δχ² ≈ 16, over the two remaining solutions by Δχ² ≳ 235, and over a 1L2S binary-source model by Δχ² ≈ 38 in I band. Because neither θ_E nor π_E is unambiguously measured, a Bayesian analysis with a Galactic model gives M_host = 0.49(+0.38/−0.25) M_Sun, M_planet = 2.2(+1.9/−1.1) M_J, D_L = 7.1(+0.8/−2.4) kpc, and projected separation r_perp = 3.5(+1.1/−0.9) AU. CFHT imaging shows that the blended light is from unrelated stars. The paper also discusses a possible 'mass ratio desert' in the published KMTNet sample, with explicit caveats about detection efficiency and publication bias.","tokens_in":15645,"tokens_out":9025,"duration_ms":85464,"significance":"If the planetary interpretation holds, this is a well-characterized super-Jovian planet found by a high-cadence survey, adding to the growing KMTNet sample and to the microlensing mass-ratio distribution. The paper has several strengths: it explicitly treats the close-wide degeneracy, tests microlensing parallax, considers the 1L2S binary-source alternative, checks the blend with CFHT images, and is transparent about the prior-dominated nature of the physical-parameter estimates and about the preliminary status of the mass-ratio desert. The core detection, the mass ratio q ≈ 0.004, is supported by a stable MCMC solution and a clear anomaly in the light curve. The main weaknesses are the absence of a V-band test of the 1L2S hypothesis and the presentation of prior-dependent physical parameters in the abstract without the caveats given in the text.","major_comments":[{"comment":"The exclusion of the binary-source (1L2S) model is a single-band result. Section 2 states that about 10% of KMTC and 5% of KMTS images are in V-band, and Section 4.1 uses those data only for the CMD; Section 3.2's Eqs. (4)-(5) define a wavelength-dependent flux ratio qf,λ, and Table 3 reports qf,I only. Because a second source with a different V-I color would produce a V-band bump different from the achromatic 2L1S prediction, the Δχ² ≈ 38 in I band does not by itself rule out 1L2S. I recommend a joint I+V fit with free qf,V, or a quantitative statement of why the V-band data cannot constrain it, especially because the 1L2S solution has u0,2 ≈ 0.002, a nearly perfect alignment of the second source.","section":"§3.2, Table 3; §2"},{"comment":"The reported physical parameters are prior-dominated. The text states that neither θ_E nor π_E is unambiguously measured, Table 1 gives only upper limits on ρ (≤ 2.0 × 10^-3 and ≤ 2.8 × 10^-3), and Table 2 shows parallax values consistent with zero within 1σ. Consequently M_host = 0.49(+0.38/−0.25) M_Sun, M_planet = 2.2(+1.9/−1.1) M_J, D_L = 7.1(+0.8/−2.4) kpc, and r_perp = 3.5(+1.1/−0.9) AU are largely outputs of the adopted Galactic model, Kroupa IMF capped at 1.3 M_Sun, and the Gaia proper-motion prior, rather than direct measurements. The Abstract presents 'super-Jovian M_planet = 2.2...' and 'beyond the snowline' without this caveat. I request either a sensitivity test of the posteriors to the priors or a clear statement in the Abstract that these values are prior-dependent; the mass ratio q ≈ 0.004 itself is not affected.","section":"§5.1, Eqs. (10)-(11), Table 4; Abstract"}],"minor_comments":[{"comment":"The labels in the sentence 'we label them by \"Close\" (solution B, s < 1) and \"Wide\" (solution A, s > 1)' are reversed relative to Table 1, where solution A has s = 0.90 and solution B has s = 1.29; swap the labels.","section":"§3.1"},{"comment":"The sentence 'we fix logq, logs, ρ = 0.001, and free t0, u0, tE, α' lists α both as a fixed grid coordinate and as a free parameter; please clarify which parameters are varied at each stage.","section":"§3.1"},{"comment":"The claim that the mass-ratio desert 'cannot be caused by the detection efficiency of KMTNet because eight planets with logq < −3.7 have been detected' is too strong, since detection efficiency in a narrow intermediate-q range need not be bracketed by detections at lower q; the subsequent acknowledgement of publication bias is more cautious and should be reflected in this sentence.","section":"§6"},{"comment":"The statement that the east component of the parallax vector is 'well constrained' is not supported by Table 2, where πE,E ≈ 0.08 ± 0.08 for W+ and comparable for other solutions; consider rephrasing to something like 'less poorly constrained than πE,N'.","section":"§3.1, Table 2"},{"comment":"The reference entry for Mróz et al. (2017) contains a stray 'and' in the author list ('Han, C., and, et al.').","section":"References"},{"comment":"The caption phrase 'Accumulate 2' should probably read 'Cumulative Δχ²' or 'Accumulated χ²'.","section":"Figure 5 caption"}],"recommendation":"major_revision","confidential_remarks":"For the editor: the V-band 1L2S test is the main load-bearing issue and is addressable with existing data, so I do not see grounds for rejection. The prior-dependence of the physical parameters is partly a presentation problem, but the abstract's unqualified numbers should be fixed before acceptance."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a solid, workmanlike microlensing paper that reports one new super-Jovian planet and floats a tentative mass-ratio desert. The planet is plausibly real, but the paper would be stronger if the authors had used their existing V-band data to test the binary-source alternative.\n\nWhat's new: the specific planet, KMT-2016-BLG-1836Lb, with q ~ 0.004, and the suggested gap in the KMTNet planet mass-ratio distribution near log q ≈ -3.7 to -3.0. Neither is revolutionary, but both are useful. The paper does the standard things well: grid search over binary-lens parameters, four local minima, close-wide degeneracy, parallax check, source characterization via CMD, and a Bayesian estimate of physical parameters. The blended-light investigation with CFHT images is a nice extra. The cumulative chi^2 plot showing the 1L2S comparison is good practice.\n\nThe main soft spot is the one the stress-test flags. The 2L1S-vs-1L2S discrimination uses I-band data only. The V-band images, about 10% of KMTC and 5% of KMTS, are used for the CMD but not included in the light-curve fits. In a 1L2S model the flux ratio is wavelength-dependent, so the V-band data carry independent information about whether a second source can explain the bump. The reported Delta chi^2 ~ 38 is convincing in I, but the test is incomplete. A joint fit with V, letting the second-source flux ratio float, is the standard way to close this. Given the second source's impact parameter is near zero, the 1L2S solution is already a bit contrived, so I suspect the planet interpretation will survive, but the paper should show it.\n\nThe physical parameters are prior-dominated, which the authors acknowledge. Without a theta_E measurement, masses and distances come largely from the Galactic model and the chosen IMF. That is normal for this kind of paper, but it means the quoted M_host and r_perp should be read as model-dependent. The mass-ratio desert is explicitly preliminary; the authors admit they have not computed detection efficiency and that publication bias is a likely culprit. It is a useful hint, not a result.\n\nOverall, this is a competent contribution. It deserves a serious referee. I would ask for the V-band joint fit, and for the desert discussion to either be backed by an efficiency calculation or kept as a one-paragraph caveat. If the V-band check confirms the planet, this is a fine addition to the microlensing census.","headline":"A competent super-Jovian microlensing discovery whose planet interpretation rests on single-band data; valuable but needs a V-band joint fit before the 1L2S alternative is closed.","tokens_in":16293,"tokens_out":2748,"would_cite":true,"duration_ms":28151,"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":"A 1-day bump in a microlensing light curve is a 2.2-Jupiter-mass planet around a small, distant star.","keywords":["gravitational microlensing","exoplanets","super-Jovian planet","mass ratio","Galactic bulge","planetary anomaly","binary-source degeneracy","high-cadence survey"],"falsifier":"High-resolution imaging of KMT-2016-BLG-1836 in the late 2020s, when the lens and source should be separated by roughly 40 milliarcseconds, can measure the lens's light directly and compare its brightness and motion with the Bayesian prediction of a roughly 0.5 solar-mass star at about 7 kiloparsecs; a disagreement would show that the prior-dominated masses are wrong.","tokens_in":15159,"feed_emoji":"🪐","tokens_out":6201,"duration_ms":63190,"temperature":0.7,"pith_summary":"The paper reports a gravitationally lensed bump in the otherwise smooth brightening of a background star, and argues that it is the signal of a planet roughly 2.2 times Jupiter's mass orbiting a low-mass star about 7 kiloparsecs away. The claim is built from high-cadence light-curve data that caught a planetary perturbation lasting about a day, with a planet-host mass ratio of $q \\sim 0.004$. The authors show that a binary-source star cannot explain the anomaly as well as a planet does, and they convert the dimensionless light-curve fit into physical masses and distance using a Galactic model. The discovery matters because it adds a cold giant planet beyond the snow line to the microlensing census, and it hints at a gap in the planet-to-star mass-ratio distribution that planet-formation theory may explain.","feed_headline":"A one-day blip reveals a 2.2-Jupiter-mass planet","feed_subtitle":"High-cadence survey data place the planet beyond the snow line of a small star 7 kiloparsecs away.","key_machinery":"The central object is the planet-host mass ratio $q$, read directly from the light curve through binary-lens microlensing theory. In this theory a planetary companion creates a caustic or cusp-approach feature whose duration scales roughly as $t_{\\rm E}\\,\\sqrt{q}$; here $q \\approx 0.004$ turns a roughly 55-day event into the observed one-day perturbation. The argument is carried by fitting competing models, with the binary-lens planetary model winning over the binary-source model by $\\Delta\\chi^2 \\approx 38$, and then by a Bayesian weighting over a Galactic model, an initial mass function, and the measured source angular radius $\\theta_*$ to convert the dimensionless fit into masses, distance, and projected separation.","core_discovery":"In the microlensing event KMT-2016-BLG-1836, a roughly one-day perturbation near the peak of an otherwise ordinary point-lens light curve is best explained as a planetary companion with mass ratio $q \\approx 0.004$ to its host star. The paper favors the 'wide' binary-lens solution over the 'close' solution, and disfavors the binary-source (1L2S) interpretation by $\\Delta\\chi^2 \\approx 38$. Adding microlens parallax does not significantly improve the fit. A Bayesian analysis that weights the surviving solutions by their $\nchi^2$ and by a Galactic model yields a host mass of $0.49^{+0.38}_{-0.25}\\,M_\\odot$, a planet mass of $2.2^{+1.9}_{-1.1}\\,M_{\\rm J}$, a distance of $7.1^{+0.8}_{-2.4}\\,{\\rm kpc}$, and a projected planet-host separation of $3.5^{+1.1}_{-0.9}\\,{\\rm AU}$, placing the planet beyond the snow line of what is probably an M or K dwarf.","pith_inferences":["Because neither the Einstein radius nor the microlens parallax is measured, the quoted masses and distance come largely from the assumed Galactic model and stellar initial mass function; the firm result is the mass ratio, not the physical masses.","The apparent 'mass-ratio desert' may well be a publication artifact, since the paper notes that incompleteness from unpublished events could create the gap; a completeness-corrected reanalysis of the same survey season would settle this without waiting for new data.","If other high-cadence microlensing events show the same gap, the feature would become a strong test of planet-formation models; if not, it will fade once selection and detection biases are modeled fully."],"forward_implications":["If the planet is real, a cold giant planet exists beyond the snow line of a small M/K dwarf star, reinforcing that such stars can host Jupiter-class planets at wide orbits.","The apparent gap in the KMTNet mass-ratio distribution near $\\log q \\approx -3.7$ to $-3.0$ would match the predicted scarcity of roughly 30 to 100 Earth-mass cores from runaway core accretion, providing a test of planet-formation theory.","High-cadence survey observations alone can discover and characterize such planets without follow-up telescopes, which improves the completeness of the microlensing planet census.","Future adaptive-optics imaging can potentially measure the lens brightness and break the remaining close/wide degeneracy, turning the prior-dominated mass and distance estimates into direct measurements.","The close/wide degeneracy leaves two possible projected separations, and a resolved lens would choose between them, pinning down the orbital configuration."],"supporting_citations":[{"why":"Establishes the scaling that planetary perturbations last roughly $t_{\\rm E}\\sqrt{q}$, which is the basis for identifying the one-day bump as planetary.","marker":"Gould & Loeb 1992"},{"why":"Describes the event-finding algorithm that flagged KMT-2016-BLG-1836 from the survey data.","marker":"Kim et al. 2018a"},{"why":"Supplies the pySIS difference-imaging photometry package used to build the light curve.","marker":"Albrow et al. 2009"},{"why":"Provides the emcee sampler used for the Markov-chain Monte Carlo model fits.","marker":"Foreman-Mackey et al. 2013"},{"why":"Supplies the Galactic model and event-rate weighting used in the Bayesian physical-parameter estimates.","marker":"Zhu et al. 2017"},{"why":"Fixes the stellar initial mass function in the Bayesian prior.","marker":"Kroupa 2001"},{"why":"Gives the binary-source (1L2S) model that is fitted and then disfavored by the data.","marker":"Gaudi 1998"},{"why":"Provides the color/surface-brightness relation used to convert the measured source color into the angular radius $\\theta_*$.","marker":"Adams et al. 2018"}],"fun_headline_variants":["Super-Jovian planet found beyond snowline of distant dwarf","Microlensing blip reveals 2.2-Jupiter-mass planet","Distant dwarf hosts super-Jovian planet beyond snowline","High-cadence microlensing uncovers a 2.2-Jupiter planet","One-day blip points to super-Jovian planet around tiny star"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The reported mass and distance are not measured directly; they come from a Bayesian average over a model of the Milky Way's stellar density, velocities, and masses, so if that model misrepresents the lens population, the physical properties of the planet shift.","fun_headline_variants_meta":{"raw":{"variants":["Super-Jovian planet found beyond snowline of distant dwarf","Microlensing blip reveals 2.2-Jupiter-mass planet","Distant dwarf hosts super-Jovian planet beyond snowline","High-cadence microlensing uncovers a 2.2-Jupiter planet","One-day blip points to super-Jovian planet around tiny star"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000705,"raw_usage":{"total_tokens":3229,"prompt_tokens":1047,"completion_tokens":2182,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":663,"completion_tokens_details":{"reasoning_tokens":2087}},"tokens_in":663,"tokens_out":2182,"duration_ms":15916,"temperature":1.0,"reasoning_tokens":2087,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:55:34.009993+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"High-resolution imaging of KMT-2016-BLG-1836 in the late 2020s, when the lens and source should be separated by roughly 40 milliarcseconds, can measure the lens's light directly and compare its brightness and motion with the Bayesian prediction of a roughly 0.5 solar-mass star at about 7 kiloparsecs; a disagreement would show that the prior-dominated masses are wrong.","supporting_citations":[{"cited_title":"2017, , 154, 210","cited_arxiv_id":null,"evidence_quote":"Supplies the Galactic model and event-rate weighting used in the Bayesian physical-parameter estimates."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the binary-source (1L2S) model that is fitted and then disfavored by the data."}],"review_version":1}