{"id":"2f77d9d8-ecc3-4075-89f6-41a4abd03912","arxiv_id":"2505.14317","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Radial-velocity monitoring of three evolved stars yields five new giant planets, plus a likely distant substellar companion, and revises the outer planet period in the HD 121056 system.","lead":"This paper reports five new massive planets found by tracking the wobble of three evolved giant stars with the CORALIE spectrograph over 12 to 18 years. It also confirms a previously known planet and updates the orbit of its outer companion, adding rare data points to the census of planets around intermediate-mass stars.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"HD 87816 c is counted as a new planet even though its period exceeds the observing baseline and the MCMC shows it is degenerate with CORALIE offsets; the central discovery count depends on a model comparison the paper does not report.","rationale":"The reader's condition on HD 87816 c is the right load-bearing concern. The paper is otherwise careful: HD 94890 b/c, HD 102888 b, and HD 121056 b/c have full or phase-complete coverage, strong periodogram FAPs, and no activity-correlation red flags. The unique weak point is that the outermost planet in the first system has P > baseline and is explicitly correlated with the zero-point offsets. The concrete test would settle whether the long-period signal survives a simpler model. This does not change the conditional verdict; it sharpens the condition. No ad hominem is intended; the authors disclose the degeneracy, but the abstract and conclusion nevertheless count HD 87816 c among five new planets.","tokens_in":20672,"tokens_out":4407,"duration_ms":48469,"concrete_test":"Re-fit the HD 87816 RVs with one Keplerian (b) plus the same three CORALIE offsets plus a linear or quadratic acceleration term, and compare it with the published two-planet model using ΔBIC and a residual generalized Lomb-Scargle periodogram (1% FAP threshold). If the one-planet-plus-acceleration model is statistically preferred or not clearly rejected (ΔBIC < 10), or if its residual periodogram retains a peak near 7000-10000 d above the 1% FAP level, then HD 87816 c is not required by the data and should be labeled a candidate, not a confirmed planet.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The discovery claim of five new planets hinges on HD 87816 c, a 12.2 M_Jup companion at P = 7596 d. Section 4.1 itself states that this period exceeds the time span of the observations, that it cannot be easily constrained, and that Fig. B.2 shows clear correlations between P_c and the three instrument offsets. The same section labels the signals 'two candidates' before the conclusion counts them as discoveries. Because the long-period signal is essentially a long-timescale curvature in the RVs, a one-planet Keplerian plus an acceleration term may reproduce the data without requiring a second companion. If so, the abstract's 'five new massive planets' is reduced to four secure detections plus one candidate. This is a correctness risk, not a disagreement with consensus: the paper's own diagnostic plots are the evidence that the offset model and the outer period are not independently constrained.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a radial-velocity analysis of four evolved stars observed with CORALIE (and, for HD 121056, also FEROS and CHIRON) over more than a decade. Using iterative periodogram analysis followed by MCMC Keplerian fitting, the authors report the discovery of five new massive planets around three stars (HD 87816 b and c, HD 94890 b and c, HD 102888 b), a distant unseen substellar companion around HD 102888, and a revised orbital period for the outer companion in the known HD 121056 system. The analysis includes checks against stellar activity indicators (FWHM, BIS, H-alpha) and a genuine out-of-fit test for HD 87816 b through a targeted periastron observing campaign.","tokens_in":20891,"tokens_out":5864,"duration_ms":59042,"significance":"If the detections hold, the paper adds five (or four secure plus one candidate) members to the relatively sparse population of planets around intermediate-mass giant stars, including a very eccentric inner planet (HD 87816 b, e=0.78) and a possible 3:1 mean-motion resonance pair (HD 94890 b and c). A notable strength is the predicted and observed periastron passage of HD 87816 b, which provides an independent validation of the orbital solution. The paper also makes its RV data publicly available and includes MCMC corner plots for all systems. The main weakness is that one of the five claimed planets, HD 87816 c, has a period longer than the observing baseline and is strongly degenerate with the instrument zero-point offsets, as the paper itself notes.","major_comments":[{"comment":"The detection of HD 87816 c is not secure under the paper's own stated criteria: its best-fit period (7596 days) exceeds the 18-year observing baseline, and the MCMC corner plot in Fig. B.2 shows clear correlations between P_c and the three CORALIE zero-point offsets. The text explicitly says that the period cannot be easily constrained and that the large uncertainties arise from this degeneracy, and it refers to the two signals as 'two candidates.' Nevertheless, the abstract and Section 5 count HD 87816 c among 'five new massive planets.' The manuscript should either provide a formal model comparison between the two-planet model and a one-planet plus acceleration (trend) model, reporting the significance (e.g., FAP or Bayesian evidence) of the second Keplerian component after accounting for the trend, or should reclassify HD 87816 c as a candidate and adjust the abstract and conclusion accordingly.","section":"Section 4.1, Table 4, Fig. B.2"},{"comment":"The 519-day FWHM periodicity that appeared after the high-cadence periastron campaign is close to the 484-day period of HD 87816 b, and the paper states that its origin remains unclear. While the authors report that the peaks are separated in frequency and that no correlation between FWHM and RVs was found, a more quantitative demonstration would be appropriate, for instance a joint fit of the RVs with the FWHM as an activity proxy or a periodogram of the RVs after conditioning on the FWHM variation. This is not a fatal concern given the out-of-fit periastron observation, but it is load-bearing for the confirmation of HD 87816 b and should be addressed explicitly.","section":"Section 4.1, Fig. A.1a"}],"minor_comments":[{"comment":"The period ratio of HD 94890 b and c is very close to 3:1, suggesting a possible mean-motion resonance. A brief dynamical stability check or an explicit statement about the expected interaction timescale would be a useful addition, although the current two-Keplerian fit is adequate for the detection claim.","section":"Section 4.2"},{"comment":"The mass estimate for the unseen companion HD 102888 c assumes a circular orbit and a companion mass much smaller than the stellar mass. These assumptions should be stated explicitly, and a short discussion of how the estimate would change for moderate eccentricities or inclinations would help the reader judge the robustness of the 'potentially substellar' classification.","section":"Section 4.3, Eq. (1)"},{"comment":"The previously published values for the period and especially the periapsis argument of HD 121056 c vary widely (P_c from about 2130 to 3920 days, omega_c from 166 to 321 degrees). A brief explanation of why the new, longer-baseline solution, with P_c=3128 d, breaks this degeneracy (e.g., a spectral window analysis or an explicit comparison of the phase coverage) would strengthen the proposed update.","section":"Table 7"},{"comment":"Minor wording issues: 'offthe' in the abstract should be 'off the', and the reference list entry for Schwab et al. (2012) contains 'V ol.' instead of 'Vol.' These do not affect the science.","section":"Abstract and Introduction"}],"recommendation":"major_revision","confidential_remarks":"The core of the paper -- the four inner planets and the verification of HD 121056 b -- appears to be on solid ground, and the periastron prediction for HD 87816 b is a genuine asset. However, the paper's headline count of 'five new massive planets' depends on HD 87816 c, whose period exceeds the baseline and which is degenerate with instrument offsets, as the authors themselves document. The manuscript is currently internally inconsistent (Section 4.1 calls the signals 'candidates' while the abstract and conclusion call them discoveries). A major revision that either supplies the missing model comparison or downgrades HD 87816 c to candidate status would resolve the concern. The journal's readership will notice the overclaim if it is not addressed."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Straight to it: this is a careful, standard RV survey paper that likely adds four secure massive planets around three evolved stars, plus a genuinely useful period update for HD 121056 c. The fifth planet, HD 87816 c, is not on the same footing, and the paper's own text admits it. I'd send it to review, but I'd ask the authors to either demote HD 87816 c to candidate status or add a formal model comparison against a one-planet-plus-acceleration/offset model.\n\nWhat is good: the pipeline is periodogram + iterative Keplerian fitting + MCMC, executed cleanly. Activity checks (FWHM, BIS, H-alpha) are done for every target. The 519-day FWHM signal in HD 87816 is handled well: they show it is separated from the 484-day RV signal and vanishes if the last 26 points are removed, which supports the planetary interpretation of the inner signal. The periastron passage of HD 87816 b was predicted and then observed at high cadence; that is a genuine out-of-fit check, rare in this literature. HD 94890 b and c have full phase coverage and very low FAPs; those look robust. The HD 121056 re-analysis uses a longer baseline than prior work, and the new 3128-day period is a real update. Data are available on DACE.\n\nThe soft spot is exactly where the stress-test lands: HD 87816 c. The period is 7596 days, the observing baseline is about 6570 days, and the paper says the period 'cannot be easily constrained' and shows correlations between P_c and the three CORALIE offsets (Fig. B.2). Section 4.1 calls it a candidate, but the abstract and conclusion count it as one of five new planets. That is an internal inconsistency. Without a comparison between the two-Keplerian model and a one-Keplerian plus acceleration or offset-drift model, the claim of five new massive planets is too strong. The low-FAP long-period peak is real, but its interpretation as a bound companion is not unique. A referee should require either reclassification or the missing model comparison.\n\nThe HD 102888 c companion is a linear trend, not a detected orbit; the authors say 'unseen companion' and give a rough mass estimate. That is fine, though it should not be called a discovery. The 3:1 resonance suggestion for HD 94890 is speculative but flagged as such. The claim that the revised HD 121056 period 'reflects the true orbital period' is slightly overconfident, but the full phase coverage supports it.\n\nThis paper is for the RV and evolved-star community. It deserves a serious referee. My recommendation: engage, but push for the reclassification or additional modeling of HD 87816 c before publication.","headline":"Four secure planets and a useful period update, with a fifth planet that is an honest but overcounted long-period signal.","tokens_in":21611,"tokens_out":3409,"would_cite":true,"duration_ms":32311,"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":"Five new massive planets orbit three evolved giant stars, and the outer planet of a fourth system gets a revised orbit.","keywords":["radial-velocity method","giant stars","evolved stars","multi-planet systems","exoplanet demographics","CORALIE spectrograph","CASCADES survey","brown dwarf companion"],"falsifier":"Continue observing HD 87816 past one full period of the claimed 7600-day companion with the current instrument configuration alone. If the signal is real, the velocity curve should close and the fitted period and offsets should remain consistent as new data arrive; if it is an offset artifact, the apparent period will drift or break coherence. A second test: recover the 519-day FWHM signal after the next periastron passage; if it correlates with the 484-day radial-velocity signal, the inner planet's interpretation would need revision.","tokens_in":20455,"feed_emoji":"🪐","tokens_out":6178,"duration_ms":52321,"temperature":0.7,"pith_summary":"This paper reports five new massive planets detected by radial-velocity monitoring of three evolved giant stars, plus a distant candidate companion and a revised orbit for an outer planet in a fourth system. After more than ten years of CORALIE measurements per target, the authors identify periodic Keplerian signals in the velocity time series of HD 87816, HD 94890, and HD 102888, check them against activity indicators such as line width, bisector, and H-alpha, and conclude that the signals are true companions rather than stellar variability or instrumental jumps. The result matters because planets around intermediate-mass stars that have left the main sequence remain rare, so each secure system helps constrain how planet formation and survival depend on stellar mass and evolution.","feed_headline":"Five giant planets found orbiting three evolved stars","feed_subtitle":"A 19-year radial-velocity campaign adds five new worlds and revises the outer orbit of a sixth system.","key_machinery":"The machinery is an iterative periodogram-plus-Keplerian pipeline: a periodogram search finds the strongest peak, a Keplerian model is fit to it, the model is subtracted, and the search repeats until no peak with false-alarm probability below 1 percent remains. The final parameters come from a Markov chain Monte Carlo sampler that fits period, semi-amplitude, mean longitude, and eccentricity-related parameters, with separate velocity zero-points for the three generations of CORALIE, per-instrument precisions, and a global stellar jitter term. The activity diagnostics serve as the main discriminator between planetary signals and stellar oscillations or convection.","core_discovery":"On the paper's terms, HD 87816 hosts an eccentric 6.7 Jupiter-mass planet at 484 days and a 12.2 Jupiter-mass companion at roughly 7600 days; HD 94890 hosts a 2.1 Jupiter-mass planet at 824 days and an 8.9 Jupiter-mass planet at 2492 days, with a period ratio close to 3 suggesting a possible 3:1 resonance; and HD 102888 hosts a 5.7 Jupiter-mass planet at 252 days plus a linear trend consistent with an unseen substellar companion at tens of AU. The paper also confirms the 89-day inner planet of HD 121056 and fixes the outer companion at about 3128 days rather than earlier values near 2130, 2200, or 3920 days. No significant periodicity or correlation with the radial velocities is found in the FWHM, bisector, or H-alpha activity indicators for any system, which the authors take as evidence that the velocity signals are planetary.","pith_inferences":["Editorial inference: if the HD 87816 c signal is real, the system joins a small set of giant-star hosts with a very wide companion, and the correlation between its period and the instrument offsets suggests the stated uncertainty may be underestimated; independent radial-velocity data from another spectrograph could settle this.","Editorial inference: the near 3:1 period ratio in HD 94890 offers a natural test of resonance capture and orbital migration around post-main-sequence stars; a dynamical fit including planet-planet interactions could constrain the orbital inclinations and true masses.","Editorial inference: the HD 102888 linear trend could be followed with future astrometric monitoring or direct imaging; a roughly 14 Jupiter-mass companion at about 20 AU should be detectable at the system's distance.","Editorial inference: all five new planets orbit outside the expected engulfment zone, consistent with the picture that close-in planets are destroyed as their stars evolve off the main sequence."],"forward_implications":["The giant-star planet population gains five members, including two multi-planet systems, strengthening demographic statistics for intermediate-mass hosts.","The HD 94890 period ratio near 3:1 suggests a possible resonant pair; if confirmed by dynamical modelling, it would be a rare resonance among evolved-star planets.","HD 102888's linear trend implies a likely substellar companion, potentially a brown dwarf of roughly 14 Jupiter masses at 20 AU, 30 at 30 AU, or 57 at 40 AU.","The revised 3128-day period for HD 121056 c, with data spanning a full orbit, resolves earlier discrepancies and gives a more reliable minimum mass near 5 Jupiter masses.","If the roughly 7600-day signal around HD 87816 c survives further monitoring, it will be one of the longest-period planets known around a giant star."],"supporting_citations":[{"why":"Supplies the CASCADES survey sample, stellar parameters, and the original survey presentation.","marker":"Ottoni et al. (2022)"},{"why":"Provides the Kepmodel package used for periodogram analysis and Keplerian fitting.","marker":"Delisle & Ségransan (2022)"},{"why":"Provides the SAMSAM MCMC sampler used to derive the final parameter posteriors.","marker":"Delisle (2022)"},{"why":"Provides the generalized Lomb-Scargle periodogram and the false-alarm probabilities used to identify significant signals.","marker":"Zechmeister & Kürster (2009)"},{"why":"Sets the expected intrinsic radial-velocity scatter of quiet giant stars used to judge whether residuals contain additional signals.","marker":"Hekker et al. (2006c)"},{"why":"Earlier detection of the HD 121056 planets that this work confirms and refines.","marker":"Jones et al. (2015b)"},{"why":"Earlier independent solution for HD 121056 used as a comparison for the revised outer period.","marker":"Wittenmyer et al. (2015)"},{"why":"Earlier solution for HD 121056 used as a comparison for the revised outer period.","marker":"Feng et al. (2022)"},{"why":"Provides the approximation used to estimate the mass of the unseen HD 102888 companion from the linear trend.","marker":"Winn et al. (2010)"}],"fun_headline_variants":["Five giant exoplanets revealed around three evolved stars","New worlds: five giant planets orbit three evolved stars","Evolved stars host five new giant planets","RV survey adds five giant planets and revises a system"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing assumption is that the long-period signal around HD 87816 is a real Keplerian companion and not an artifact of combining three generations of instrument zero-point offsets; if that offset model is wrong, or if the 519-day FWHM variation shares a stellar origin with the 484-day radial-velocity signal, the outer planet and the total planet count would be affected.","fun_headline_variants_meta":{"raw":{"variants":["Five giant exoplanets revealed around three evolved stars","New worlds: five giant planets orbit three evolved stars","Evolved stars host five new giant planets","RV survey adds five giant planets and revises a system"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00027,"raw_usage":{"total_tokens":1726,"prompt_tokens":1148,"completion_tokens":578,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":764,"completion_tokens_details":{"reasoning_tokens":516}},"tokens_in":764,"tokens_out":578,"duration_ms":24262,"temperature":1.0,"reasoning_tokens":516,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T15:36:56.357027+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Continue observing HD 87816 past one full period of the claimed 7600-day companion with the current instrument configuration alone. If the signal is real, the velocity curve should close and the fitted period and offsets should remain consistent as new data arrive; if it is an offset artifact, the apparent period will drift or break coherence. A second test: recover the 519-day FWHM signal after the next periastron passage; if it correlates with the 484-day radial-velocity signal, the inner planet's interpretation would need revision.","supporting_citations":[{"cited_title":"& Kürster, M","cited_arxiv_id":null,"evidence_quote":"Provides the generalized Lomb-Scargle periodogram and the false-alarm probabilities used to identify significant signals."},{"cited_title":"A., Wang, L., Liu, F., et al","cited_arxiv_id":null,"evidence_quote":"Earlier independent solution for HD 121056 used as a comparison for the revised outer period."},{"cited_title":"N., Johnson, J","cited_arxiv_id":null,"evidence_quote":"Provides the approximation used to estimate the mass of the unseen HD 102888 companion from the linear trend."}],"review_version":1}