{"id":"f9c22fde-e0d4-408d-bc27-e0ed33d3897a","arxiv_id":"2507.00150","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"The paper presents 149 new hypervelocity star candidates from Gaia DR3, with follow-up spectroscopy for 23, showing that escape status and orbital origin depend strongly on the assumed Milky Way potential.","lead":"Using Gaia astrometry and new VLT spectroscopy, the authors identified 149 candidate hypervelocity stars, stars moving fast enough to leave the Milky Way, and measured radial velocities for 23 of them. The results suggest most do not come from the Galactic center and about one-third may have originated outside the Galaxy, though the classification depends heavily on the assumed Galactic potential model.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proper-motion systematics are the load-bearing risk: the paper's own §4 flag (vT significantly exceeding sqrt(2) vt) is never resolved, and without a control or independent astrometry the 149-candidate census inherits any faint-source PM bias.","rationale":"The reader's weakest assumption points to the same place: the paper's central discovery claim is only as good as the Gaia DR3 proper motions for faint, RV-less sources. The authors are explicit about the red flag but do not resolve it with an external astrometric check or a control sample; this is exactly the kind of residual risk that justifies CONDITIONAL rather than ACCEPT. I considered whether the extra-Galactic-origin classification is the more serious issue, since it rests on a permissive definition ('past orbits do not intersect the plane') and a 1 Gyr integration, but the origin claim is explicitly presented as tentative and is secondary to the candidate census. I also noted the potential dependence in §4.2: under McMillan (2017) most of the 23 sources are no longer classified as unbound. That is an important caveat, but it is quantified and disclosed, so it does not by itself require a different verdict. The missing 149-source catalog is a reproducibility problem, not a demonstrated correctness failure. The paper deserves credit for transparently reporting the sqrt(2) vt anomaly, the silver/golden split, and the potential comparison; those disclosures are why the appropriate outcome is a conditional acceptance with an astrometric verification requirement, not rejection.","tokens_in":20749,"tokens_out":11383,"duration_ms":151948,"concrete_test":"Resolve the 23 targets with independent astrometry—Gaia DR4 when available, or an external PM catalog (e.g., VLBI/HST or a re-reduction of Gaia astrometry)—and recompute vT and Pesc using the same pipeline. If the median vT drops by more than the quadrature sum of the formal uncertainties, or if any target's Pesc falls below 50% in MWPotential2014, then the PM-overestimation concern is confirmed and the 149-source census must be re-derived; if the velocities are unchanged within errors, the concern is retired.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim—149 HVS candidates and a significant sample escaping the Galaxy—depends directly on Gaia DR3 proper motions for faint sources. The candidate selection uses vT_min derived from those PMs and parallaxes, and the 'confirmation' of the 23 follow-up stars inherits the same PMs; the FORS2 radial velocities add only one velocity component. The authors flag the critical symptom in §4: 'the fact that vT significantly exceeds sqrt(2) vt for most of our targets may indicate potentially overestimated proper motion values', citing Palladino et al. (2014) and Scholz (2024). Because the sample deliberately excludes stars with Gaia radial velocities, there is no internal control population to calibrate this excess. Selection on a high measured PM (Pesc,0≥50%) preferentially retains upward astrometric noise; if faint-source PMs carry even a modest positive bias or a heavy error tail, the 149 list can be substantially contaminated and the Pesc values in Tables 2 and 3 would be systematically too high. This concern is compounded by the paper's own potential comparison: under McMillan (2017), only two (three with a bar) of the 23 targets retain Pesc≥50%, so the headline count is also model-dependent. The paper is transparent about both limitations, but the PM question is the load-bearing one because it affects every derived velocity and origin classification.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper searches for hypervelocity star (HVS) candidates in Gaia DR3 among sources lacking Gaia radial velocities. The authors apply astrometric quality cuts, estimate for each source a minimum total Galactocentric velocity by choosing the unmeasured line-of-sight velocity that minimizes vT, and define an escape probability Pesc,0 from a Monte Carlo propagation of astrometric uncertainties. They identify 149 candidates with Pesc,0 ≥ 50%, of which 23 were followed up with VLT/FORS2 spectroscopy to measure radial velocities. Using the measured RVs, they recompute escape probabilities and integrate orbits backward under several Galactic potentials, classifying the stars according to whether their past orbits cross the disk, stay in the disk, or avoid the plane, and they associate eight stars with a tentative extra-Galactic origin. The paper also estimates that 93 of the remaining candidates satisfy vt ≥ vesc under the most conservative potential considered.","tokens_in":21050,"tokens_out":4036,"duration_ms":54426,"significance":"If the candidate list is robust, the paper would significantly enlarge the census of potential unbound stars and provide valuable targets for follow-up. The main strengths are the transparent presentation of the method, the use of multiple Galactic potentials, and the acquisition of new spectroscopic radial velocities for 23 sources. The paper also ships Monte Carlo uncertainty propagation and publicly available tables. However, the central claim of 149 candidates rests on a minimum-velocity proxy and on Gaia DR3 proper motions that the authors themselves flag as potentially overestimated in Section 4. The strong dependence of the escape classification on the assumed potential—only two or three of the 23 followed-up stars retain Pesc ≥ 50% under the McMillan (2017) model—means the headline number is not a robust measure of the true unbound population.","major_comments":[{"comment":"The abstract and conclusions state that the authors identified 149 HVS candidates with Pesc > 50%, but this number is based on the minimum-velocity estimate Pesc,0 under MWPotential2014. Table 3 shows that under the McMillan (2017) potential only two (three with a bar) of the 23 spectroscopically followed stars have Pesc ≥ 50%. The central claim is therefore strongly potential-dependent and should be rephrased to state explicitly that the 149 candidates are only candidates under a particular potential and a minimum-velocity assumption, and the abstract should quote the number that survives under the most conservative potential.","section":"Section 4.2, Table 3"},{"comment":"The authors note that vT significantly exceeds √2 vt for most targets, 'may indicate potentially overestimated proper motion values', and cite Palladino et al. (2014) and Scholz (2024). This concern is never resolved. Because the selection preferentially retains sources with positive proper-motion noise, a modest systematic bias in Gaia DR3 proper motions for faint sources could substantially contaminate the 149-candidate list and bias the Pesc values in Tables 2 and 3. The paper should provide a quantitative diagnostic—for example, a comparison of Gaia DR2 and DR3 proper motions for the candidates, or a control sample of faint stars—to place an upper limit on the likely PM bias, or it should downgrade the confidence in the candidate list accordingly.","section":"Section 4, first paragraph"},{"comment":"The quantity Pesc,0 is not the probability that the true total velocity exceeds the escape velocity; it is the probability that the minimum possible total velocity (obtained by setting the unknown radial velocity to the value that minimizes vT) exceeds vesc. This is a conservative lower limit for an individual star, but it means the headline '149 candidates with Pesc > 50%' is a statement about a lower envelope, not about the actual unbound fraction. The abstract and conclusions use 'Pesc > 50%' without this qualifier, which is misleading; the qualifier 'minimum-velocity-based' should be added wherever the 149 number is quoted.","section":"Section 2, Pesc,0 definition"},{"comment":"The claim that approximately one-third of the sample may have an extra-Galactic origin is based on orbit classifications (disk crossing, plane avoidance, retrograde motion) computed under MWPotential2014 alone. Section 4.2 tests alternative potentials only for rmin and Pesc, not for the plane-crossing statistics or the disk/extra-Galactic classification. Since the classification is a central result and is qualitative, the robustness of the one-third fraction to the choice of potential should be assessed, or the claim should be explicitly limited to the adopted potential model.","section":"Section 4, 'extra-Galactic' classification"}],"minor_comments":[{"comment":"There is a typo in the Introduction: 'MV center' should read 'MW center'.","section":"Introduction"},{"comment":"The text refers to 'HV01' when discussing the fastest star; this should be 'HVS01' for consistency with Table 1.","section":"Section 4"},{"comment":"The abstract defines Pesc without stating that it is the minimum-velocity-based initial probability Pesc,0; please define the symbol in the abstract or in the first occurrence in the Introduction.","section":"Abstract and Section 2"},{"comment":"The statement that 93 of the remaining candidates have vt ≥ vesc,MM should note that this is a necessary but not sufficient condition for escape, and that vt is a lower limit to the true transverse-plus-radial velocity; the current phrasing could be read as a confirmed count.","section":"Section 4.3"},{"comment":"The exposure-time footnote in Table 1 is clear, but the caption of Figure 2 should specify whether absolute magnitudes use a distance from parallax or from a prior; the text later says distances are from parallax inversion, so this is a minor clarity issue.","section":"Section 3"}],"recommendation":"major_revision","confidential_remarks":"The paper is suitable for A&A in scope, and the new FORS2 data are a useful contribution. The main concern is that the abstract and conclusions present the 149-candidate number and the one-third extra-Galactic fraction without the strong caveats that appear in the body of the paper. The authors are transparent in Sections 4 and 4.2, but the central claims need to be brought in line with the model dependence and the unresolved proper-motion concern. I would not reject the paper; a major revision that adds a PM-bias diagnostic and rephrases the headline claims would make it publishable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Worth a read if you work on HVSs, but the headline number is softer than it looks. The genuinely new things are 23 FORS2 radial velocities for candidates that deliberately lack Gaia RVs, and a 149-source candidate list with Pesc,0 > 50%. The spectroscopy is real, and the authors deserve credit for checking escape probabilities under five different potentials, including LMC-live/frozen models. That comparison honestly shows the fragility of their main claim: under McMillan (2017), only two or three of the 23 stars retain Pesc >= 50%. They also flag the proper-motion worry themselves in Section 4, citing Palladino and Scholz, and there is no circularity because the RVs are independent of the selection.\n\nThe soft spots are real but mostly acknowledged. The load-bearing one is the unresolved proper-motion issue: the selection uses Gaia DR3 PMs for faint sources, and the vT >> sqrt(2) vt excess the authors note in Section 4 is never calibrated against anything. There is no control sample or external astrometric check, so the 149-candidate count could be substantially inflated by upward PM noise. The six targets called \"control\" are just lower-Pesc stars included by mistake, not a control for astrometric bias. The one-third extra-Galactic fraction rests on orbit classifications (no plane crossings in 1 Gyr, retrograde orbits) with no comparison to a halo-star sample, so it is suggestive but not robust. And the full 149-source catalog is not in the paper; only the 23 observed stars appear in the tables. For a candidate catalog, that is a genuine omission.\n\nWho is this for? People working on HVS statistics, Gaia astrometric systematics, or Milky Way escape-velocity modeling. It deserves a serious referee, not a desk reject, with the refereeing focused on the PM systematics and the missing catalog. A revision that either strengthens the astrometric case or explicitly demotes the list to \"needs confirmation\" would be a solid citable contribution.\n\nRecommendation: engage with it, but treat the 149 number as an upper bound until the PM issue is tackled.","headline":"A cautious HVS candidate search with real follow-up RVs, but the 149-candidate census is soft because the selection depends on the same proper motions the authors suspect are biased.","tokens_in":21603,"tokens_out":3558,"would_cite":true,"duration_ms":40140,"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":"Gaia DR3 identifies 149 new hypervelocity star candidates, and follow-up spectra show many are fast enough to leave the Milky Way.","keywords":["hypervelocity stars","Gaia DR3","stellar kinematics","Galactic potential","escape velocity","proper motions","FORS2 spectroscopy","Milky Way halo"],"falsifier":"Measure high-precision radial velocities for all 149 candidates and recompute $v_T$ with the measured $v_{\\rm los}$; if, as the paper's own caution about $v_T \\gg \\sqrt{2}\\,v_t$ suggests, a large fraction of candidates then drop below the McMillan (2017) escape velocity, the 149-candidate census would shrink substantially. A future Gaia data release providing improved proper motions for these exact faint sources would settle whether the selection was inflated by astrometric systematics.","tokens_in":20569,"feed_emoji":"⭐","tokens_out":7966,"duration_ms":85023,"temperature":0.7,"pith_summary":"The paper sets out to find hypervelocity stars, meaning stars moving faster than the Milky Way's local escape velocity, using only Gaia DR3 astrometry for objects that lack measured radial velocities. By assuming the radial velocity that minimizes each star's total Galactocentric speed, it derives conservative total velocities and assigns escape probabilities through Monte Carlo sampling of the astrometric uncertainties. This yields 149 candidates with more than a 50% chance of being unbound. Follow-up VLT/FORS2 spectroscopy of 23 candidates supplies the missing radial velocities and confirms total velocities between roughly 500 and 900 km/s, with many stars unbound depending on the adopted Galactic potential. The paper concludes that most of these stars did not come from the Galactic center and that about a third may have an extra-Galactic origin.","feed_headline":"149 new hypervelocity star candidates found in Gaia DR3","feed_subtitle":"Follow-up spectra confirm high speeds and hint that a third may come from outside the Galaxy.","key_machinery":"The load-bearing device is the minimum-total-velocity estimator: for each star, the unknown heliocentric radial velocity is set to the value $v_{{\\rm los},0}$ that minimizes the total Galactocentric speed $v_T$, making $v_T$ a lower limit that still depends on sky position. This is combined with 1,000 multivariate-Gaussian realizations that propagate Gaia parallax and proper-motion uncertainties and their correlations through the covariance matrix presented in the appendix, with the local escape velocity $v_{\\rm esc}$ computed in the MWPotential2014 Galactic potential. The fraction of realizations with $v_T > v_{\\rm esc}$ defines $P_{\\rm esc,0}$, and later FORS2 spectra replace the assumed $v_{{\\rm los},0}$ with a measured radial velocity, turning a two-dimensional tangential motion into a full three-dimensional orbit that can be integrated backward to test possible origins.","core_discovery":"On its own terms, the paper's central claim is that a conservative, astrometry-only selection from Gaia DR3 can identify a large new population of hypervelocity star candidates. For each of roughly 190 million high-quality sources without Gaia radial velocities, the authors build 1,000 Monte Carlo realizations of the astrometry and take the radial-velocity value that minimizes the total Galactocentric velocity, so the resulting $v_T$ distributions are lower limits. Comparing those distributions with the local escape velocity in the MWPotential2014 model gives an escape probability $P_{\\rm esc}$; 149 sources have $P_{\\rm esc} > 50\\%$. FORS2 spectra for 23 candidates yield the missing line-of-sight velocities and show total velocities of $500$--$900$ km/s, with most stars remaining classified as hypervelocity or high-velocity stars. Orbital backtracking over 1 Gyr finds that none of the followed-up stars, except possibly HVS07, came from within 1 kpc of the Galactic center, and about one-third have past orbits that never intersect the Galactic disk, which the authors interpret as evidence for an extra-Galactic origin.","pith_inferences":["If the proper-motion bias the authors worry about is absent, the true number of unbound stars among the 149 could approach the full list, because the minimum-total-velocity construction means every measured radial velocity can only raise $P_{\\rm esc}$.","The retrograde, very metal-poor orbits of the extra-Galactic candidates make them promising tracers of past accretion events; high-resolution spectroscopy could test whether their abundances match known dwarf-galaxy debris.","Re-running the same selection with the next Gaia data release would directly separate astrometric noise from real high velocities and would also reveal whether the preference for retrograde orbits is physical or a selection artifact."],"forward_implications":["The 149 candidates, 113 of them in the astrometrically cleaner golden subsample, expand the catalog of stars that may be unbound from the Milky Way but lack Gaia radial velocities.","Because $v_T$ was built as a lower limit, adding measured radial velocities cannot reduce a candidate's escape probability below $P_{\\rm esc,0}$, so the 23 followed-up stars retain their classifications.","None of the 23 followed-up stars has a past orbit that clearly passes within 1 kpc of the Galactic center, so the Milky Way's central black hole is not demanded as the main ejection site.","Roughly one-third of the followed-up sample has past orbits that avoid the Galactic disk entirely, which the authors read as evidence for extra-Galactic origins.","Under the most conservative alternative potential considered, 93 of the 132 remaining candidates still have transverse speeds above the local escape velocity, so further spectroscopy could substantially grow the confirmed hypervelocity star population."],"supporting_citations":[{"why":"Supplies the Gaia DR3 astrometry, including parallaxes and proper motions, on which the whole candidate search is built.","marker":"Gaia Collaboration et al. (2023b)"},{"why":"Provides the recipe for correcting Gaia parallax zero-points before distances and velocities are derived.","marker":"Lindegren et al. (2021a)"},{"why":"Defines the astrometric fidelity_v2 parameter used to filter out poor astrometric solutions.","marker":"Rybizki et al. (2022)"},{"why":"Provides the MWPotential2014 Galactic potential used to compute local escape velocities and integrate orbits.","marker":"Bovy (2015)"},{"why":"Supplies the additional astrometric reliability checks and the earlier catalog of hypervelocity candidates used for comparison.","marker":"Scholz (2024)"},{"why":"Introduces the binary-disruption ejection mechanism whose Galactic-center origin prediction the orbital backtracking tests.","marker":"Hills (1988)"},{"why":"Provides the alternative, heavier Galactic potential that gives the most conservative escape probabilities.","marker":"McMillan (2017)"},{"why":"Supplies the classification scheme that separates extra-Galactic candidates, disk-crossing stars, and in-disk stars.","marker":"Marchetti et al. (2019)"}],"fun_headline_variants":["Gaia DR3 yields 149 hypervelocity star candidates","149 new hypervelocity stars from Gaia, spectra confirm","Hypervelocity star hunt: 149 candidates, one-third from beyond","Gaia DR3 and FORS2 reveal 149 fast stars, some extragalactic","149 hypervelocity candidates, many likely escape the Milky Way"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The candidate list assumes the Gaia DR3 proper motions of these faint stars are not systematically overestimated; the paper itself warns that the observed excess of total velocities over $\\sqrt{2}$ times the transverse velocity may indicate just such a bias.","fun_headline_variants_meta":{"raw":{"variants":["Gaia DR3 yields 149 hypervelocity star candidates","149 new hypervelocity stars from Gaia, spectra confirm","Hypervelocity star hunt: 149 candidates, one-third from beyond","Gaia DR3 and FORS2 reveal 149 fast stars, some extragalactic","149 hypervelocity candidates, many likely escape the Milky Way"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000467,"raw_usage":{"total_tokens":2392,"prompt_tokens":1072,"completion_tokens":1320,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":688,"completion_tokens_details":{"reasoning_tokens":1230}},"tokens_in":688,"tokens_out":1320,"duration_ms":10401,"temperature":1.0,"reasoning_tokens":1230,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T21:22:22.214233+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure high-precision radial velocities for all 149 candidates and recompute $v_T$ with the measured $v_{\\rm los}$; if, as the paper's own caution about $v_T \\gg \\sqrt{2}\\,v_t$ suggests, a large fraction of candidates then drop below the McMillan (2017) escape velocity, the 149-candidate census would shrink substantially. A future Gaia data release providing improved proper motions for these exact faint sources would settle whether the selection was inflated by astrometric systematics.","supporting_citations":[{"cited_title":"M., Rix, H.-W., et al","cited_arxiv_id":null,"evidence_quote":"Defines the astrometric fidelity_v2 parameter used to filter out poor astrometric solutions."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the additional astrometric reliability checks and the earlier catalog of hypervelocity candidates used for comparison."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces the binary-disruption ejection mechanism whose Galactic-center origin prediction the orbital backtracking tests."},{"cited_title":"2019, MNRAS, 490, 157","cited_arxiv_id":null,"evidence_quote":"Supplies the classification scheme that separates extra-Galactic candidates, disk-crossing stars, and in-disk stars."}],"review_version":1}