{"id":"f1ddfdec-0d94-4372-b604-27599807853f","arxiv_id":"2412.11616","paper_version":1,"verdict":"REJECT","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"high","formal_verification":"none","parameter_count":5,"one_line_summary":"A ZTF-based pipeline yields a list of 9 to 24 unconfirmed spider pulsar candidates near 4FGL unassociated sources, with a large search radius and no chance-coincidence analysis.","lead":"The authors cross-matched 2,157 Fermi gamma-ray sources without known counterparts against Zwicky Transient Facility light curves to find binary stars that might host millisecond 'spider' pulsars. They report 24 gold sample candidates, later narrowed to 19 and then 9, but none are confirmed and the selection lacks a false-positive control.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The gold-sample candidate list has no chance-coincidence or control-field estimate; the same cuts admit field ellipsoidal binaries and ELM-WD progenitors, so the 194/24/19/9 counts are uninterpretable as a spider-pulsar population.","rationale":"I agree with the reader's weakest assumption: the absence of a chance-coincidence or control-field analysis is the critical gap. The paper does provide real supporting evidence for the pipeline: twelve known spider pulsars are recovered with plausible periods and light-curve morphologies, which validates the period-search methodology. However, recovering known systems does not quantify false positives in the search sample. The Fermi 95% error ellipse does not cure this problem, because a 6-arcminute search radius is deliberately larger than most localization ellipses and will include many field stars; indeed, the paper itself notes that five gold-sample objects fall outside the 4FGL ellipse. The internal inclusion of a white dwarf candidate and the acknowledged overlap with evolved CVs/proto-ELM WDs reinforce the concern that the final list is contaminated. The missing control experiment is straightforward and would settle the matter: if random positions yield similar numbers of EV-type and gold-sample variables, then proximity to Fermi sources is not informative. Since the central claim is currently not supported without that measurement, I do not see a reason to alter the reader's rejection.","tokens_in":25717,"tokens_out":9158,"duration_ms":93605,"concrete_test":"Run the complete pipeline on a control set of 1,351 positions, each offset by 20-30 arcminutes in a random direction from a real 4FGL-DR3 unassociated source, or drawn from the same Galactic-latitude distribution within the ZTF footprint. Compare the numbers of EV-type variables, H-R gold-sample selections, 4FGL-ellipse matches, and RPM final candidates to the 194/24/19/9 found around the real sources. If the control fields produce a comparable fraction of these counts, the positional association is dominated by chance coincidence and the candidate list cannot support the claimed spider-pulsar interpretation.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The single most load-bearing assumption is that a ZTF variable falling within 6 arcminutes of a 4FGL-DR3 unassociated source is plausibly its counterpart. The paper never tests this. The search volume is approximately 1,351 sources times pi*(0.1 deg)^2, about 42 square degrees, and the cuts select generic ellipsoidal variability with amplitude >15%, a Gaia H-R strip, and a reduced-proper-motion boundary. Section 4.4 explicitly concedes that the H-R selection region overlaps the 'Birth of the ELMs' population of evolved CVs and proto-ELM white dwarfs, and the paper mentions but does not apply Antoniadis (2021)'s additional white-dwarf filter H_g <= 6.9(color)+13.0. In fact, the gold sample includes an object the paper itself flags as a white dwarf candidate (ZTF J2144+7714, Section 4.3). Without a background estimate, the numbers 194 EV variables, 24 gold candidates, 19 within the 4FGL ellipse, and 9 after the RPM cut are not evidence of an excess over random field binaries. The only purity argument is qualitative comparison of corner plots and cumulative period distributions (Figures 11-12), not a measured false-positive rate. Because the candidate list is the paper's principal product, this omission is load-bearing: the list's scientific value depends on knowing how many similar objects appear at random positions.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper describes a systematic search for spider pulsar (redback and black widow) candidates by combining unassociated gamma-ray sources from the 4FGL-DR3 catalog with ZTF time-series photometry. The authors cross-match 1,351 Fermi unassociated sources to ZTF targets within a 6-arcminute radius, identify 194 ellipsoidal (EV) variables and two irradiation-type binaries, apply Gaia H-R diagram and reduced proper-motion cuts to define a \"gold sample\" of spider pulsar candidates, and further select 19 objects within the 4FGL 95% confidence error ellipse and nine objects after the RPM cut. The paper also fits an orbital period-optical luminosity relation for known spider pulsars and places the gold sample on this diagram.","tokens_in":26082,"tokens_out":11932,"duration_ms":92037,"significance":"If the candidate list is reliable, it would provide a valuable target list for radio, X-ray, and optical follow-up, and the test-sample recovery of known spider systems demonstrates that the ZTF-based pipeline can detect the expected variability. The use of archival ZTF data to screen 4FGL unassociated sources is a practical and potentially productive approach. However, as written, the central claim of a meaningful candidate list is not supported by the absence of any chance-coincidence or control-field analysis, and the many internal inconsistencies in the headline numbers undermine confidence in the sample.","major_comments":[{"comment":"The search radius of 6 arcminutes adopted in Section 2.1.2 is very large relative to typical 4FGL 95% error ellipses, and the selection criteria (peak-to-peak amplitude greater than 15%, the H-R strip of Eq. 1, and the RPM boundary of Eq. 3) are broad enough to admit ordinary field binaries, evolved CVs, and proto-ELM white dwarfs. The paper never estimates the expected number of chance coincidences, for example by repeating the same selection around random sky positions or around offset 4FGL positions. Without such a control, the reported counts (194 EV-type variables, 24/19/9 gold candidates) cannot be interpreted as an excess over background, and the candidate list is not a scientifically interpretable product.","section":"Section 2.1.2 and Section 3.3"},{"comment":"The numbers that define the paper's main result are inconsistent across sections and tables. Section 2.4 reports 11 periodic variables in the test sample (5 redbacks and 6 black widows), while Table 1 and Figure 3 show 12, and the conclusion reports 12 (5 redbacks and 7 black widows). Section 3.3.1 says 22 EV-type binary stars satisfy the H-R criteria and constitute the golden sample, but the abstract, Table 2, and Section 3.3.3 say 24. For the 4FGL ellipse membership, Section 3.3.3 states 19 of 24 are inside and lists five outside, yet Table 4 appears to mark ZTFJ1957+1233 and ZTFJ1816+1747 (two of the listed five) as inside, with 21 rows carrying a Fermi ellipse checkmark. Similarly, Section 3.3.2 says the RPM cut leaves nine candidates, but Table 4 shows only seven rows with the color-Hg selection checkmark. These discrepancies make it impossible to know exactly which objects constitute the gold sample and which survive the final cuts; this must be corrected and the final list presented unambiguously.","section":"Sections 2.4, 3.3.1, 3.3.2, 3.3.3, Table 4, Conclusion"},{"comment":"Section 3.3.2 explicitly notes that Antoniadis (2021) proposed the additional white-dwarf filter Hg <= 6.9(GBP - GRP) + 13.0, but the paper does not apply it. This is pertinent because Section 4.3 reports that one gold-sample object, ZTFJ2144+7714, is a white dwarf candidate, and Section 4.4 acknowledges substantial overlap with the 'Birth of the ELMs' population of evolved CVs and proto-ELM white dwarfs. The qualitative comparisons in Figures 11-12 do not quantify the contamination rate. The authors should either apply the white-dwarf filter or provide a quantitative estimate of the expected white-dwarf/proto-ELM fraction in the gold sample.","section":"Section 3.3.2 and Section 4.3"},{"comment":"The period-finding pipeline is described qualitatively in Section 2.3, but the paper does not state the false-alarm probability threshold used to accept a Lomb-Scargle peak, the searched period range, or the minimum number of observations required. Without these, it is unclear how many of the 1,566 periodic variables are spurious detections. A quantitative statement of the detection significance is needed to make the test-sample recovery in Section 2.4 a meaningful validation of the pipeline.","section":"Section 2.3 and Section 2.4"}],"minor_comments":[{"comment":"The number of 4FGL-DR3 unassociated sources is given as 2,157 in Section 2.1.2 and as 2,179 in the conclusion; please unify the number.","section":"Section 2.1.2 vs. Section 5"},{"comment":"The sentence claiming that '22 Spider pulsars from the golden sample were matched with the pulsars detected in the TRAPUM L-band survey' is confusing, since the immediately following text reports no radio pulse detections; please clarify what was actually matched.","section":"Section 4.1"},{"comment":"There are several typographical and language issues, such as 'the faction of radio luminosity' (should be 'fraction') and the unusual spacing in the title 'T ransient F acility'; a careful language edit is recommended.","section":"Throughout"},{"comment":"Table 4 would be clearer if the checkmarks for the Fermi ellipse and color-Hg selections were explicitly placed in separate columns with unambiguous symbols; the current formatting leads to the counting ambiguities noted in the major comments.","section":"Table 4"}],"recommendation":"major_revision","confidential_remarks":"I am recommending major revision rather than acceptance because the missing chance-coincidence analysis is load-bearing for the paper's principal product. If the authors cannot provide a control or a quantitative estimate of the expected background, the paper should not proceed to publication. The internal inconsistencies in the headline numbers also suggest that the manuscript needs a careful audit before it can be considered reliable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: the paper builds a working pipeline and compiles a new candidate list, but as written the list is not interpretable because there is no chance-coincidence analysis. Searching within 6 arcminutes of 1,351 unassociated Fermi sources and accepting generic ellipsoidal variables with amplitude >15% will pick up field binaries and ELM progenitors. The paper itself shows substantial overlap with the 'Birth of the ELMs' population, and it flags one gold-sample object (ZTF J2144+7714) as a white dwarf candidate, yet it doesn't apply the Antoniadis white-dwarf filter it cites. That omission is load-bearing: the 194/24/19/9 counts are not shown to be an excess over random field stars.\n\nWhat's genuinely good: the test-sample validation on 12 known spider pulsars is credible evidence that the Lomb-Scargle pipeline recovers known orbital periods. The MCMC re-fit of the period–luminosity relation is a reasonable update of Koljonen & Linares. And the compiled table of 24 candidates with Gaia parameters and periods is a useful starting point for follow-up, even though two of the three highlighted objects (ZTF J0336+7502, ZTF J1838+3224) were already reported in earlier papers.\n\nThe soft spots, in proportion: the internal inconsistencies are real but mostly minor. The test sample is said to yield 11 periodic light curves in one place and 12 in another, with a 5+6 vs 5+7 split. More concerning, Table 2 lists ZTF J2144+7714 with a Gaia parallax of 15.65 mas (~64 pc) while Table 4 gives a median distance of 3.38 kpc; that looks like a data-handling error and casts some doubt on the other distance/luminosity estimates. The selection boundaries in Eqs. 1 and 3 are empirical lines drawn around known spiders, so the candidates are pre-screened by definition; that's a legitimate search strategy, but it means the 'gold sample' is not independently validated.\n\nBottom line: this is a useful search attempt, not a confirmed sample. The central claim that these are spider pulsar candidates is unsupported as written. The fix is straightforward: add a control-field or random-position analysis, apply the white-dwarf filter, correct the tables. With those changes the list would be a worthwhile resource. I would send it to a referee with those specific requests rather than desk-reject. For a reading group, it's a good case study in why chance-coincidence controls matter.","headline":"A working pipeline and a useful starting candidate list, but the list has no false-positive estimate and the data products have internal inconsistencies; a control sample is needed before the counts mean anything.","tokens_in":26614,"tokens_out":4308,"would_cite":false,"duration_ms":36777,"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":"The paper claims that a periodicity search of ZTF light curves around 4FGL-DR3 unassociated gamma-ray sources, followed by Gaia color-magnitude and reduced proper-motion cuts, yields 24 spider pulsar candidates, 19 inside the 4FGL 95%…","keywords":["spider pulsars","redback","black widow","millisecond pulsar","Zwicky Transient Facility","4FGL-DR3","ellipsoidal variables","gamma-ray unassociated sources"],"falsifier":"Count how many ZTF ellipsoidal variables within 6 arcminutes of random sky positions pass the same Gaia H-R and reduced proper-motion cuts; if that random expectation is comparable to the 24 gold-sample candidates, the list is not evidence of a genuine association with Fermi sources.","tokens_in":25547,"feed_emoji":"🕷️","tokens_out":7300,"duration_ms":66788,"temperature":0.7,"pith_summary":"Spider pulsars are millisecond pulsars that orbit a low-mass companion, which is heated and tidally distorted so the binary's optical light varies with the orbital period. The paper claims that a systematic periodicity search of Zwicky Transient Facility light curves around 2,157 gamma-ray sources in the 4FGL-DR3 catalog that have no identified counterpart recovers known spider systems and produces a shortlist of 24 candidate spider pulsars. Sifting the candidates by Gaia position in the Hertzsprung-Russell diagram leaves 24 gold-sample candidates, requiring position inside the 4FGL 95% error ellipse leaves 19, and applying a reduced proper-motion cut leaves 9. If the candidates are real, they would enlarge the known population of black widows and redbacks and give a targeted sample for radio, X-ray, and optical follow-up that tests the recycling model of millisecond pulsar formation.","feed_headline":"Spider pulsar hunt sifts 1,351 Fermi sources down to 9 top candidates","feed_subtitle":"Periodic ZTF stars around unassociated gamma-ray sources, filtered by Gaia, form a target list for radio and X-ray follow-up.","key_machinery":"The machinery has five components. First, a Lomb-Scargle periodogram pipeline applied to ZTF $g$- and $r$-band light curves downloaded within a 6-arcminute radius of each unassociated 4FGL-DR3 source. Second, a morphological classification that selects ellipsoidal variables (quasi-sinusoidal, two minima per cycle, more than 15 percent peak-to-peak amplitude in the $g$/$r$ bands) and irradiation-type binaries (a single sinusoidal peak per orbit). Third, the Gaia Hertzsprung-Russell selection strip defined by $M_G \\le 2.7(G_{BP}-G_{RP})+10.5$ and $M_G > 3.7(G_{BP}-G_{RP})+2.4$, which isolates the sub-main-sequence region occupied by millisecond-pulsar companions. Fourth, the reduced proper-motion cut $H_G > 5.3(G_{BP}-G_{RP})+5.9$, a kinematic filter that does not require parallax. Fifth, membership inside the 4FGL 95 percent confidence error ellipse. These filters are load-bearing: the first finds the periodic signal, the second and third reject field binaries, and the fourth removes high-proper-motion contaminants such as white dwarfs.","core_discovery":"The central claim is that spider pulsar binaries can be recognized in archival time-domain photometry without radio detection: an ellipsoidal or irradiation-modulated light curve with an orbital-period signal, located within 6 arcminutes of an unassociated Fermi gamma-ray source and lying in the Gaia sub-main-sequence band populated by millisecond-pulsar companions, is a credible spider pulsar candidate. Applying this scheme to 1,351 Fermi unassociated sources with ZTF coverage, the paper identifies 194 ellipsoidal variables and two irradiation binaries, then reduces 22 of the ellipsoidal variables plus the confirmed irradiation systems to a 24-object gold sample using the Antoniadis (2021) Gaia H-R selection strip. Requiring membership inside the 4FGL 95% confidence error ellipse cuts the sample to 19, and the color-reduced proper-motion criterion $H_G > 5.3(G_{BP}-G_{RP})+5.9$ cuts it to 9. Two of the irradiation binaries reproduce independently confirmed black widow systems, validating the pipeline. The paper also refits the orbital period-optical luminosity correlation for known spiders and shows that the gold sample's period distribution overlaps the redback and black widow distributions rather than the evolved-CV/proto-ELM white dwarf population.","pith_inferences":["A chance-coincidence check would test whether the 6-arcminute positional matches are meaningful: count comparable ZTF ellipsoidal variables with the same Gaia cuts around random sky positions, and compare with the 24 found around Fermi sources.","The nine candidates surviving the color-reduced proper-motion cut are the natural first targets for deep radio timing searches; a blind search of those positions would settle the question more directly than the archival cross-matching used here.","The same pipeline could be applied to future Fermi catalogs and to deeper time-domain surveys in the southern sky, extending the search beyond ZTF's declination limit and testing whether the candidate rate scales with survey depth."],"forward_implications":["The 24 gold-sample candidates, and particularly the 9 that survive the reduced proper-motion cut, are the targets the paper proposes for radio, X-ray, and optical-spectroscopic follow-up.","If confirmed, the candidates expand the known spider pulsar population and give new systems against which the recycling model of millisecond pulsar formation can be tested.","The refitted orbital period-optical luminosity relation predicts Gaia G-band magnitudes for candidate spiders, providing a statistical screening tool for future searches.","The gold sample's orbital-period distribution overlaps that of known redbacks and black widows and is distinct from evolved-CV/proto-ELM white dwarf systems, indicating that the selection is not dominated by those contaminants."],"supporting_citations":[{"why":"Established the protocol of searching inside 4FGL 95% error ellipses for black widow and redback analogs, providing the positional-selection foundation.","marker":"Braglia et al. (2020)"},{"why":"Supplies the Gaia Hertzsprung-Russell and color-reduced proper-motion selection boundaries for millisecond-pulsar companions that the gold-sample cuts are based on.","marker":"Antoniadis (2021)"},{"why":"Provides the orbital period-optical luminosity correlation and Gaia distance estimation method used to compute luminosities and re-fit the relation.","marker":"Koljonen & Linares (2023)"},{"why":"Demonstrated the feasibility of detecting spider-pulsar orbital periodicities in time-domain photometry, validating the search approach.","marker":"Pal et al. (2020)"},{"why":"Defines the ellipsoidal-variability selection criterion and provides the evolved-CV and proto-ELM white dwarf comparison sample.","marker":"El-Badry et al. (2021)"},{"why":"Supplies the ZTF light-curve classification scheme and prior ZTF binary-search methodology used to categorize the periodic sources.","marker":"Ren et al. (2023)"},{"why":"Confirms ZTF J0336+7502 as a black widow binary, validating the pipeline's period recovery against an independent measurement.","marker":"Li et al. (2021)"},{"why":"Confirms ZTF J1838+3224 as a flaring spider pulsar candidate, matching one of the irradiation binaries independently found in this search.","marker":"Zyuzin et al. (2024)"}],"fun_headline_variants":["ZTF and Fermi narrow spider pulsar search to 9 gold candidates","Time-domain hunt finds 9 spider pulsar candidates in Fermi data","Spider pulsar search: 1,351 sources to 9 prime targets","Gaia and ZTF sift Fermi unassociated sources for spider pulsars","9 promising spider pulsar candidates from archival surveys"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole list rests on the premise that a periodically varying star within six arcminutes of an unassociated Fermi source is plausibly that source's counterpart, yet the paper does not calculate how many such coincidences would occur by chance.","fun_headline_variants_meta":{"raw":{"variants":["ZTF and Fermi narrow spider pulsar search to 9 gold candidates","Time-domain hunt finds 9 spider pulsar candidates in Fermi data","Spider pulsar search: 1,351 sources to 9 prime targets","Gaia and ZTF sift Fermi unassociated sources for spider pulsars","9 promising spider pulsar candidates from archival surveys"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000292,"raw_usage":{"total_tokens":1782,"prompt_tokens":1099,"completion_tokens":683,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":715,"completion_tokens_details":{"reasoning_tokens":590}},"tokens_in":715,"tokens_out":683,"duration_ms":6071,"temperature":1.0,"reasoning_tokens":590,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T14:46:21.905137+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Count how many ZTF ellipsoidal variables within 6 arcminutes of random sky positions pass the same Gaia H-R and reduced proper-motion cuts; if that random expectation is comparable to the 24 gold-sample candidates, the list is not evidence of a genuine association with Fermi sources.","supporting_citations":[{"cited_title":"P., Belfiore, A., et al","cited_arxiv_id":null,"evidence_quote":"Established the protocol of searching inside 4FGL 95% error ellipses for black widow and redback analogs, providing the positional-selection foundation."},{"cited_title":"S., Tam, P","cited_arxiv_id":null,"evidence_quote":"Demonstrated the feasibility of detecting spider-pulsar orbital periodicities in time-domain photometry, validating the search approach."}],"review_version":1}