{"id":"f9becb1b-b508-411b-8875-eaa8499f95ac","arxiv_id":"2509.04666","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"An open-source software package, Kete, accurately predicts positions and magnitudes of known small solar system bodies in survey images, recovering millions of archival observations.","lead":"Kete is a new open-source Python/Rust package that predicts where and how bright known asteroids and comets appear in telescope images. It runs on a laptop, handling over 11 million images from ZTF and WISE in under an hour per year.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"WISE 'recovered all observations' claim conflicts with the Section 2 diameter-subset restriction; the all-observation validation is not independently verifiable as stated.","rationale":"I agree the paper is promising and the public code is a real strength: the reproducibility, runtime, and ZTF/WISE demonstrations are valuable. But the strongest evidence—full recovery of all WISE observations—is internally contradicted by the explicit subset restriction in Section 2. The reader's weakest assumption about the two-body propagation is legitimate, but the subset issue is more direct: no amount of integrator accuracy can recover observations of asteroids omitted from the simulation. The matching criteria and magnitude offset issues noted in the reader's rationale are also real. A conditional verdict is appropriate pending clarification/correction of the WISE 'all' claim.","tokens_in":11649,"tokens_out":7402,"duration_ms":77850,"concrete_test":"Using the public kete repository and the same input catalogs, rerun the WISE simulation without the diameter-subset restriction (or, minimally, cross-match the 1.77M WISE MPC observations against the simulation's input asteroid list). Count how many MPC-reported WISE observations belong to numbered asteroids lacking a diameter in the subset. If that count is nonzero, the phrase 'recovered all WISE MPC-submitted observations' is unsupported and the abstract must be qualified. Also report the source and fraction of diameter values that were originally derived from WISE/NEOWISE data.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The load-bearing problem is an internal inconsistency in the validation. Section 2 states the WISE simulation 'utilized a subset of the numbered asteroids for which diameters are either directly measured or reliably fitted,' because NEATM requires diameters. Yet Section 3.2 says Kete 'successfully recovered all 3.89 million ZTF and 1.77 million WISE MPC-submitted observations,' and the Abstract claims prediction of 'all observations of every numbered asteroid seen by WISE.' If any WISE MPC observation belongs to a numbered asteroid excluded by the diameter subset, the 'all' claim is false. The paper gives no count of excluded asteroids or of WISE observations attached to them, no source catalog for the diameters, and no demonstration that the subset covers all observed WISE asteroids. In addition, if those diameters come from WISE/NEOWISE fits, the NEATM magnitude comparison in Fig. 4b is not independent. The two-body batch approximation noted by the reader is a secondary accuracy question; this subset issue directly undermines the headline recovery claim.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents Kete, an open-source Python/Rust package for predicting positions and magnitudes of asteroids and comets in large survey images, with applications to observation planning, pre-discovery, and dust modeling. The central demonstration is a full-year simulation for ZTF (2021) and for the WISE cryogenic phase (2010): Kete computes ephemerides and magnitudes for hundreds of thousands of numbered asteroids across more than 11 million images and claims to have recovered all 3.89 million ZTF and 1.77 million WISE MPC-submitted observations, with residuals matching the known precision of each survey. Additional sections showcase pre-discovery identification on a 1950 Palomar plate, Finson-Probstein cometary dust modeling, Monte Carlo orbital evolution of comet 103P/Hartley, and observation planning for (2) Pallas.","tokens_in":11929,"tokens_out":3680,"duration_ms":38862,"significance":"If the validation claims hold, Kete would be a significant open-source contribution: it offers a fast, high-throughput, modular tool for labeling known small bodies in current and upcoming surveys (e.g., LSST, NEO Surveyor), and the public code and documentation lower the barrier for community adoption. The paper also demonstrates a useful precovery workflow and a dust-tail model integrated with n-body propagation. The main strength is the breadth of capabilities and the claimed computational performance on commodity hardware. However, the headline 'full recovery' claim is currently under-specified and, for WISE, internally inconsistent with the diameter-subset restriction, so the validation does not yet support the abstract's strongest statements.","major_comments":[{"comment":"The WISE simulation is explicitly restricted to a 'subset of the numbered asteroids for which diameters are either directly measured or reliably fitted' (Section 2), yet Section 3.2 states that Kete 'successfully recovered all 3.89 million ZTF and 1.77 million WISE MPC-submitted observations' and the Abstract claims prediction of 'all observations of every numbered asteroid seen by WISE.' The paper gives no count of excluded numbered asteroids, no count of WISE MPC observations attached to them, and no demonstration that the diameter subset covers all WISE-observed numbered asteroids. If any MPC observation belongs to an excluded asteroid, the 'all' claim is false. Also, the source of diameters and the NEATM parameters (e.g., the beaming parameter eta) are not stated; if the diameters come from WISE/NEOWISE fits, the NEATM magnitude comparison in Fig. 4b is not independent.","section":"Section 2 vs Section 3.2 and Abstract"},{"comment":"The term 'successfully recovered' is never defined. The paper does not specify the matching criterion between predicted positions and MPC-reported astrometry: association radius or chi-square threshold, handling of multiple detections per exposure, treatment of duplicates, or how non-recovered observations were counted. Figure 3 shows residual distributions but lacks quantitative statistics (RMS, median, percentiles) and error bars. Without this information, the claim that residuals 'match the known precision for ZTF and WISE MPC submissions' cannot be verified, and the 'full recovery' number is not reproducible from the text.","section":"Section 3.2, Fig. 3"},{"comment":"The hybrid propagation strategy is central to the performance claim, but its accuracy is not quantified. Section 2 states that batches of FoVs 'do not span more than 3 days of time' and that within each batch motion is approximated by two-body dynamics, with the approximation interval treated as a tunable parameter. No test is reported comparing this batch-two-body scheme against a full n-body integration for representative orbits, including high-eccentricity NEOs or objects undergoing close encounters. Please provide the default parameter values used in the ZTF and WISE simulations and a quantitative bound (e.g., maximum and percentile astrometric error) for the two-body approximation. Otherwise, the claimed sub-arcsecond residuals cannot be attributed to the propagation algorithm.","section":"Section 2, adaptive propagation"},{"comment":"The validation uses MPC astrometry as ground truth, but the orbit catalog used by Kete is itself fitted to those same MPC observations (including ZTF and WISE data). The paper does not address the resulting potential circularity: agreement with MPC observations may largely reflect the quality of the input orbit fits rather than Kete's propagation or matching logic. A stronger validation would use withheld observations, or an independent dataset such as Gaia asteroid astrometry or stellar occultations. At minimum, the paper should explicitly discuss this limitation and quantify how much of the residual can be attributed to orbit uncertainty versus Kete's own error.","section":"Section 3.2, validation independence"}],"minor_comments":[{"comment":"The text references 'Figure 3.1' but the figure is numbered 'Figure 3'; please fix the cross-reference.","section":"Section 3.2"},{"comment":"'Astometry.net' should be 'Astrometry.net'.","section":"Section 4.1"},{"comment":"Harris 1998 appears twice as Harris 1998a and 1998b but both refer to the same Icarus article (pp. 291); similarly Burns et al. 1979a and 1979b are the same paper. Please collapse these duplicate entries.","section":"References"},{"comment":"The 'intrinsic band offsets' for ZTF magnitudes are described qualitatively but never tabulated. Please provide the offset values and state whether they were fitted or adopted from external calibrations.","section":"Fig. 4a"},{"comment":"The Abstract says 'over 11 million images' while Section 2 specifies 9,976,181 ZTF images plus WISE images; clarify whether the 11 million figure is the combined total and define the WISE image counting convention (e.g., unique sky patches vs per-band exposures).","section":"Abstract / Section 2"},{"comment":"The residual plots would benefit from showing encircled-energy radii or survey astrometric uncertainty ellipses, and from separating residuals by asteroid population (e.g., main belt vs NEO).","section":"Fig. 3"}],"recommendation":"major_revision","confidential_remarks":"The WISE subset-versus-'all observations' inconsistency is the main blocker; if the authors can show that the diameter-restricted subset covers all WISE-observed numbered asteroids (or revise the claims accordingly), and if they specify the matching criteria and error budget, the paper could become acceptable. The reliance on a co-authored prior study (Masiero et al. 2023) as the precision standard also deserves explicit acknowledgment of the lack of independence."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Kete is a real piece of engineering, not a toy. The Rust backend, batched n-body/two-body propagation, custom SPICE reader, and spherical-polygon FoV filter are all sensible, and the performance numbers (756k asteroids, 11M images on a laptop) are credible. The 1950 pre-discovery of (382632) is a nice demonstration. The code is public and the force-model appendix is detailed enough to reproduce the integrator. That is genuine credit.\n\nThe soft spot is the validation section, and the stress-test note is right: Section 2 says the WISE simulation used only numbered asteroids with measured or reliably fitted diameters, because NEATM needs a diameter. Then Section 3.2 and the abstract claim all WISE MPC-submitted observations were recovered. Unless that diameter subset contains every numbered asteroid WISE actually observed, the 'all' is false. The paper never gives the count of excluded asteroids or of WISE observations attached to them. It may be that the subset is nearly complete, but the paper doesn't show it. That's an internal inconsistency, not a minor omission.\n\nAlso, the matching criteria for 'recovered all observations' are undefined – no residual threshold, no statistics. The residual plots have no error bars. NEATM parameters (beaming, etc.) are not stated, and the magnitude comparison uses fitted band offsets, so it's not a clean prediction test. The MPC ground truth is somewhat circular – the orbit catalog was fit to those same observations – though for an ephemeris service that is an acceptable internal-consistency check, just not an independent validation.\n\nThe two-body-per-batch approximation is a secondary accuracy question. It could be quantified and isn't, but the WISE residuals look consistent with the known ~6\" PSF, so I wouldn't reject on that alone.\n\nWho gets value from this: anyone doing survey planning, pre-discovery, or labeling known objects in Rubin, NEOS, or SPHEREx data. It deserves a serious referee and probably acceptance after a minor revision that fixes the 'all' claim, defines the matching criteria, and states the thermal-model parameters. I'd cite it as a software reference and would bring it to a reading group for the performance-engineering discussion.","headline":"Genuinely useful open-source engineering, but the 'recovered all WISE observations' claim conflicts with the diameter-subset restriction in Section 2; fix the validation details and this is a solid software paper.","tokens_in":12386,"tokens_out":3942,"would_cite":true,"duration_ms":37021,"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":"The paper presents Kete, an open-source tool that predicts positions and magnitudes of known asteroids in survey images, and demonstrates full recovery of every reported asteroid observation from a year of ZTF and the cryogenic WISE mission","keywords":["Kete","asteroid ephemerides","survey simulations","small-body identification","n-body propagation","adaptive time-stepping","optical and thermal flux models","pre-discovery"],"falsifier":"Pick a set of asteroids that pass near Jupiter or have very eccentric orbits and recompute their positions at every image time using full n-body integration over the entire year; compare those positions with Kete's batched adaptive predictions. If the two-body-within-batch approximation shows differences larger than a few hundred milliarcseconds for a meaningful fraction of the sample, the unconditional recovery claim fails.","tokens_in":11593,"feed_emoji":"🔭","tokens_out":7250,"duration_ms":64581,"temperature":0.7,"pith_summary":"The paper presents Kete, an open-source tool that predicts where and how bright every known asteroid will appear in astronomical images. Its central demonstration is a full-year simulation for two telescopes: it recovered all 3.89 million ZTF and 1.77 million WISE asteroid observations submitted to the standard clearinghouse, with position and magnitude residuals matching the surveys' known precision. The claim matters because upcoming surveys will produce billions of detections, and cross-identifying known asteroids at that scale requires ephemerides computed quickly and accurately. The paper's answer is an adaptive propagation scheme that does full n-body integration between batches of images and a two-body approximation within each batch, cutting the number of integration steps by orders of magnitude.","feed_headline":"Software matches 5.66 million asteroid detections in two surveys","feed_subtitle":"Its adaptive batched propagator places 756,999 asteroids in over 11 million images with survey-level residual errors.","key_machinery":"The central mechanism is the adaptive-precision propagator: it groups field-of-view records into time-ordered batches no wider than a few days, runs a full n-body integration (planets, main-belt perturbers, general relativity, oblateness) only to the batch midpoint, then evolves each object to individual image times with a short two-body update and light-time correction. This keeps the number of costly integration steps low—about 7,100 times fewer steps for WISE's twelve-second cadence—while the geometric field-of-view inclusion test, a set of dot products against great-circle plane normals, quickly filters which of hundreds of thousands of objects appear in each image.","core_discovery":"The paper claims that a single propagation strategy can scale to survey-sized problems without losing accuracy: batch images by time, integrate the full n-body problem only to the midpoint of each batch, then approximate motion inside the batch with two-body dynamics, including light-time correction. Applied to a year of ZTF and the cryogenic phase of WISE, this reproduces every reported asteroid observation—3.89 million and 1.77 million respectively—with astrometric and photometric residuals consistent with each survey's known precision. Kete also predicts fluxes in the optical and infrared by combining an H-G phase model with thermophysical models, and uses geometric field-of-view tests ba","pith_inferences":["The two-body batch approximation is most likely to strain for asteroids in close encounters or with very high eccentricity; a targeted population-level test of those objects would show whether the survey-wide agreement holds where the dynamics are hardest.","The same batching strategy could be applied to alert streams, attaching known-object labels to candidate detections before human review, since the cost is low enough for a laptop-scale run.","The per-facet thermal model opens a route to using Kete as a forward model in shape and albedo inversion: if errors are dominated by facet geometry rather than dynamics, fitting those facets to residuals could turn the simulator into an estimator.","Because the geometry check is a simple dot-product test on spherical polygons, it should transfer directly to irregular, multi-detector footprints like those of upcoming survey missions."],"forward_implications":["Archival survey images can be systematically labeled with every known asteroid they contain, recovering pre-discovery detections such as the 1950 photographic plate where one object was identified 52 years before its official discovery.","Observation planning for small bodies becomes a local, all-sky computation: any site or telescope with defined footprints can be queried for visibility windows and magnitudes without a server-side ephemeris service.","Survey-scale validation is possible in under an hour on a laptop, so known-object labeling can become part of routine image processing rather than an offline task.","The same machinery applies to cometary dust modeling and orbital-evolution studies, making the ephemeris engine reusable outside pure detection."],"supporting_citations":[{"why":"Defines the ZTF survey whose metadata and reported observations are used as the optical validation dataset.","marker":"Bellm et al. (2018)"},{"why":"Defines the WISE mission and its infrared bands, providing the space-based validation dataset and calibration context.","marker":"Wright et al. (2010)"},{"why":"Supplies the RADAU integrator that Kete's modified propagation algorithm builds on.","marker":"Everhart (1985)"},{"why":"Supplies the IAS15 convergence criteria used to remove the predictor step and speed up the integrator.","marker":"Rein & Spiegel (2015)"},{"why":"Provides the DE440 planetary ephemeris giving positions of massive bodies during integrations.","marker":"Park et al. (2021)"},{"why":"Provides the NEATM thermal model used to predict WISE infrared magnitudes.","marker":"Harris (1998a)"},{"why":"Provides the H-G magnitude system used for optical V-band predictions and phase behavior.","marker":"Bowell et al. (1989)"},{"why":"Establishes the expected astrometric precision for ZTF and WISE submissions against which Kete's residuals are judged.","marker":"Masiero et al. (2023)"}],"fun_headline_variants":["Kete predicts 756,999 asteroids in 11M images","Matches 5.66M asteroid detections across two surveys","Open-source Kete predicts asteroid positions and magnitudes","Software places 756,999 asteroids in survey images"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The load-bearing premise is that asteroid motion within each batch of images (up to three days) is well approximated by two-body dynamics after a full n-body integration to the batch midpoint; if that approximation produces errors above a few hundred milliarcseconds for a substantial share of asteroids, the claim of recovering every reported observation would break down.","fun_headline_variants_meta":{"raw":{"variants":["Kete predicts 756,999 asteroids in 11M images","Matches 5.66M asteroid detections across two surveys","Open-source Kete predicts asteroid positions and magnitudes","Software places 756,999 asteroids in survey images"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000908,"raw_usage":{"total_tokens":3698,"prompt_tokens":659,"completion_tokens":3039,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":403,"completion_tokens_details":{"reasoning_tokens":2984}},"tokens_in":403,"tokens_out":3039,"duration_ms":23067,"temperature":1.0,"reasoning_tokens":2984,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T05:55:40.908276+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Pick a set of asteroids that pass near Jupiter or have very eccentric orbits and recompute their positions at every image time using full n-body integration over the entire year; compare those positions with Kete's batched adaptive predictions. If the two-body-within-batch approximation shows differences larger than a few hundred milliarcseconds for a meaningful fraction of the sample, the unconditional recovery claim fails.","supporting_citations":[{"cited_title":"L., Eisenhardt, P","cited_arxiv_id":null,"evidence_quote":"Defines the WISE mission and its infrared bands, providing the space-based validation dataset and calibration context."},{"cited_title":"1985, in International Astronomical Union Colloquium, Vol","cited_arxiv_id":null,"evidence_quote":"Supplies the RADAU integrator that Kete's modified propagation algorithm builds on."},{"cited_title":"& Spiegel, D","cited_arxiv_id":null,"evidence_quote":"Supplies the IAS15 convergence criteria used to remove the predictor step and speed up the integrator."},{"cited_title":"S., Folkner, W","cited_arxiv_id":null,"evidence_quote":"Provides the DE440 planetary ephemeris giving positions of massive bodies during integrations."},{"cited_title":"1989, in Asteroids II, ed","cited_arxiv_id":null,"evidence_quote":"Provides the H-G magnitude system used for optical V-band predictions and phase behavior."},{"cited_title":"R., Dahlen, D","cited_arxiv_id":null,"evidence_quote":"Establishes the expected astrometric precision for ZTF and WISE submissions against which Kete's residuals are judged."}],"review_version":1}