{"id":"e3fc0455-8bbb-4672-be9f-0e2a91bb7934","arxiv_id":"1908.05352","paper_version":1,"verdict":"UNVERDICTED","confidence":"MODERATE","novelty_score":2.0,"correctness_risk":"high","formal_verification":"none","parameter_count":4,"one_line_summary":"A status report on a 3D maximum-likelihood method for VERITAS that models extended gamma-ray sources using a third shower-shape dimension, with no validated results presented.","lead":"Gamma-ray telescopes like VERITAS struggle to see sources larger than their field of view because standard background subtraction needs empty sky. This paper describes a maximum-likelihood method, still under development, that uses a third event-shape dimension to model extended sources, but it reports no finished validation or detection yet.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 3D likelihood factorizes spatial and MSW dimensions despite the paper's own evidence that the PSF depends on MSW; this admitted misspecification blocks the extended-source claim.","rationale":"The reader's weakest-assumption identification is exactly the load-bearing concern: the factorized model assumes spatial coordinates and MSW are uncorrelated, while the paper's own Section 3.2 demonstrates the PSF varies with MSW. This is a correctness risk, not merely an incomplete-status issue, because the admitted misspecification affects the source model used for both point-like and extended sources. The reader's UNVERDICTED verdict is appropriate: the paper is a status report with planned validation rather than a completed, validated method, and the central claim is not yet supported. I see no reason to change the verdict; the concern reinforces it. I credit the paper for transparently reporting the residual bias and the planned fix, but transparency does not remove the misspecification. The proposed concrete test would determine whether the factorized model can be salvaged or whether an MSW-dependent PSF is mandatory before any extended-source result can be trusted.","tokens_in":5906,"tokens_out":2228,"duration_ms":23034,"concrete_test":"Run a likelihood-ratio test on gamma-ray simulations: fit the current factorized model (King PSF independent of MSW, as in Eq. 3.4) against a model with PSF parameters sigma and lambda allowed to vary independently in each MSW bin (e.g., 0.8–0.9, 0.9–1.1, 1.1–1.2, 1.2–1.3) on a simulated extended source of known morphology and spectrum. If the delta-log-likelihood exceeds the Wilks threshold for the added parameters, or if best-fit morphology (source radius, flux) shifts by more than the statistical error when the MSW-PSF coupling is included, the factorized assumption is rejected and the central claim fails for the current implementation.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim—that the 3D MLM determines morphology and spectra of extended sources—rests on the model construction in Sec. 3, where the source spatial term is a product of a King-function PSF (Eq. 3.4) and an MSW distribution, based on 'the assumption that these parameters and MSW are uncorrelated' (Sec. 3). Section 3.2 reports the opposite: for Crab data, model-subtracted residuals change sign with MSW—over-subtraction of core and under-subtraction of tail for MSW 1.1–1.3, reversed for MSW < 1.1—and Fig. 4 shows the gamma-ray PSF radial distribution depends on MSW. Thus the source model is misspecified even for point sources; the spatial likelihood is not p(r, MSW) = p(r) p(MSW). Any extended-source morphology or spectrum fit with this factorized model inherits this bias, and no quantitative validation is presented: Secs. 5–6 list planned tests, with the Barlow-Beeston implementation 'currently being rewritten.' The paper explicitly concludes that the PSF dependence is 'not yet taken into account,' so the strongest claim is an aspiration rather than a demonstrated result. This is not a disagreement with external consensus; it is an internally acknowledged misspecification of the model under discussion.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper describes a three-dimensional maximum-likelihood method for VERITAS that models gamma-ray source and hadronic background events using two spatial coordinates plus the mean-scaled-width (MSW) shower parameter. The motivation is to enable analysis of extended sources, for which the standard ring-background and reflected-region methods fail because they require source-free field-of-view regions. The likelihood construction follows Mattox et al. with a King-function point-spread function (PSF), an MSW-based background model built from gamma-ray-free observations, and a binned energy treatment. The paper reports preliminary point-source consistency checks from prior work, identifies a PSF dependence on MSW that is not yet modeled, describes an ongoing Barlow-Beeston implementation to handle finite Monte Carlo statistics, and lists planned validation tests including Monte Carlo injections, the Crab, 1ES 1218+304/1ES 1215+303, dark-matter fields, and IC 443. The text is framed as a status report from ICRC2019.","tokens_in":6133,"tokens_out":5333,"duration_ms":55049,"significance":"If the method were completed and validated, it would address a genuine gap in IACT analysis: a likelihood framework capable of fitting extended sources up to the field-of-view scale and of combining multiple sources, with an alternative data-driven background model. The paper is clearly written and honestly discloses the main known limitation (MSW-dependent PSF) and the fact that validation is ongoing. However, as a journal contribution, the central claim that the method currently determines the morphology and spectra of extended sources is not supported by the evidence presented; the only quantitative-looking validation shown, the Crab residual maps, actually demonstrates a bias that is internally acknowledged. The paper is therefore more a progress report than a demonstration of a working method, and it needs either substantial model revision or a careful reframing of its claims.","major_comments":[{"comment":"The source model is explicitly built on the assumption that the spatial parameters and MSW are uncorrelated, but Section 3.2 presents direct evidence against this assumption: for Crab data the residual sky maps show over-subtraction of the core and under-subtraction of the tail for MSW 1.1-1.3, with the opposite pattern for MSW < 1.1, and Figure 4 shows that the gamma-ray PSF radial distribution depends on MSW. Since the likelihood factorizes into a spatial term and an MSW term, the current point-source PSF is misspecified, and any extended-source morphology or spectrum fitted with this model will inherit that bias. The manuscript itself concludes that the PSF dependence is \"not yet taken into account,\" so the opening claim of Section 3 that the method \"optimizes model parameters to determine the morphology and spectra of extended sources\" is not currently supported. Please either implement and validate an MSW-dependent PSF or restrict the claims to a status report with this limitation clearly stated as blocking the extended-source capability.","section":"Section 3 and Section 3.2, Eq. (3.2), Eq. (3.4)"},{"comment":"No quantitative validation of the method is presented. Section 4 states that the Barlow-Beeston likelihood calculation is \"currently being rewritten\" and that first tests will be done on bright point sources and empty fields; Section 5 describes the Monte Carlo extended-source tests, the Crab validation, and the IC 443 test only as planned. The only quantitative consistency check cited is the Cardenzana (2017) thesis comparison with standard analysis, which predates the MSW-dependent PSF issue and is not repeated or extended here. An abstract that advertises \"current status and performance\" should be accompanied by at least one closed validation result, such as the recovered morphology and spectrum of an injected extended source, or a comparison of the current implementation with standard analysis on a source such as the Crab.","section":"Sections 4 and 5"},{"comment":"The finite-statistics issue in the background model is identified as a concern, but the remedy is not yet part of the code. Since the background model is derived from observations and the paper states that the 10:1 model-to-data sample ratio will not be satisfied in many bins, the method as described currently ignores an acknowledged source of systematic uncertainty in the likelihood. Please make the status of the Barlow-Beeston implementation explicit in the abstract and conclusions, and either quantify the expected impact of the approximation or present a test of the current likelihood before claiming enhanced sensitivity to extended sources.","section":"Section 4, Eq. (4.3)"}],"minor_comments":[{"comment":"The acronym MSW is used in the opening of Section 3 before it is defined in Section 3.1; please define it at first use.","section":"Section 3"},{"comment":"The final legend entry \"(with MSW)\" appears to be a placeholder; please replace it with the actual MSW range labels corresponding to the plotted curves.","section":"Figure 4"},{"comment":"The terms \"over-subtraction\" and \"under-subtraction\" are qualitative; specifying the residual significance or event counts per MSW bin would make the reported bias measurable and reproducible.","section":"Section 3.2"},{"comment":"Please check and clarify the index conventions in Eqs. (4.1)-(4.3); in particular, the relationship between p_j, A_ji, and the sums over j and i should be stated explicitly so that the likelihood expression is unambiguous.","section":"Section 4, Eqs. (4.1)-(4.3)"},{"comment":"Reference [6] is an unpublished Ph.D. thesis; if it contains essential derivations or validation details, please provide a persistent link or make the relevant equations self-contained in this manuscript.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"This is an ICRC proceedings paper rather than a full methods article, and some of the shortfalls I identify are typical of status reports at conferences. However, the claimed capability in the title and in Section 3 is stronger than the presented evidence, and the self-admitted MSW-PSF misspecification is load-bearing. I believe the manuscript can be made publishable by reframing it as a progress report and adding at least one quantitative validation result, but in its current form the central claim is not supported. The reliance on the same-group thesis [6] for validation is defensible but should be supplemented with independently checkable results."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Nothing here is a completed result, and the authors don't pretend otherwise. This is a conference status report: the 3D likelihood method itself comes from Cardenzana's thesis (ref [6]), and the genuinely new content is a preliminary observation that the VERITAS PSF depends on mean scaled width (MSW), plus a plan to implement Barlow-Beeston. If you want a demonstration that the method works, this is not the paper.\n\nWhat it does well: the writing is clear, the motivation for extended-source analysis is convincing, and the residual maps in Sec. 3.2 are honest. Those maps show a real problem—for MSW 1.1–1.3 the model over-subtracts the core and under-subtracts the tail; below 1.1 the pattern reverses. The paper reports this rather than hiding it.\n\nThe soft spot is load-bearing. The spatial model is built on the assumption that spatial coordinates and MSW are uncorrelated (Sec. 3). Section 3.2 and Fig. 4 show the opposite: the PSF depends on MSW. So the factorized likelihood p(r, MSW) = p(r) p(MSW) is misspecified even for point sources, and any extended-source morphology or spectrum fit with this model would inherit the bias. The paper acknowledges this in the conclusion: 'a dependence of PSF on MSW not yet taken into account.' That means the abstract's claim that the method determines morphology and spectra of extended sources is not supported by the present text. It is an aspiration, not a result.\n\nThere is no quantitative validation here. Sections 4–5 list planned tests, and the Barlow-Beeston implementation is 'currently being rewritten.' The earlier agreement with standard analysis on two point sources comes from the thesis, not this paper. So the novelty over the thesis is incremental, and the one new quantitative result is a negative one.\n\nWho is this for? Someone in VERITAS or IACT methods who wants the current status. A general reader won't get much. For a journal, this would be a desk reject; for ICRC proceedings it's an acceptable status report. My recommendation: don't send it to journal review as is. If a future paper includes real validation and a model that accounts for the MSW dependence, that would be worth reading. This one is a progress note.","headline":"A candid status report, not a result: the paper's own residual maps contradict its core spatial-MSW factorization assumption, so the extended-source claim is unsupported as written.","tokens_in":6695,"tokens_out":3000,"would_cite":false,"duration_ms":26509,"reading_group":"no","serious_thinker":"yes","would_accept_peer_review":false},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proposes a 3D maximum-likelihood method that lets VERITAS measure the morphology and spectra of extended gamma-ray sources too large for standard background subtraction.","keywords":["VERITAS","3D maximum likelihood","extended gamma-ray sources","mean scaled width","point spread function","imaging atmospheric Cherenkov telescope","Barlow-Beeston method","very-high-energy gamma-ray astronomy"],"falsifier":"Analyze Crab data with the current 3D likelihood split by mean scaled width: if the paper's residual pattern—core over-subtracted for MSW 1.1-1.3 and under-subtracted below 1.1—persists after refitting the King function's $\\sigma$, then the spatial model is biased and extended-source morphologies recovered from it cannot be trusted.","tokens_in":5653,"feed_emoji":"🔭","tokens_out":10750,"duration_ms":97238,"temperature":0.7,"pith_summary":"Standard VERITAS analysis estimates the cosmic-ray background with ring or reflected-region methods that demand large source-free patches of the camera, so they fail for sources approaching the 3.5-degree field of view. This paper reports the development, for VERITAS observations taken after September 2009, of a three-dimensional maximum-likelihood method that fits source and background simultaneously in two spatial coordinates plus mean scaled width, the shower-image width variable that separates gamma-ray events from hadronic background. If the method works, VERITAS could analyze extended galactic sources such as IC 443 and the Geminga pulsar wind nebula, extracting both morphology and spectrum from 316 GeV to 5.01 TeV. The paper is a status report: previous point-source validation agrees with standard analysis, but a known point-spread function dependence on mean scaled width still has to be modeled before extended-source results can be trusted.","feed_headline":"3D likelihood fit lets VERITAS detect sources as big as its field of view","feed_subtitle":"Standard background subtraction needs empty camera patches; fitting shower width opens sources like IC 443 and Geminga.","key_machinery":"The load-bearing object is the three-dimensional likelihood itself, where the third dimension is the mean scaled width (MSW) of each air-shower image, computed by summing shower widths across telescopes and normalizing by simulated widths for that image size and impact distance. Gamma-ray events peak near MSW close to 1 while hadronic background events spread to larger values, so MSW replaces the need for a spatial OFF region. The source spatial model follows the standard likelihood formulation of equation (3.2), a convolution of intrinsic source morphology with the telescope point-spread function, energy response, and effective area, summed over reconstructed energy bins. The background MSW distribution is taken from VERITAS observations of gamma-ray-quiet fields, and the Poisson likelihood is augmented with the Barlow-Beeston method to account for the finite statistics of those model samples. The point-spread function is a symmetric King function with fitted $\\sigma$ and fixed $\\lambda$.","core_discovery":"The paper's central claim is that a maximum-likelihood analysis in three dimensions—two spatial coordinates plus mean scaled width—can estimate the morphology and spectrum of gamma-ray sources whose extent approaches the VERITAS field of view, which standard ring-background and reflected-region methods cannot do because they need a source-free patch of the camera for background estimation. The source model is built from simulations using a convolution of intrinsic source morphology with the point-spread function, energy response, and effective area, summed over energy bins, while the background model is built from deep observations of gamma-ray-quiet fields with source regions excluded. A binned-likelihood fit, with data unbinned in spatial and mean-scaled-width dimensions and binned in reconstructed energy, optimizes the model parameters. The method is presented as a work in progress: agreement with standard analysis has been shown on point sources, but full tests on known extended sources and implementation of the Barlow-Beeston correction for finite background statistics are ongoing.","pith_inferences":["The direct next step implied by the paper's own residual maps is to make the King-function point-spread parameters functions of MSW; the displayed Crab residuals suggest separate $\\sigma$ and $\\lambda$ values per MSW band would remove the over- and under-subtraction pattern.","If the MSW-dependent point-spread function is corrected, the same likelihood structure could be extended to a joint fit of VERITAS, Fermi-LAT, and HAWC data, since all three are Poisson likelihoods over spatial and energy dimensions; this would let one extended-source model be constrained from 100 MeV to tens of TeV.","A decisive end-to-end test not yet reported in the paper would be to inject simulated extended sources with known spectra into real VERITAS background fields and check that the fitted morphology and flux are unbiased in every MSW band."],"forward_implications":["Standard VERITAS analyses could be extended to sources almost as large as the telescope's 3.5-degree field of view, including supernova remnants and pulsar wind nebulae whose emission fills the camera.","Morphology and spectrum are fitted in one optimization, so the spatial shape and the spectral index no longer need separate treatments with different background regions.","The same likelihood structure can treat several overlapping sources in one field, as already demonstrated on the two point sources 1ES 1218+304 and 1ES 1215+303.","Using MSW values up to 1.3125 rather than the standard 1.1 cut gives more information on the hadronic background distribution and improves signal extraction.","Successful validation on sources such as IC 443 would connect IACT measurements to extended sources already catalogued by Fermi-LAT and HAWC, bridging the GeV and TeV bands."],"supporting_citations":[{"why":"defines the ring-background and reflected-region background models whose source-free-patch requirement is the limitation the new method must overcome.","marker":"[4]"},{"why":"supplies the likelihood formulation and source-model convolution used for the source spatial model in equation (3.2).","marker":"[5]"},{"why":"defines the original VERITAS 3D MLM framework, including energy binning, the post-September-2009 data selection, and the earlier validation on two point sources.","marker":"[6]"},{"why":"provides the finite-Monte-Carlo statistical correction that the likelihood is being rewritten to incorporate for background models built from limited observations.","marker":"[8]"},{"why":"documents the VERITAS field of view, energy range, and angular resolution that set the problem constraints the method must satisfy.","marker":"[3]"}],"fun_headline_variants":["3D fit sees gamma-ray sources as big as VERITAS field of view","No empty patches: 3D likelihood finds extended sources for VERITAS","Shower-width fitting extends VERITAS to field-filling sources","3D analysis detects extended sources in VERITAS without background holes","VERITAS 3D likelihood maps extended sources without source-free gaps"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the width of an air-shower image is statistically independent of the gamma ray's arrival direction in the camera; the paper's own Crab residual maps show the point-spread function changing with image width, and if that coupling is real the fitted source shape is biased.","fun_headline_variants_meta":{"raw":{"variants":["3D fit sees gamma-ray sources as big as VERITAS field of view","No empty patches: 3D likelihood finds extended sources for VERITAS","Shower-width fitting extends VERITAS to field-filling sources","3D analysis detects extended sources in VERITAS without background holes","VERITAS 3D likelihood maps extended sources without source-free gaps"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000315,"raw_usage":{"total_tokens":1782,"prompt_tokens":939,"completion_tokens":843,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":555,"completion_tokens_details":{"reasoning_tokens":746}},"tokens_in":555,"tokens_out":843,"duration_ms":7392,"temperature":1.0,"reasoning_tokens":746,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:16:31.347857+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Analyze Crab data with the current 3D likelihood split by mean scaled width: if the paper's residual pattern—core over-subtracted for MSW 1.1-1.3 and under-subtracted below 1.1—persists after refitting the King function's $\\sigma$, then the spatial model is biased and extended-source morphologies recovered from it cannot be trusted.","supporting_citations":[{"cited_title":"Mattox et al., The Likelihood Analysis of EGRET Data, ApJ 461, 396","cited_arxiv_id":null,"evidence_quote":"supplies the likelihood formulation and source-model convolution used for the source spatial model in equation (3.2)."},{"cited_title":"Cardenzana, A 3D Maximum Likelihood Analysis for Studying Highly Extended Sources in VERITAS Data, PhD diss., Iowa State University","cited_arxiv_id":null,"evidence_quote":"defines the original VERITAS 3D MLM framework, including energy binning, the post-September-2009 data selection, and the earlier validation on two point sources."},{"cited_title":"Barlow and C","cited_arxiv_id":null,"evidence_quote":"provides the finite-Monte-Carlo statistical correction that the likelihood is being rewritten to incorporate for background models built from limited observations."},{"cited_title":"Park, Performance of the VERITAS experiment, in proceedings of 34th ICRC, PoS (ICRC2015) 771","cited_arxiv_id":null,"evidence_quote":"documents the VERITAS field of view, energy range, and angular resolution that set the problem constraints the method must satisfy."}],"review_version":1}