{"id":"cfb0dcee-f213-4b88-b65f-f3b7b00956af","arxiv_id":"1908.03085","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"low","formal_verification":"none","parameter_count":1,"one_line_summary":"For three extreme blazars, VERITAS finds band-to-band correlations partly consistent with one-zone emission, while the TeV opacity analysis cannot rule out secondary gamma-ray production.","lead":"VERITAS telescope data on three extreme blazars were checked for correlations between very-high-energy gamma rays, X-rays, and lower-energy gamma rays over up to 11 years. The study finds the data cannot yet rule out a model where some TeV light comes from secondary particles produced along the line of sight.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The toy study in Section 3 draws a single random secondary lightcurve; one ~3σ realization does not establish that a 13% primary contribution reproduces the observed ~4σ correlation.","rationale":"The central argument is a compatibility claim: the observed ~4σ opacity-bin correlation is consistent with a secondary-gamma-ray model in which primaries contribute only 13% to the tau>2 bin. The only quantitative support is the toy study in Section 3. The text reports a single simulated secondary lightcurve and a single resulting ZDCF significance (~3σ). Because the secondary is generated as a random PSD realization, the reported significance is one sample from a distribution. A robust demonstration would require many realizations and a statement of how often the toy correlation reaches the observed level. This is not a question of model physics but of statistical support for the paper's strongest claim. The reader's weakest assumption concerned the smoothness and representativeness of the secondary PSD; that is related but distinct. Even if the PSD choice is accepted, the single-draw issue remains and is sufficient to make the toy result anecdotal. The overall conclusion that the current dataset cannot set constraints on the secondary model is a null result and is not overturned by this flaw; however, the specific '13% is sufficient' argument is not demonstrated. Therefore the reader's CONDITIONAL verdict is appropriate and unchanged.","tokens_in":8677,"tokens_out":8826,"duration_ms":92432,"concrete_test":"Repeat the Section 3 toy study with at least 1000 independent realizations of L_sim_secondary drawn from the same PSD of the flare-removed tau<1 lightcurve. For each realization, construct L_sim_tau>2 = 0.13 * L_tau<1 + 0.87 * L_sim_secondary and compute the ZDCF significance between L_sim_tau>2 and the tau<1 lightcurve. Report the median significance and the fraction of realizations reaching or exceeding the observed ~4σ. If only a small fraction (e.g., <5%) reach 4σ, the single ~3σ result does not support the claim that 13% primary is sufficient.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing step is the toy study in Section 3. The text describes constructing one simulated lightcurve, L_sim_tau>2 = 0.13 * L_tau<1 + 0.87 * L_sim_secondary, where L_sim_secondary is a single random realization drawn from the PSD of the flare-removed tau<1 lightcurve, and then reports 'Correlation at the ~3σ level is observed.' Because L_sim_secondary is stochastic, this significance is one draw from a distribution. A different realization could yield a lower or higher ZDCF significance, so the conclusion that a 13% primary contribution is 'sufficient to produce the ~4σ correlation' is not established by one realization. The paper does not report how many secondary realizations were generated, nor the spread of resulting ZDCF significances. Without that distribution, the toy result may be a chance alignment or an unrepresentative quiet secondary rather than a robust property of the 13%/87% composition. This directly affects the central claim that the observed correlation cannot rule out the secondary gamma-ray model.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This proceedings paper presents a variability study of extreme HBLs using VERITAS data, focusing on three sources: 1ES 1011+496, 1ES 1218+304, and PG 1553+113. For each source, multiwavelength light curves (VERITAS, Fermi-LAT, Swift-XRT) are analyzed with Bayesian blocks and the z-transformed discrete cross-correlation function (ZDCF), with significance estimated from Monte Carlo simulations using PSD-based light-curve generation. The novel element is an opacity-binned analysis of the VERITAS light curve of 1ES 1011+496, dividing the VHE band into τ<1, 1<τ<2, and τ>2 bins. A ~4σ correlation is found between the τ<1 and τ>2 bins. To interpret this, the authors construct a toy light curve for the τ>2 bin as 13% of the observed τ<1 light curve plus 87% of a simulated smooth secondary component, and report a ~3σ correlation with the τ<1 light curve. They conclude that the observed correlation cannot rule out the production of secondary gamma rays from ultra-high-energy proton interactions along the line of sight, and that the current dataset cannot constrain this model.","tokens_in":8888,"tokens_out":5275,"duration_ms":57436,"significance":"If the result is correct, it is a valuable null result: it shows that the observed short-timescale correlation between opacity-binned VHE light curves of 1ES 1011+496 is consistent with a secondary gamma-ray model in which primaries contribute only 13% of the highest-opacity flux. The strength of the paper is its careful, conservative wording: the central claim is explicitly a non-detection and non-constraint, and the correlation significances are estimated with a standard Monte Carlo procedure using 100,000 PSD-simulated light curves. The toy study is a transparent, first-step consistency check. The main weakness is the statistical robustness of the toy study, which currently rests on a single stochastic realization and on an unvalidated assumption about the variability properties of the secondary component.","major_comments":[{"comment":"The toy study uses a single random realization of L_sim_secondary, and the reported '~3σ correlation' is one draw from a distribution. Since L_sim_secondary is generated stochastically from a PSD, different draws will produce different ZDCF significances, and a single realization does not establish that 13% primary is 'sufficient' to reproduce the observed ~4σ correlation. The paper should simulate many realizations (e.g., 1000) and report the distribution of the resulting ZDCF significances, ideally giving the fraction of realizations that produce a correlation at or above the observed value. Without this, the robustness of the central non-constraint claim is not quantified.","section":"Section 3, toy study (Eq. L_sim_tau>2 = 0.13*L_tau<1 + 0.87*L_sim_secondary)"},{"comment":"The simulated secondary component is derived from the PSD of the flare-removed τ<1 light curve, which is assumed to represent the smooth secondary emission. However, the model cited in the paper predicts that secondary-dominated emission should show no variability on timescales shorter than ~0.1 year. If the quiescent τ<1 light curve contains significant power on shorter timescales, the simulated secondary may be more variable than physically allowed, which would bias the toy toward higher correlation with the primary and make the conclusion that the secondary model cannot be ruled out easier to reach. The authors should justify that the flare-removed PSD is dominated by long timescales, or repeat the toy with a smoothed secondary component to test whether the conclusion is robust.","section":"Section 3, construction of L_sim_secondary"}],"minor_comments":[{"comment":"The caption lists Pearson correlation coefficients between opacity bins, while the text states that correlations are tested with the ZDCF. Please clarify whether the Pearson values are shown for illustration or as an alternative statistic, and ensure the notation distinguishes the two methods.","section":"Figure 6 caption"},{"comment":"The text says a Bayesian block analysis is used to determine significant (3σ) change points, but Bayesian blocks do not directly produce a significance level. Please rephrase to avoid implying a false equivalence, for example by stating that blocks correspond to flux changes at an approximate 3σ level as estimated from the data.","section":"Section 2"},{"comment":"The EBL model adopted is Franceschini et al. (2008). Since opacity bins depend on the EBL model, it would be helpful to state whether the conclusions are sensitive to the choice of EBL model (e.g., using a different recent model such as Gilmore et al. or Domínguez et al.).","section":"Section 3"},{"comment":"The abstract states that 24 of the 39 detected AGNs are HBLs, while the introduction cites 51 HBLs detected so far by the TeV community. Clarify whether the 24 refers only to VERITAS detections, and make the numbers consistent between the abstract and the body.","section":"Abstract / Introduction"},{"comment":"The sentence 'A similar short-term variability and correlation needs to be observed with the τ > 3 bin' is grammatically awkward; consider rewording to 'If a similar short-term variability and correlation were observed in the τ > 3 bin, the model would be constrained.'","section":"Section 3, final paragraph of Conclusions"}],"recommendation":"major_revision","confidential_remarks":"This is a proceedings contribution, and the standard for statistical rigor is commensurate with that format. However, the toy study is the only quantitative support for the central non-constraint claim, and its single-realization nature is a genuine gap that can be fixed with additional simulations. I recommend major revision to require the authors to report a distribution over realizations and to discuss the PSD assumption for the secondary component. If the authors can supply a brief additional figure or table showing the spread of toy significances, the paper would be acceptable for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You can skim the first half if you have seen one VERITAS AGN proceeding: standard ZDCF, Bayesian blocks, the usual multiwavelength correlation plots. What is actually new is Section 3: splitting the 1ES 1011+496 lightcurve into three opacity bins, finding a ~4σ correlation between the τ<1 and τ>2 bins, and then a toy model indicating that a 13% primary contribution could reproduce that level of correlation under the secondary-gamma-ray hypothesis. That negative result is the real contribution: it warns the community that this dataset, with its current flare-dominated coverage, cannot constrain the UHE-proton secondary gamma-ray scenario.\n\nCredit where it is due: the analysis is honestly reported. The ZDCF significances are estimated from PSD-simulated Monte Carlo lightcurves, the Bayesian block choices are standard, and the conclusions are carefully worded as non-detections. The PG 1553+113 VHE-X-ray correlation at 3.8σ is a useful confirmation of the SSC picture for that object.\n\nThe main soft spot is exactly what the stress-test flags. The toy study appears to generate one simulated secondary lightcurve. A single realization that gives ~3σ does not establish that 13% primary is sufficient to reproduce the ~4σ correlation; you need the distribution of ZDCF significances over many realizations. The paper says ‘a simulated lightcurve is constructed’ and then ‘Correlation at the ~3σ level is observed’ — no mention of the spread or the number of draws. The wording overreaches a bit, but it is not fatal because the bottom line is genuinely “cannot rule out”. That conclusion survives even with the uncertainty, since a 3σ correlation still does not exclude the model. Still, the specific claim that 13% is sufficient to produce the ~4σ is not actually proven.\n\nA related limitation: the secondary component is assumed smooth, built from the PSD of the flare-removed τ<1 lightcurve. If secondary emission can track the flare on short timescales, the toy is not informative. The paper removes the flare to avoid introducing fast variability, but that assumption is doing real work and deserves a sensitivity test.\n\nAlso, only three of seven sources are shown, and no data tables or code are included. For a conference proceeding that is normal, but it limits scrutiny. The citation pattern looks appropriate for a collaboration proceeding; the relevant model papers are cited.\n\nWho is this for? People modeling extreme HBLs and testing cascade or secondary emission scenarios will want the opacity-bin correlation and the caveat. It is a negative result, but a useful one. I would not cite it in my own work unless I specifically needed a reference for “current VERITAS data cannot constrain this model”. If the authors expand this into a full paper, it deserves peer review — mainly to force them to report the distribution over toy realizations and to show all sources. As a proceeding, it is a solid contribution with one methodological loose end.","headline":"A modest but honest VERITAS variability report: the opacity-bin correlation result is new, and the 'cannot rule out' conclusion holds, but the toy study leans on a single stochastic realization.","tokens_in":9405,"tokens_out":2197,"would_cite":false,"duration_ms":25700,"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":"A ~4σ correlation between low- and high-opacity VERITAS light curves of 1ES 1011+496 is consistent with secondary gamma-ray production, because a toy model with 13% primary emission reproduces it.","keywords":["blazar variability","very-high-energy gamma rays","extreme HBL","VERITAS","opacity binning","secondary gamma rays","ultra-high-energy protons","cross-correlation"],"falsifier":"Look at the $\\tau>3$ light curve of 1ES 1011+496 around a flare like February 2014. The secondary-gamma-ray model predicts a primary fraction below 5% in that bin, so a significant correlation between the $\\tau>3$ and $\\tau<1$ light curves would refute the paper's conclusion that the current data cannot constrain the model; a smooth, uncorrelated $\\tau>3$ light curve would confirm it.","tokens_in":8471,"feed_emoji":"🔭","tokens_out":9618,"duration_ms":85639,"temperature":0.7,"pith_summary":"VERITAS observations spanning up to 11 years are used to test whether very-high-energy gamma rays from extreme BL Lac objects could be produced away from the source, as secondary particles from ultra-high-energy protons interacting with background photons. The paper's central demonstration is about 1ES 1011+496: the ~4σ correlation between low-opacity and high-opacity TeV light curves, which at first looks like primary emission dominating all energies, is reproduced by a toy model in which only 13% of the high-opacity flux is primary and 87% is smooth secondary emission. Because of this, the paper concludes that the current dataset cannot constrain the ultra-high-energy proton scenario. Along the way, the study also reports a 3.8σ VHE–X-ray correlation for PG 1553+113 and no VHE–X-ray correlation for 1ES 1218+304, using both to weigh one-zone synchrotron self-Compton interpretations. The reason to care is that a convincing rejection of the secondary-emission model would close a major alternative explanation for the highest-energy gamma rays seen from distant blazars.","feed_headline":"Blazar's 4σ correlation can't rule out proton-produced gamma rays","feed_subtitle":"Toy light curve with 13% primary and 87% smooth secondary emission reproduces the correlation, so the model stays open.","key_machinery":"The machinery is the z-transformed discrete cross-correlation function (ZDCF) applied to sparsely sampled light curves, together with an opacity-binning scheme. Light curves of each source are split by $\\tau$ bins set by the Franceschini et al. (2008) EBL model, and the ZDCF measures correlation versus lag, with significance estimated from Monte Carlo light curves generated from each light curve's power spectral density (PSD). The decisive object is the toy construction $L_{\\tau>2}^{\\rm sim} = 0.13\\,L_{\\tau<1}^{\\rm primary} + 0.87\\,L_{\\rm secondary}^{\\rm sim}$: the primary piece is the real $\\tau<1$ data, and the secondary piece is a PSD realisation of the same light curve after the flare is removed, so that it is smooth. The toy shows that a small primary flare contribution can imprint a correlation even when most of the high-opacity flux is secondary.","core_discovery":"The paper claims that the observed correlations between the opacity-binned VERITAS light curves of 1ES 1011+496 — $\\tau<1$, $1<\\tau<2$, and $\\tau>2$, where $\\tau$ is the opacity of the extragalactic background light — do not discriminate between primary and secondary gamma-ray production. The February 2014 flare appears in all three bins and the light curves are correlated at about $4\\sigma$. Yet a synthetic $\\tau>2$ light curve built from $0.13\\,L_{\\tau<1}^{\\rm primary} + 0.87\\,L_{\\rm secondary}^{\\rm sim}$, with a smooth secondary component simulated from the flare-removed $\\tau<1$ power spectrum, reproduces the correlation at the $\\sim3\\sigma$ level. The paper therefore concludes that the data cannot rule out the model in which ultra-high-energy protons produce secondary gamma rays along the line of sight, and that a $\\tau>3$ bin, where the predicted primary fraction is below 5%, would be needed for a real constraint.","pith_inferences":["A natural extension, not pursued in the paper, is to map how the required primary fraction changes with redshift and with the adopted EBL model; if the critical primary fraction needed to reproduce correlations drops below the model prediction for other xHBLs, those sources could constrain the proton scenario even without $\\tau>3$ data.","The toy's reliance on a smooth secondary component is the hinge: if intergalactic magnetic fields are near the upper limit of $\\lesssim10^{-14}$ G or the proton beam is broad, cascade emission could vary on shorter timescales, and the 13% test could either under- or over-estimate the correlation.","The same opacity-binned ZDCF analysis could be applied to the full VERITAS xHBL sample (e.g. 1ES 0229+200, RGB J0710+591) to check whether a small primary fraction always reproduces the observed correlations; a source where it does not would be the most promising place to look for a constraint."],"forward_implications":["The ~4σ opacity-bin correlation in 1ES 1011+496 cannot by itself serve as evidence against the ultra-high-energy proton scenario; a 13% primary fraction during the flare is enough to produce it.","A meaningful test requires a $\\tau>3$ opacity bin, where the secondary model predicts the primary contribution is below 5% and no correlation with the $\\tau<1$ light curve is expected.","For PG 1553+113, the 3.8σ VHE–X-ray correlation is compatible with a one-zone SSC interpretation, while the absence of VHE–HE correlation adds tension to that model.","For 1ES 1218+304, uncorrelated VHE and X-ray short-term variability suggests that different emission zones or particle populations may be at work.","An overall consequence is that variability correlation measurements of these HBLs can point toward single-zone or multi-zone scenarios, but only the highest-opacity bins can cleanly separate primary from secondary emission."],"supporting_citations":[{"why":"Proposes that the gamma-ray flux from distant blazars is dominated by secondary gamma rays produced in proton interactions with background photons, the hypothesis under test.","marker":"[8]"},{"why":"Provides the Bayesian block algorithm used to determine significant change points and to identify the flare period that is later removed from the $\\tau<1$ light curve.","marker":"[15]"},{"why":"Provides the ZDCF method used for all cross-correlation measurements and their significance estimates in this work.","marker":"[16]"},{"why":"Supplies the PSD-based algorithm for simulating light curves, used both for significance levels and for the smooth secondary component in the toy model.","marker":"[17]"},{"why":"Develop the ultra-high-energy proton scenario in which protons from the blazar interact with background photons to generate VHE secondary gamma rays along the line of sight.","marker":"[24–26]"},{"why":"Provides the prediction that secondary-dominated emission should not show variability on timescales shorter than about 0.1 year, the physical basis of the opacity-binned correlation test.","marker":"[27]"},{"why":"Establishes the opacity-binning strategy used to separate energy ranges according to the expected attenuation of primary photons.","marker":"[28]"},{"why":"Supplies the Franceschini et al. EBL model, $\\tau(E,z)$, used to set the energy boundaries of the three opacity bins.","marker":"[29]"}],"fun_headline_variants":["Blazar 4σ correlation fails to exclude proton-driven gamma rays","Extreme blazar data keep secondary gamma-ray model viable","Proton-induced gamma rays survive blazar variability test","Blazar light curves can't pin down gamma-ray origin"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the secondary gamma-ray component is smooth, so it can be represented by a power-spectrum simulation of the flare-removed low-opacity light curve; if secondary emission can track the flare on short timescales, the 13% primary test would not show what the paper says it shows.","fun_headline_variants_meta":{"raw":{"variants":["Blazar 4σ correlation fails to exclude proton-driven gamma rays","Extreme blazar data keep secondary gamma-ray model viable","Proton-induced gamma rays survive blazar variability test","Blazar light curves can't pin down gamma-ray origin"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000836,"raw_usage":{"total_tokens":3644,"prompt_tokens":937,"completion_tokens":2707,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":553,"completion_tokens_details":{"reasoning_tokens":2639}},"tokens_in":553,"tokens_out":2707,"duration_ms":22955,"temperature":1.0,"reasoning_tokens":2639,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:25:02.087736+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Look at the $\\tau>3$ light curve of 1ES 1011+496 around a flare like February 2014. The secondary-gamma-ray model predicts a primary fraction below 5% in that bin, so a significant correlation between the $\\tau>3$ and $\\tau<1$ light curves would refute the paper's conclusion that the current data cannot constrain the model; a smooth, uncorrelated $\\tau>3$ light curve would confirm it.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the Bayesian block algorithm used to determine significant change points and to identify the flare period that is later removed from the $\\tau<1$ light curve."},{"cited_title":"Alexander, vol","cited_arxiv_id":null,"evidence_quote":"Provides the ZDCF method used for all cross-correlation measurements and their significance estimates in this work."},{"cited_title":"Timmer and M","cited_arxiv_id":null,"evidence_quote":"Supplies the PSD-based algorithm for simulating light curves, used both for significance levels and for the smooth secondary component in the toy model."},{"cited_title":"On weak redshift dependence of gamma-ray spectra of distant blazars","cited_arxiv_id":"1111.0815","evidence_quote":"Establishes the opacity-binning strategy used to separate energy ranges according to the expected attenuation of primary photons."}],"review_version":1}