{"id":"ac306e94-14f5-45a1-940c-b62e68530b52","arxiv_id":"2501.16328","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Weak-emission-line quasars show systematically weaker optical variability (factor ~1.76 in amplitude) on month/year timescales than matched normal quasars.","lead":"This paper compares how much the optical brightness of 76 weak-emission-line quasars changes over months to years, using six years of ZTF data and seven years of CRTS data, and compares them with 603 matched normal quasars. It finds that WLQs vary about 1.7 times less in amplitude, a new and independent clue to the nature of these enigmatic objects.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The qualitative result (WLQs are less variable) appears robust, but the headline factor 1.76±0.05 is not yet supported because the quoted errors come from pair-level scatter rather than object-level variance.","rationale":"After reading the full manuscript, the strongest result is the consistent offset between WLQs and normal QSOs in two independent surveys (ZTF and CRTS), supported by KS tests on ψ. The luminosity-line-contamination alternative is the most dangerous systematic, but the authors' argument in Section 4 is quantitatively sound: even taking the full K-correction range (0.6 mag) as an upper limit on line brightening, the known luminosity–variability anti-correlation would require roughly a factor of 20 in continuum luminosity to explain a 1.76× variability deficit, which is implausible. I therefore do not press that concern. The real weakness is statistical: the quoted 1.76±0.05 and the invisible 'smaller than symbol' error bars are computed from pair-level scatter, not from the independent objects, so the uncertainty is almost certainly underestimated. The qualitative claim survives any reasonable widening of the errors, so the paper remains conditionally acceptable; the authors should replace the pair-level error bars with object-level bootstrap confidence intervals and report the uncertainty on the factor explicitly.","tokens_in":10705,"tokens_out":9122,"duration_ms":94902,"concrete_test":"Resample the 76 WLQs and 603 QSOs with replacement at the object level (e.g., 10,000 bootstrap replicates), recompute the ensemble structure function in a fixed rest-frame lag bin (say Δt ≈ 500 days) for each replicate, and form the 95% bootstrap confidence interval on the WLQ/QSO SF ratio. Also compute per-object SF values and report the standard error of the ratio of their means. If the interval is much wider than ±0.05 or includes 1.0, the headline factor should be revised to a range and the claim reworded; if the interval remains narrow and excludes 1.0, the quantitative claim is verified.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claim—the factor 1.76±0.05 in Section 4—rests on the ensemble structure function defined by Eq. (2), yet the error bars on the SF points are stated to be 'smaller than the symbol size' (Figs. 2a, 2b, 3) and are neither displayed nor tabulated. The only error propagation described in Section 3 averages photometric uncertainties over all magnitude-difference pairs within a rest-frame lag bin. That procedure yields the standard error of a mean over many highly correlated pairs drawn from only 76 independent WLQs and 603 independent QSOs; it does not capture object-to-object variance. Given the known broad scatter of AGN optical variability, the 95% confidence interval on the WLQ/QSO SF ratio is likely far wider than ±0.05. The KS tests on the per-object ψ distributions (p ≈ 1e-15) do establish that the two samples differ, so the qualitative conclusion is safe; what is not established is the precise factor and especially its stated uncertainty. The alternative luminosity-offset interpretation discussed in Section 4 is a real systematic, but the authors' K-correction upper limit of 0.6 mag makes the required factor-of-20 luminosity difference implausible, so that is not the main weak point.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper compares the optical variability of 76 radio-quiet weak-emission-line quasars (WLQs) with a matched control sample of 603 radio-quiet normal QSOs using ZTF g- and r-band light curves spanning 2018-2024, and independently verifies the result with CRTS V-band light curves for 51 WLQs and 361 control QSOs. Variability is quantified by the per-object amplitude ψ (Eq. 1) and by the ensemble structure function (Eq. 2) over rest-frame lags of roughly 50-1000 days. The authors find that WLQs are systematically less variable, with the ensemble structure functions differing by a factor of about 1.76±0.05 in amplitude on month/year-like timescales. They argue that this difference is not an artifact of emission-line contribution to the r-band photometry of normal QSOs, using a K-correction range as an upper limit and the luminosity-variability anti-correlation. The paper concludes with a speculative scenario in which the clumpiness of torus material feeding the central engine simultaneously explains the weak emission lines and the reduced variability.","tokens_in":10903,"tokens_out":8997,"duration_ms":85524,"significance":"The qualitative result—that WLQs are milder optical variables than normal radio-quiet QSOs on month/year timescales—is well supported by the highly significant KS tests on the per-object ψ distributions (p ≈ 2.7×10^-15 in r, 4.3×10^-14 in g) and by the consistency between ZTF and CRTS. The use of two independent surveys, a matched control sample, and public data are strengths. The principal weakness is statistical: the quoted uncertainty on the 1.76 factor is derived from pair-level photometric errors rather than object-level variance, and the SF error bars are not displayed or tabulated. Thus the precise factor, and especially its stated error, is not yet established. With a proper object-level resampling the paper could make a solid contribution to the debate on WLQ nature.","major_comments":[{"comment":"The uncertainty budget for the ensemble structure functions is not adequate to support the headline factor of 1.76±0.05. The text states that error bars are smaller than the symbol size, but the only described error propagation averages photometric uncertainties over all magnitude-difference pairs within a rest-frame lag bin. Because pairs drawn from the same light curve are correlated, the effective independent sample size is the number of objects (76 WLQs and 603 QSOs), not the number of pairs. This procedure estimates the standard error of a mean of many correlated pairs and does not capture the object-to-object variance in variability amplitude, which is known to be large among AGN. Please compute object-level bootstrap or jackknife uncertainties on the SF points (or per-object SF distributions) and quote the resulting confidence interval on the WLQ/QSO SF ratio. Without such an analysis, the ±0.05 attached to the 1.76 factor is not justified.","section":"Section 3, Eq. (2) and Section 4, Fig. 2"},{"comment":"The paper quotes a single factor 1.76, but the SF curves in Figs. 2a, 2b and 3 do not show a lag-dependent ratio; the ratio may vary across the probed Δt range. Please present the ratio as a function of rest-frame lag with object-level uncertainties, or explicitly define the averaging used to obtain 1.76 (e.g., the range of lags and whether it is a mean or median of bin-wise ratios). The KS tests on ψ establish that the two samples differ, but they do not by themselves calibrate the magnitude of the difference on month/year timescales.","section":"Section 4"}],"minor_comments":[{"comment":"Section 2 states that ZTF-DR22 light curves are used, but the Data Availability section mentions ZTF DR16; please correct the inconsistency.","section":"Section 2 / Data Availability"},{"comment":"The abstract quotes 1.76±0.05 while Section 5 says '~1.7'; please reconcile these values.","section":"Abstract / Section 5"},{"comment":"For light curves where (Amax−Amin)^2 < 2σ², ψ is undefined; state explicitly how such cases were treated (e.g., set to zero or excluded).","section":"Section 3, Eq. (1)"},{"comment":"The caption of Fig. 2a is confusing: it says the SF is derived from r-band light-curves but describes insets comparing g and r bands; also 'CTRS' in Section 4 and Fig. 3 should be 'CRTS'.","section":"Section 4 and Fig. 3"},{"comment":"The power-law slopes are quoted with uncertainties, but the fitting procedure (e.g., least squares over which lag range) is not described; please specify.","section":"Section 4"},{"comment":"The phrase 'factor of ~1.76±0.05 in amplitude' is ambiguous: the structure function is in magnitudes, so the ratio is a ratio of magnitude amplitudes, not flux amplitudes; please clarify.","section":"Section 4"},{"comment":"The x-axis labels of Fig. A2 appear garbled; please fix the formatting.","section":"Appendix A"}],"recommendation":"major_revision","confidential_remarks":"The paper is a solid observational study that needs a statistical revision. The qualitative result is robust and the required analysis (object-level resampling of the structure-function uncertainties) is straightforward. I recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The thing to know: this paper gives the sharpest evidence yet that weak-emission-line quasars are systematically less optically variable than matched normal QSOs on month-to-year timescales. The two-survey design (ZTF and CRTS) and the matched control sample are the strengths. The KS tests on per-object variability amplitudes are decisive (p ~ 1e-15), so the qualitative result is solid and independently confirmed. The quantitative factor of 1.76±0.05 is less well supported than the prose suggests, because the quoted uncertainty comes from pair-level scatter, not object-level variance. I agree with the stress-test note: with only 76 WLQs and the known broad object-to-object scatter in AGN variability, the 95% interval on the ratio is probably much wider than ±0.05. That is a real but contained flaw—it does not threaten the main conclusion, only the precision claim. The luminosity-offset alternative is handled honestly; their K-correction upper limit of 0.6 mag means an order-of-magnitude luminosity difference would be needed to explain the effect, which is implausible. The torus-clumpiness speculation at the end is flagged as simplified and is not load-bearing. Minor issues: the SF error bars not being tabulated is annoying, and the paper leans on an earlier hint from Kumar et al. 2018 without overclaiming. The citation pattern is relevant and not self-serving. Who gets value: anyone working on AGN variability, WLQ phenomenology, or quasar structure functions. It deserves a serious referee, mostly to pin down the error propagation and make the SF uncertainties explicit. I would cite it for the qualitative WLQ/QSO variability difference, not for the precise factor.","headline":"A credible, well-matched observational claim that WLQs are milder optical variables than normal QSOs, though the headline factor's error bar is overstated.","tokens_in":11485,"tokens_out":465,"would_cite":true,"duration_ms":5927,"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":"Weak-emission-line quasars are systematically less optically variable than normal quasars, by a factor of about 1.76 in amplitude.","keywords":["weak emission line quasars","optical variability","structure function","Zwicky Transient Facility","Catalina Real-Time Transient Survey","radio-quiet quasars","accretion disks","torus clumpiness"],"falsifier":"Match WLQs and normal QSOs by continuum luminosity measured at line-free rest-frame wavelengths (or by mid-infrared luminosity) instead of by r-band magnitude, and recompute the ensemble structure functions; if the ~1.76 amplitude ratio disappears, the central claim was a selection artifact.","tokens_in":10493,"feed_emoji":"🔭","tokens_out":10451,"duration_ms":86071,"temperature":0.7,"pith_summary":"Using six years of Zwicky Transient Facility light curves for 76 radio-quiet weak-emission-line quasars (WLQs) and 603 normal radio-quiet QSOs matched in redshift and r-band magnitude, this paper shows that WLQs vary more mildly than normal QSOs on month/year-like time scales, by a factor of about $1.76\\pm0.05$ in amplitude. The same factor is recovered independently from V-band light curves of 51 WLQs and 361 controls from the Catalina Real-Time Transient Survey. If the difference is real, it adds a new observational handle on an enigmatic quasar subclass and suggests that the weak broad emission lines and the calm optical continuum share a common cause, such as a smoother, less clumpy inflow of torus material feeding the central engine. The authors argue that the effect is far too large to be explained by emission-line contamination of the control sample magnitudes.","feed_headline":"Quasars with weak emission lines vary 1.76 times less in brightness","feed_subtitle":"Six years of ZTF light curves show a clean new split between two quasar classes, hinting at clumpy inflow as the cause.","key_machinery":"The load-bearing tool is the ensemble structure function, defined as $\\mathrm{SF}(\\Delta t)=\\sqrt{\\frac{\\pi}{2}\\langle |m(t+\\Delta t)-m(t)|\\rangle^2 - \\langle\\sigma^2\\rangle}$, computed in bins of rest-frame time lag across all light curves of a sample; because the bins are independent, it gives a per-timescale measure of average variability. The companion statistic is the per-object variability amplitude $\\psi=\\sqrt{(A_{\\max}-A_{\\min})^2-2\\sigma^2}$. These are applied to two matched samples built from 76 WLQs and 603 normal QSOs (10 per WLQ, matched in redshift and r-band magnitude), and the structure functions are compared between surveys, with the line-contamination concern handled by a bounded K-correction argument.","core_discovery":"The paper's central claim is that, as a class, WLQs have intrinsically milder optical continuum variability than normal radio-quiet QSOs. In the ZTF data, median variability amplitudes are $\\psi = 0.35\\pm0.03$ (r band) and $0.40\\pm0.04$ (g band) for WLQs, versus $0.62\\pm0.01$ and $0.72\\pm0.01$ for the matched controls, and the ensemble structure functions differ by a factor of $\\sim1.76\\pm0.05$ in amplitude over rest-frame lags of about 100 to 1000 days. The CRTS V-band data confirm the pattern, ruling out a filter-specific or survey-specific artifact. After bounding the possible emission-line contribution through K-corrections and comparing with the known luminosity-variability anti-correlation, the authors conclude that the milder variability is intrinsic to the quasar continuum, not a luminosity-matching artifact.","pith_inferences":["A direct extension would be to check whether the variability amplitude correlates continuously with Ly-alpha+NV equivalent width across the WLQ-to-normal-QSO sequence, which would tie the two defining properties quantitatively.","The variability floor could be used photometrically to pre-select WLQ candidates in time-domain surveys before spectroscopy.","If smooth inflow is the cause, WLQs should show comparatively subdued variability in UV/X-ray bands as well, since those bands probe the innermost disk; a multi-wavelength campaign could test this.","A complementary test is to search for rare high-variability WLQs; they would be natural transitional objects in which the inflow becomes clumpy while the lines are still weak."],"forward_implications":["WLQs form a distinct variability class: on month/year-like time scales they are about 1.76 times less variable in optical amplitude than normal radio-quiet QSOs.","The variability difference is a new observational discriminator that is at least as clean as the equivalent-width criterion that defines WLQs.","The difference persists across two independent surveys and two filter systems, so it is not an artifact of one photometric band or cadence.","The size of the effect rules out emission-line contamination of r-band magnitudes as its cause, pointing to an intrinsic continuum property.","If the proposed torus-clumpiness scenario is correct, weak broad emission lines and mild optical variability are twin symptoms of the same inflow condition."],"supporting_citations":[{"why":"ZTF survey reference; supplies the g- and r-band light curves that carry the main variability measurement.","marker":"Bellm et al. 2018"},{"why":"CRTS survey reference; supplies the independent V-band light curves used to confirm the result.","marker":"Drake et al. 2009"},{"why":"SDSS DR14 quasar catalog; provides the parent list from which the redshift-magnitude matched control sample is drawn.","marker":"Pâris et al. 2018"},{"why":"Compilation of 90 bona fide WLQs from which the 76-source sample is drawn.","marker":"Kumar et al. (2023)"},{"why":"Parent WLQ compilation; defines the WLQ subclass and supplies many of the sample candidates.","marker":"Plotkin et al. (2010)"},{"why":"Second parent WLQ compilation; supplies additional sample candidates and sample definition context.","marker":"Meusinger & Balafkan (2014)"},{"why":"K-correction table used to bound the maximum emission-line contribution to r-band photometry.","marker":"Richards et al. (2006)"},{"why":"Luminosity-variability anti-correlation fit used to show the 1.76 factor cannot be a luminosity-matching artifact.","marker":"Laurenti et al. (2020)"},{"why":"Gives the variability amplitude definition psi used for the per-band median statistics.","marker":"Heidt & Wagner (1996)"},{"why":"Gives the modified structure function definition on which the ensemble analysis is built.","marker":"Di Clemente et al. (1995)"}],"fun_headline_variants":["WLQ quasars vary 1.76x less, hinting at clumpy inflow","Weak-lined quasars are calmer: 1.76x less variability","New clue: weak-emission quasars show milder brightness swings","Quasar puzzle: weak emission lines come with quieter light curves","Why are weak-emission quasars so calm? Torus clumpiness may explain"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The claim stands on the match between the 603 normal QSOs and the 76 WLQs being representative in continuum luminosity, and on the external calibrations that say emission-line contamination can brighten a normal QSO by at most about 0.6 magnitudes while a factor-of-1.76 variability gap would need a factor-of-20 luminosity offset.","fun_headline_variants_meta":{"raw":{"variants":["WLQ quasars vary 1.76x less, hinting at clumpy inflow","Weak-lined quasars are calmer: 1.76x less variability","New clue: weak-emission quasars show milder brightness swings","Quasar puzzle: weak emission lines come with quieter light curves","Why are weak-emission quasars so calm? Torus clumpiness may explain"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000188,"raw_usage":{"total_tokens":1371,"prompt_tokens":1021,"completion_tokens":350,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":637,"completion_tokens_details":{"reasoning_tokens":250}},"tokens_in":637,"tokens_out":350,"duration_ms":4089,"temperature":1.0,"reasoning_tokens":250,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T13:29:53.306553+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Match WLQs and normal QSOs by continuum luminosity measured at line-free rest-frame wavelengths (or by mid-infrared luminosity) instead of by r-band magnitude, and recompute the ensemble structure functions; if the ~1.76 amplitude ratio disappears, the central claim was a selection artifact.","supporting_citations":[{"cited_title":"C., et al., 2018, PASP , 131, 018002","cited_arxiv_id":null,"evidence_quote":"ZTF survey reference; supplies the g- and r-band light curves that carry the main variability measurement."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Compilation of 90 bona fide WLQs from which the 76-source sample is drawn."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Second parent WLQ compilation; supplies additional sample candidates and sample definition context."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Luminosity-variability anti-correlation fit used to show the 1.76 factor cannot be a luminosity-matching artifact."},{"cited_title":"J., 1996, , http://adsabs.harvard.edu/abs/1996A","cited_arxiv_id":null,"evidence_quote":"Gives the variability amplitude definition psi used for the per-band median statistics."},{"cited_title":"The Variability of Quasars. II. Frequency Dependence","cited_arxiv_id":"astro-ph/9512159","evidence_quote":"Gives the modified structure function definition on which the ensemble analysis is built."}],"review_version":1}