{"id":"440b832d-eb1a-4ec0-81dc-050ed6961473","arxiv_id":"2607.19485","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"In five changing-state AGNs, the X-ray bolometric correction follows one mass-independent relation with Eddington ratio, with intrinsic scatter of only 0.05 dex.","lead":"Using more than a thousand simultaneous X-ray and UV/optical observations of five rapidly changing black-hole systems, this study finds that the fraction of energy emitted as X-rays tracks the Eddington ratio — how hard the black hole is eating — on a single, very tight curve. If it holds, it gives a practical way to estimate total AGN luminosity from X-ray data alone.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq. 3's 'unambiguous' driver conclusion is undercut by shared-Lbol algebra and 5-source clustering; the partial correlation is essentially a between-source MBH effect.","rationale":"The reader's weakest-assumption diagnosis is correct and I would not move the verdict away from CONDITIONAL. The covariance between kappa and lambda through L_bol is not a minor nuisance; it is the central inferential threat to the paper's 'unambiguously' wording. The algebra makes this precise: with M_BH fixed per source, partialling out L_bol leaves a source-level residual for lambda, so the reported partial correlation cannot support an epoch-by-epoch causal claim. The paper does provide genuine supporting evidence — individual-source tracks are broadly consistent, removing the X-ray component from L_bol leaves a still-tight 0.11-dex scatter, and the trend agrees at high lambda with earlier population studies. These make the qualitative claim plausible, but they do not justify the 'unambiguous' language or the specific 0.05-dex scatter as a physical law. The suggested r-lambda refit is a decisive and inexpensive check: it removes the shared numerator while preserving the same physical content. I also flag the missing Paper II reference as an editorial issue. Overall, conditional acceptance remains appropriate, with the added requirement to report cluster-robust statistics and present the covariance-corrected scatter as the headline.","tokens_in":18616,"tokens_out":8741,"duration_ms":102500,"concrete_test":"Refit the relation using the algebraically non-shared pair x = log(L_bol/L_Edd) and y = log(L_X/L_Edd), which is exactly equivalent to Eq. 3 (log kappa = log x - log y, log lambda = log x). Fit y as a function of x with a linear mixed model that includes random intercepts for the 5 sources, or use cluster bootstrap by source to estimate the slope significance and residual scatter with only 5 independent clusters. If the residual scatter exceeds ~0.2 dex or the slope is not significant after clustering, the 0.05-dex kappa-lambda tightness is largely a shared-L_bol artifact and the 'unambiguous lambda_Edd driver' claim should be softened.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The headline 0.05-dex tightness of Eq. 3 is not, by itself, evidence that lambda_Edd is the primary driver, because both axes share L_bol in the numerator: log kappa = log L_bol - log L_X, log lambda = log L_bol - log M_BH. With M_BH constant per source, any within-source relation between kappa and lambda can be dominated by the common L_bol term. The authors' own covariance test in Section 3.2 — scatter increasing from 0.05 to 0.11 dex when the X-ray continuum is removed from L_bol — shows that the quoted scatter is partly mathematical, and even the retest leaves a shared L_bol' in both variables. The partial-correlation evidence is also fragile: after partialling out L_bol, the residual of log lambda is essentially a five-level source effect (log M_BH is constant per object), so rho_{kappa,lambda | L_bol} = 0.93 mostly measures between-source correlations of L_X with M_BH, not an intra-source, epoch-by-epoch coupling. Treating 214 stacked epochs as independent inflates the significance; the effective sample size may be closer to five. Thus the central claim of a single mass-independent, 'unambiguous' driver is not secure. Secondary: 'Paper II' (Jana et al. 2026b) is cited in Section 2 but absent from the reference list.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper analyzes multi-epoch Swift UVOT/XRT observations of five changing-state AGNs (NGC 1566, NGC 2617, Mrk 590, Mrk 1018, IRAS 23226−3843), stacking to 214 epochs and fitting three-component SEDs to estimate Lbol and the 2–10 keV continuum luminosity. From these it defines the X-ray bolometric correction κ2−10 and Eddington ratio λEdd, and reports a tight quadratic relation (Eq. 3) with intrinsic scatter ~0.05 dex, together with partial-correlation results that are interpreted as showing λEdd is the primary driver of κ2−10. The paper also presents a κ2−10–αOX relation and discusses implications for accretion-flow structure, X-ray weakness in high-z sources, and use of the relation to estimate Lbol from X-ray luminosity and black hole mass.","tokens_in":18988,"tokens_out":8251,"duration_ms":89287,"significance":"If Eq. (3) is robust, it would provide a remarkably simple, mass-independent description of disk-corona coupling and a practical route to estimating Lbol from LX and MBH. The data set is genuinely valuable: five changing-state AGNs with ~1000 epochs of simultaneous UV/optical/X-ray coverage, careful host-galaxy subtraction, and a physically motivated SED decomposition. The authors also deserve credit for explicitly attempting a covariance check in §3.2 and for comparing with previous samples. However, because the headline scatter and the partial correlations are affected by the definitional sharing of Lbol and by the five-source clustering, the current analysis does not yet establish the 'unambiguous' driver claim stated in the abstract. The needed analyses—within-source correlations, cluster-aware statistics, and a per-source fit table—are feasible with the existing data, so the question is well posed even though the present evidence is not yet conclusive.","major_comments":[{"comment":"The tightness of Eq. (3) is partly definitional: log κ2−10 = log Lbol − log LPC2−10 and log λEdd = log Lbol − log LEdd share Lbol, and LEdd is constant per source. The authors' X-ray-free test (scatter increasing from 0.05 to 0.11 dex) is a good-faith check, but Lbol′ = Ldisk + LSE still enters both axes, so the shared term is not removed. Likewise, after partialling out Lbol, the residual of log λEdd is essentially the source-level quantity −log MBH (plus noise), so the reported partial correlation of 0.93 is dominated by between-source differences rather than epoch-by-epoch coupling. Please report within-source (demeaned or mixed-effects) correlations and the within-source scatter, and quantify how much of the 0.05 dex depends on the exact construction of Lbol.","section":"§3.2, Eq. (3)"},{"comment":"The 214 stacked epochs are treated as independent in the Spearman tests, p-values, and scatter estimates, but they are strongly clustered within five objects and are temporally autocorrelated; with only five MBH values, the effective sample size for the between-source comparison is close to five. The p≪10^−10 values and the 0.05 dex intrinsic scatter therefore overstate the significance of the relation. Please use cluster-robust bootstrap or a mixed-effects model with source as a random effect, report the effective number of independent epochs, and test whether Eq. (3) survives with a random source intercept. Without this, the claim that λEdd is 'the' driver rather than a source-level proxy is not established.","section":"§2, §3.2 (statistical methods)"},{"comment":"The central assertion that all five sources follow the same κ2−10–λEdd relation is not quantitatively documented: the per-source fits are described only as 'consistent within uncertainties,' with no coefficients, intrinsic scatters, residuals, or λEdd coverage given for individual objects. Since the sample contains only five sources with MBH spanning 6.8–7.8 dex, a single mass-independent law cannot be inferred from this sample alone. Please provide a per-source fit table (coefficients, scatter, residual RMS, λEdd range), show within-source residuals against Eq. (3), and state explicitly which parts of the relation are anchored by which sources. In addition, the recommendation to invert Eq. (3) for Lbol is an in-sample calibration; an out-of-sample or cross-validation check would be needed before this can be called 'reliable.'","section":"§3.2, Fig. 2"}],"minor_comments":[{"comment":"Paper II (Jana et al. 2026b) is cited in §2 and elsewhere but does not appear in the reference list. The same is true for 'Jana et al. 2025b,c' cited near the end of §5. Please add the missing references or revise the citations.","section":"§2, §5, References"},{"comment":"The caption begins with 'Figure 5.raction' (typo) and the energy ranges listed for the three panels are permuted relative to the definitions in §2: Ldisk is 10^−7–0.5 keV, LSE is 0.001–10 keV, and LPC is 0.1–500 keV. Please correct the caption.","section":"Appendix A, Fig. 5 caption"},{"comment":"SciPy is credited as 'F. M. Vincentelli et al. 2020'; the standard citation for SciPy is Virtanen et al. (2020). Please correct.","section":"Software paragraph"},{"comment":"The word 'unambiguously' in the abstract is stronger than the evidence supports, given the covariance and clustering issues. Consider softening to 'strongly suggests' or adding the quantitative caveats from §3.2.","section":"Abstract and §4"}],"recommendation":"major_revision","confidential_remarks":"To the editor: the reader's conditional verdict and the skeptic's main concern are, in my reading, both on target. The manuscript's own covariance test shows that the headline scatter is not purely physical, and the partial correlation is heavily influenced by the five source-level MBH values. I would not reject: the data set is uniquely well suited to the question and the authors have already begun the right robustness checks. The needed fixes—within-source or mixed-effects analysis, cluster-robust significance, a per-source fit table, and the missing references—are feasible within a revision. Therefore I recommend major revision rather than rejection or acceptance at this stage."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a genuinely useful empirical study of five changing-state AGNs. It gives the AGN community a new tight κ2-10–λEdd calibration (Eq. 3) that extends down to log λEdd ≈ -3.6, well beyond the coverage of earlier samples. The multi-epoch, simultaneous UV/optical/X-ray SED modeling of 214 stacked epochs is careful, and the demonstration that individual sources follow mass-offset tracks in Lbol yet superimpose onto the same λEdd relation is a real and interesting result. The paper is also transparent about the covariance issue—it explicitly tests the relation with the X-ray continuum removed from Lbol, which is more honesty than many papers in this area show.\n\nWhere I push back is the headline interpretation. Both κ2-10 and λEdd are constructed from the same Lbol, so part of the tiny scatter is definitional. The authors show this themselves: when Lbol excludes the X-ray continuum, the scatter grows from 0.05 to 0.11 dex. Even that reduced scatter still contains shared Lbol, so the absolute tightness is not a clean test of the physical driver. More importantly, the partial-correlation analysis that supports λEdd over Lbol is fragile: after partialling out Lbol, log λEdd is essentially -log MBH plus a constant for each source, so the reported ρ = 0.93 mostly measures between-source correlations of L_X with MBH, not an epoch-by-epoch coupling. With only five objects and MBH constant per source, the effective sample size for the 'driver' claim is closer to five than 214, yet the paper treats the stacked epochs as independent in the fits and significance tests. The 'unambiguously' in the abstract overstates what the data can show; 'strongly suggests' would be defensible.\n\nSecondary but real: Paper II (Jana et al. 2026b) is cited in Section 2 and elsewhere but is missing from the reference list. The narrow MBH range (6.8–7.8) also limits how strongly one can claim mass independence, though the offset tracks in Lbol do help.\n\nBottom line: this deserves serious peer review and, with revisions, likely publication. The data and the multi-epoch approach are valuable, and Eq. 3 will be cited as a convenient bolometric correction regardless of the interpretation. The authors should soften the abstract, quantify the effective number of independent epochs, present the between-source versus within-source evidence more carefully, and fix the missing reference. The physics is plausible and the paper is honest about its main limitation—it just overclaims in the final step.","headline":"A useful multi-epoch κ2-10–λEdd calibration from five changing-state AGNs, but the 'unambiguous' driver claim overreaches because both variables share Lbol and the partial-correlation evidence is mostly a between-source MBH effect.","tokens_in":19589,"tokens_out":4356,"would_cite":true,"duration_ms":43316,"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":"Tracking five wildly variable active galactic nuclei across 1,000 epochs, this paper shows that the X-ray bolometric correction—the ratio of total to X-ray luminosity—is set almost entirely by the Eddington ratio.","keywords":["active galactic nuclei","X-ray bolometric correction","Eddington ratio","changing-state AGN","accretion disk-corona coupling","spectral energy distribution","AGN variability"],"falsifier":"Compute the intrinsic scatter of the same κ2-10–λEdd relation using bolometric luminosities derived from the optical/UV disk component alone, excluding the X-ray continuum; if the scatter rises from 0.05 dex to the ~0.3 dex typical of heterogeneous samples, the 'Eddington ratio alone' claim collapses into shared-variable covariance.","tokens_in":18503,"feed_emoji":"🕳️","tokens_out":6655,"duration_ms":69015,"temperature":0.7,"pith_summary":"The paper tries to establish that the X-ray bolometric correction—how much of an AGN's total luminosity emerges as 2–10 keV X-rays—is governed by a single parameter: the Eddington ratio, the accretion rate relative to the Eddington limit. Using more than a thousand simultaneous optical/UV/X-ray observations of five changing-state AGNs, stacked into 214 epochs spanning three orders of magnitude in Eddington ratio, the authors find all five sources land on one curve, log κ2-10 = 1.957 + 0.654 log λEdd + 0.074 (log λEdd)^2, with intrinsic scatter of only 0.05 dex. A sympathetic reader cares because this would mean bolometric luminosity can be recovered from X-ray luminosity plus black hole mass without full spectral coverage, and because it identifies the dimensionless accretion rate—not luminosity or mass—as the fundamental control knob for how much power the corona emits.","feed_headline":"One ratio predicts AGN X-ray output to 0.05 dex","feed_subtitle":"Five violently variable galactic nuclei trace one shared curve, so X-rays alone can now reveal a black hole's true energy output.","key_machinery":"The engine of the argument is a sample of five changing-state AGNs—nuclei that swing between optical spectral types as their accretion rates surge—monitored simultaneously in X-ray, UV, and optical bands over more than 1,000 epochs, binned into 214 SEDs. Each SED is decomposed into a multi-temperature disk blackbody, a soft-excess blackbody, and a cutoff power law for the hot corona, giving L_bol and the 2–10 keV intrinsic continuum luminosity. These feed the two quantities at the heart of the paper: κ2-10 = L_bol/L2-10 and λEdd = L_bol/L_Edd. The relation (Eqn 3) is what carries the claim: it collapses the five sources onto one curve, erasing the mass-dependent offsets seen in luminosity sp","core_discovery":"The central discovery is a mass-independent empirical law connecting the X-ray bolometric correction κ2-10 to the Eddington ratio λEdd across 10^-3.6 to 10^-0.5. The authors construct bolometric luminosities by fitting a thermal disk, soft-excess, and cutoff power-law model to simultaneous UV/optical and X-ray spectra, then show that while individual sources trace different κ2-10 versus luminosity tracks whose zero-points shift with black hole mass, they all converge on the same κ2-10–λEdd curve (Eqn 3). Partial-correlation analysis—controlling for bolometric luminosity—leaves κ2-10 and λEdd strongly correlated (coefficient 0.93), while the luminosity–κ2-10 correlation drops to –0.56 once λE","pith_inferences":["An implication the paper leaves implicit: Eqn 3 can be inverted to solve for black hole mass given independent X-ray and bolometric luminosity, turning the relation into a single-epoch mass estimator for AGNs where reverberation mapping is impractical.","The same monitoring strategy applied to a sample spanning a wider range of black hole mass (e.g., 10^5 to 10^10 solar masses) would test whether the relation is truly mass-independent or carries a hidden mass term that five similar-mass objects cannot reveal.","If the curve extends to super-Eddington rates, it predicts that the most luminous high-redshift quasars—which are often X-ray weak—owe part of that weakness to their accretion state rather than obscuration, a distinction upcoming X-ray and infrared surveys could separate."],"forward_implications":["If the relation holds, bolometric luminosity can be estimated from a single X-ray measurement and a black hole mass, without ultraviolet/optical coverage, by inverting Eqn 3.","The small scatter (0.05 dex) implies that Eddington ratio, not luminosity or black hole mass, is the primary regulator of the fraction of energy emitted by the corona—a direct constraint on disk-corona models.","At high Eddington ratio the disk supplies more than 90% of bolometric output and the 2–10 keV corona only about 3%; at low Eddington ratio the corona rises to ~30%, confirming the 'X-ray loud' state of low-accretion AGNs.","The UV-to-X-ray spectral index αOX correlates with κ2-10 at 0.18-dex scatter, giving a cheaper single-color proxy for bolometric correction.","A break in the relation at log λEdd ≈ –2.36 aligns with changes in photon index and soft-excess behavior, marking a restructure of the accretion flow near λEdd ~ 0.01."],"fun_headline_variants":["Eddington ratio, not mass, sets AGN X-ray bolometric correction","Mass-independent law: AGN X-ray output tied to Eddington ratio","Eddington ratio alone predicts AGN X-ray output to 0.05 dex","Black hole mass irrelevant? Eddington ratio controls AGN X-rays"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The load-bearing premise is that the 0.05-dex tightness is real physics rather than an arithmetic echo of defining both κ2-10 and λEdd from the same L_bol, and that the 214 stacked epochs act as independent measurements rather than five sources repeating one trend.","fun_headline_variants_meta":{"raw":{"variants":["Eddington ratio, not mass, sets AGN X-ray bolometric correction","Mass-independent law: AGN X-ray output tied to Eddington ratio","Eddington ratio alone predicts AGN X-ray output to 0.05 dex","Black hole mass irrelevant? Eddington ratio controls AGN X-rays"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000939,"raw_usage":{"total_tokens":3920,"prompt_tokens":879,"completion_tokens":3041,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":623,"completion_tokens_details":{"reasoning_tokens":2956}},"tokens_in":623,"tokens_out":3041,"duration_ms":18519,"temperature":1.0,"reasoning_tokens":2956,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T12:36:16.304553+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the intrinsic scatter of the same κ2-10–λEdd relation using bolometric luminosities derived from the optical/UV disk component alone, excluding the X-ray continuum; if the scatter rises from 0.05 dex to the ~0.3 dex typical of heterogeneous samples, the 'Eddington ratio alone' claim collapses into shared-variable covariance.","supporting_citations":[],"review_version":1}