{"id":"ca74f169-4388-494c-b092-3f03c3554a00","arxiv_id":"2412.12420","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"A bluer-when-brighter color pattern in SDSS type 2 AGNs produced 73 candidates, and follow-up spectroscopy of nine confirmed five changing-look AGNs.","lead":"Astronomers tested a simple color-magnitude trick for finding changing-look active galaxies: these objects tend to look bluer when they get brighter. They selected 73 candidates from the SDSS catalog, observed nine with new spectra, and confirmed five as changing-look AGNs.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Host-galaxy dilution can explain the apparent type 2→1 transitions: three of four confirmed cases already had broad Hα in SDSS and the paper itself concedes they were type 1.9.","rationale":"The reader's weakest assumption—that BWB and broad-line appearance trace the same intrinsic accretion-state change rather than variable obscuration or host dilution—is precisely the load-bearing point. My analysis adds specificity by pointing to the paper's own admission in Section 5 that three of the four CLAGNs were type 1.9 with broad Hα already present, and that host-galaxy emission was strong or dominant. This is not an external critique; it is an internal concession that directly contradicts the abstract's categorical 'type 2 to type 1' language. The concrete test I propose would settle the matter by checking whether a constant broad-line flux can reproduce the observed spectral evolution. If it can, the method's efficiency claim collapses from 'found 5 CLAGNs' to 'found 1-2 CLAGNs plus 3-4 variable AGNs with persistent broad lines.' However, the paper still provides a useful candidate list and a photometric preselection technique, and the authors already acknowledge the alternative scenario; thus the appropriate disposition is to keep the reader's CONDITIONAL verdict rather than reject the paper outright. The reader's emphasis on sample-size selection bias and the reliance on Wang et al. (2024) are secondary to this physical interpretation issue, which is why I mark full agreement with their weakest-assumption identification.","tokens_in":18780,"tokens_out":3026,"duration_ms":30226,"concrete_test":"Quantitatively test the dilution hypothesis for J0751+4948, J1020+2437, and J1344+5126. Re-fit the SDSS and LJT spectra jointly with (i) a host-galaxy template, (ii) a power-law AGN continuum, (iii) narrow lines, and (iv) a broad-line component whose flux is constrained to be identical between epochs (or scaled by a single factor). Compare this constant-BLR model against a model that allows the broad-line flux to vary freely, using Δχ² or an F-test. If the constant-BLR model fits both epochs acceptably, the apparent 'turn-on' is fully explained by continuum brightening against a fixed BLR, and these three objects are not genuine type 2→1 transitions. A similar test should be applied to J1150+3503 using the Wang et al. (2024) spectra and the poor-quality DOT spectrum before counting it as a success.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim (Section 5) that the method 'effectively find CLAGNs' with five of nine type 2→1 transitions is weakened because the confirmations do not establish a genuine type transition. Section 5 explicitly concedes: 'We note that three CLAGNs in our cases (except J1203+6053) had a broad Hα component in their SDSS spectra, and they were mostly type 1.9, not pure type 2.' It also states that host-galaxy emission was strong or dominant in the SDSS spectra of the four CLAGNs. Under these conditions, the observed spectral change can be produced by a brightening AGN continuum (the BWB pattern) that outshines a persistent, weak BLR against the host galaxy—an apparent CL phenomenon, not a physical turn-on of broad lines. The paper's own discussion of J. Li et al. (in preparation) describes exactly this scenario: the AGN is intrinsically type 1 throughout, and host-galaxy dilution makes it appear type 1.8/1.9 or type 2 at low continuum states. If this interpretation holds, only J1203+6053—whose DOT spectrum is noisy and partial—might be a genuine type 2→1 case, and even that is uncertain. Consequently, the abstract's 'five among them showed the CL transition from type 2 to type 1' overstates what the data demonstrate; the method may instead be selecting strongly variable AGNs with persistent broad lines. This does not invalidate the useful 73-candidate list or the photometric selection itself, but it undermines the claimed success rate as a measure of efficiency for finding true changing-look AGNs.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a photometric pre-selection method for finding changing-look AGNs (CLAGNs). For SDSS type 2 AGNs, the authors fit the ZTF zg-zr colour versus zr magnitude relation and select sources with bluer-when-brighter slopes k >= 0.1, yielding 73 candidates after visual inspection. They obtained LJT/DOT spectra for nine of these candidates and report five CLAGNs (three newly confirmed here, two previously reported in Wang et al. 2024), and they observed three type 1 AGNs with redder-when-brighter slopes but found no type 1 to type 2 transitions. The paper also compares optical and MIR variability of successes and failures and proposes refined selection criteria for future follow-up.","tokens_in":19104,"tokens_out":8386,"duration_ms":71042,"significance":"If the interpretation is accepted, the method is valuable because it uses wide-field time-domain photometry to catch recent CL events that spectroscopic surveys may miss, and the 73-object candidate list is a useful resource. The paper's transparency about the limitations (the type 1.9 nature of three confirmed cases, the strong host-galaxy contribution, and the alternative dilution/EUV scenario) is commendable. However, these same limitations mean that the headline claim of five type 2 to type 1 transitions is not established, and the reported success rate cannot yet be read as a validated efficiency. The core photometric selection remains promising and the observational material is useful, but the claim needs to be reframed and quantified.","major_comments":[{"comment":"The abstract's claim that five of nine candidates 'showed the CL transition from type 2 to type 1' is stronger than the data support. Table 3 lists broad H-alpha detections in the SDSS spectra of J0751+4948, J1020+2437, and J1344+5126, and Section 5 concedes that these sources were mostly type 1.9, not pure type 2. The observed spectral evolution is therefore mainly a strengthening of an already-present broad H-alpha and the appearance of broad H-beta; this is a type 1.9 to type 1 change or an apparent CL phenomenon, not a type 2 to type 1 transition. Because the stated success rate is the central quantitative result, the abstract and Section 5 should be rewritten to state the actual measured changes and to separate pure type 2 to type 1 events from type 1.9 to type 1 events.","section":"Abstract; Section 5; Table 3"},{"comment":"The count of five successes also includes J1150+3503, whose DOT spectrum is described in Section 4 as being of bad quality and whose CL identification is taken from Wang et al. (2024). The tally should be presented as three new confirmations plus two previously reported sources, one of which is not re-confirmed in this paper. The current phrasing ('two have already been identified as CLAGNs in Wang et al. 2024, and three are newly discovered by us') obscures this important distinction.","section":"Section 5; Table 1"},{"comment":"The paper itself advances an alternative interpretation that undermines the method's specificity: it states that host-galaxy emission was strong or dominant in the SDSS spectra, and cites J. Li et al. (in preparation) for a scenario in which short-term EUV variations make an intrinsically type 1 AGN appear type 1.8/1.9 or type 2 at low continuum states, so that 'the CLAGNs have always been type 1.' Under this scenario, both the BWB colour pattern and the apparent broad-line turn-on are driven by continuum brightening diluting a persistent BLR, not by a genuine BLR state change. The paper does not provide host-subtracted or continuum-normalized broad-line luminosity measurements to distinguish these cases; the EWs in Table 3 are not sufficient because they depend on the continuum. The conclusion that the method 'can effectively find CLAGNs' should therefore be softened to 'CL-like spectral variability,' or the analysis should quantify how many of the five objects show intrinsic broad-line turn-on rather than dilution.","section":"Section 5"},{"comment":"J1203+6053 is one of the three newly confirmed CLAGNs, but its identification rests on a noisy DOT spectrum that captures only half of H-alpha, and no difference spectrum was obtained. The text says the presence of broad H-alpha and H-beta 'suggests' this is a CLAGN. Given that the overall success count depends on this object, the paper should either present a better-quality spectrum or explicitly classify J1203+6053 as a tentative CLAGN, and the success rate should be reported with confidence levels per object.","section":"Section 4.1.3; Figure 5"},{"comment":"The selection thresholds (k >= 0.1, N_zr > 43, visual inspection, and k <= -0.7 for type 1) are chosen from the data distributions, and no comparison sample of type 2 AGNs without BWB patterns was observed. As a result, the 5/9 rate cannot be interpreted as an efficiency or specificity of the method; it is a proof-of-concept demonstration that the method can find CL-like objects. The phrase 'results prove that this rather simple method can effectively find CLAGNs' overstates what a nine-object pilot with hand-tuned thresholds can establish. A control sample or a blinded prediction would be needed to support an efficiency claim.","section":"Section 2.2; Section 5"}],"minor_comments":[{"comment":"The first sentence is garbled; the SQL query description ('We used the SQL query tool of the SDSS SkyServer have decent spectra...') needs rewriting.","section":"Section 2.2"},{"comment":"The horizontal axis is labelled 'A values'; it should be 'k values'.","section":"Figure 2"},{"comment":"'none CL-AGNs' should be 'non-CL-AGNs', 'greater than 50 per cent' is colloquial, and '~2000 day' should be '~2000 days'.","section":"Section 5; Section 4.3"},{"comment":"The sentence 'the above systematic uncertainties were include' should read 'included'.","section":"Section 3.2"},{"comment":"The quoted k uncertainties (±0.001 to ±0.04) appear to come from simple linear fits that ignore the correlated nature of the magnitude errors; the fitting method should be stated briefly.","section":"Table 1"},{"comment":"The text says 'PYTHON QSO fitting code'; the language should be written as 'Python'.","section":"Section 4"}],"recommendation":"major_revision","confidential_remarks":"To the editor: The paper is a useful pilot study with a valuable candidate list, and I found no evidence of misconduct. The main issue is framing: the abstract and conclusions claim type 2 to type 1 transitions and an 'effective' selection method, while the body concedes that three of four spectroscopically confirmed sources were type 1.9 and that host dilution may explain the apparent changes. The authors should be asked to reframe the claims, quantify the dilution scenario, and be explicit about the J1150+3503 and J1203+6053 confirmation status. A control sample would strengthen the efficiency claim but may be beyond this pilot's scope."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First, what the paper actually delivers: a simple, reproducible color-magnitude screen over ~17,000 SDSS type 2 AGN spectra, producing 73 BWB candidates, 13 of which were already known CLAGNs. They followed up nine with new spectroscopy and found three new convincing CLAGNs (J0751+4948, J1344+5126, and J1203+6053 with caveats) plus two independently confirmed by Wang et al. (2024). The 73-candidate list is a concrete resource. The method is clearly described and the photometric analysis is straightforward. Credit also goes to the authors for including the discussion of the host-galaxy dilution alternative in Section 5.\n\nThe soft spot is the headline claim. The abstract states that five of nine candidates showed the CL transition from type 2 to type 1. That is not what the data show. Three of the four sources confirmed with the authors' own spectra already had broad Hα in their SDSS spectra; the paper classifies them as type 1.9, not pure type 2. One of the five (J1150+3503) is credited to Wang et al. because the DOT spectrum was bad. The one possible true turn-on, J1203+6053, rests on a noisy partial spectrum. Section 5 then concedes that host-galaxy emission was strong or dominant in the SDSS spectra and that the apparent CL phenomenon can be explained by a persistently type 1 AGN whose weak BLR is diluted by the host at low continuum states. Under that interpretation, the method is selecting strongly variable AGNs with existing broad lines, not type 2→1 transitions.\n\nThis does not destroy the paper's value. The BWB screen is a useful triage tool for finding CLAGN candidates and the candidate list stands. But the success rate 'greater than 50 per cent' is not an estimate of the method's efficiency for finding true type changes, because the targets were brightness-selected and the confirmations are heterogeneous in quality. The hand-chosen thresholds (k ≥ 0.1, k ≤ −0.7) are reasonable but not justified quantitatively.\n\nWho is this for? People building CLAGN samples from archival photometry will want to know about the 73 candidates and the BWB criterion. It is a solid application paper, not a breakthrough. With a careful revision that states the transitions are mostly from weak/intermediate broad lines to strong broad lines, and that the success rate applies to candidate selection rather than physical state changes, this deserves peer review and publication in a specialist journal. As it stands, I would recommend sending it to referees with the request to make the abstract match the evidence.","headline":"Useful candidate screen, but the 'five of nine type 2→1' claim is overstated and the paper itself concedes the main caveat.","tokens_in":19834,"tokens_out":2374,"would_cite":true,"duration_ms":20429,"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":"Five of nine colour-selected candidates showed the type 2 to type 1 changing-look transition.","keywords":["changing-look AGN","bluer-when-brighter","colour-magnitude diagram","AGN variability","type 2 AGN","broad emission lines","photometric selection","accretion mode"],"falsifier":"Obtain spectra of the remaining unobserved bluer-when-brighter candidates and compare the confirmed fraction with the 5-in-9 rate reported here; if the success rate drops well below that level, the method's efficiency does not generalise.","tokens_in":18525,"feed_emoji":"🔭","tokens_out":5806,"duration_ms":48281,"temperature":0.7,"pith_summary":"The paper tests a simple selection recipe: among AGNs previously classified as type 2, pick those whose optical colour gets bluer as they brighten. Following that pattern through recent photometric surveys produced 73 candidates, and spectroscopy of nine of them revealed five type 2 to type 1 transitions, including two already reported elsewhere. The authors argue that the bluer-when-brighter slope is an efficient and direct way to mine time-domain photometry for changing-look AGNs, and that adding flare-like amplitude and mid-infrared brightening can sharpen the selection further.","feed_headline":"Five of nine colour-selected AGN candidates changed type","feed_subtitle":"A bluer-when-brighter pattern in old type-2 AGNs flags real transitions, cheaply.","key_machinery":"The central object is the colour-magnitude diagram slope $k$ from the linear fit $z_g - z_r = k\\,z_r + c$ to ZTF photometry. A type 2 AGN with a positive slope, meaning bluer when brighter, is a candidate; the paper uses $k \\geq 0.1$ and then visually verifies the pattern. The companion diagnostic is the multi-band light-curve shape: confirmed transitions showed a flare-like brightening of $\\Delta z_g \\gtrsim 1$ mag and simultaneous mid-infrared brightening, with the W1-W2 colour moving from galaxy-like to AGN-like, while false positives had smaller optical changes and flat or decaying mid-infrared emission.","core_discovery":"The central claim is that a bluer-when-brighter colour-magnitude pattern in previously type 2 AGNs is a practical, low-cost selector of changing-look AGNs. Analysing recent $z_g-z_r$ versus $z_r$ photometry from ZTF for more than ten thousand SDSS type 2 AGNs, the paper selects 73 sources with the strongest such pattern; 13 of these were already known changing-look AGNs. New spectroscopy of nine candidates confirmed five (over 50 per cent) as type 2 to type 1 transitions, with three newly discovered. The failures, plus the three observed type 1 AGNs with redder-when-brighter slopes, show that the simple slope criterion alone is not perfect, but the colour pattern acts as a strong initial sieve.","pith_inferences":["If the bluer-when-brighter slope is a genuine precursor of broad-line turn-on, next-generation time-domain surveys could pre-select AGNs for spectroscopy years before the transition is spectroscopically confirmed.","Three of the five confirmed sources already had weak broad H-alpha in their older SDSS spectra, so the method may partly be detecting host-galaxy dilution of a persistent broad-line region; high-spatial-resolution or spectropolarimetric observations could test whether the broad lines ever truly disappear.","The false positives show that a pure colour slope can arise without a type change; combining the slope with a threshold on the amplitude of the blue flare may convert the sieve into a more precise classifier, a test the authors say they plan."],"forward_implications":["The selection pipeline can be re-run on any wide-field time-domain survey with two optical bands, producing large candidate lists for follow-up spectroscopy.","Adding a minimum blue-flare amplitude and a mid-infrared brightening condition should cut the false-positive rate seen in the four failed candidates.","The confirmed changing-look AGNs all have Eddington ratios near $10^{-2}$, supporting the picture that these transitions track a switch between accretion modes rather than extreme obscuration changes.","The same colour idea can be reversed, looking for redder-when-brighter type 1 AGNs as candidates for type 1 to type 2 transitions, though the first three observed sources did not change."],"supporting_citations":[{"why":"Supplies the initial discovery of four CLAGNs and the suggestion that the BWB pattern marks type 2 to type 1 transitions.","marker":"Zhu et al. 2024"},{"why":"Independently identified two of the five confirmed sources (J1020+2437 and J1150+3503) as changing-look AGNs, validating the selection.","marker":"Wang et al. 2024"},{"why":"Documents bluer-when-brighter behaviour in known CLAGNs and the argument that combined optical/MIR variability excludes variable obscuration.","marker":"Yang et al. 2018"},{"why":"Machine-learning selection of reclassified type 1 AGNs among type 2, giving the comparison success rate of at least 66 per cent.","marker":"López-Navas et al. 2022"},{"why":"Prior variability-based method for identifying changing-look candidates from optical light curves.","marker":"MacLeod et al. 2016b"},{"why":"Establishes mid-infrared variability and W1-W2 colour as diagnostics of accretion-rate changes in CLAGNs.","marker":"Sheng et al. 2020"},{"why":"Provides the accretion-mode transition scenario at Eddington ratio near 0.01 that the measured values support.","marker":"Noda & Done 2018"},{"why":"The PyQSOFIT fitting code used to measure broad lines and derive black-hole masses and Eddington ratios.","marker":"Guo et al. 2018"}],"fun_headline_variants":["Colour pattern flags changing-look AGNs: 5 of 9 confirmed","Bluer-when-brighter selects changing-look AGNs: 5/9 hits","Type-2 AGN colour pattern predicts type switch: 5 of 9","Low-cost colour test finds 5 changing-look AGNs among 9","Bluer-when-brighter AGN pattern: 5 of 9 turn out to change type"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The bluer-when-brighter colour pattern is assumed to trace the same intrinsic accretion change that turns on broad emission lines, rather than variable dust extinction or dilution by the host galaxy.","fun_headline_variants_meta":{"raw":{"variants":["Colour pattern flags changing-look AGNs: 5 of 9 confirmed","Bluer-when-brighter selects changing-look AGNs: 5/9 hits","Type-2 AGN colour pattern predicts type switch: 5 of 9","Low-cost colour test finds 5 changing-look AGNs among 9","Bluer-when-brighter AGN pattern: 5 of 9 turn out to change type"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00074,"raw_usage":{"total_tokens":3322,"prompt_tokens":982,"completion_tokens":2340,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":598,"completion_tokens_details":{"reasoning_tokens":2227}},"tokens_in":598,"tokens_out":2340,"duration_ms":16061,"temperature":1.0,"reasoning_tokens":2227,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T14:07:10.206929+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Obtain spectra of the remaining unobserved bluer-when-brighter candidates and compare the confirmed fraction with the 5-in-9 rate reported here; if the success rate drops well below that level, the method's efficiency does not generalise.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The PyQSOFIT fitting code used to measure broad lines and derive black-hole masses and Eddington ratios."}],"review_version":1}