{"id":"27d1faa1-39c9-4957-9d19-7400c85de7ea","arxiv_id":"2412.00855","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"A SkyMapper-based selection using reconstructed uvg colours and three-exposure u-band variability discovers 12 new symbiotic stars, four with accretion-disc flickering, and implies at least 20% of symbiotics flicker.","lead":"Using archival SkyMapper survey photometry, the authors identify 12 new symbiotic star binaries and 10 candidates, including four with optical accretion-disc flickering and two with hard X-rays. The work offers a survey method that avoids the H-alpha line selection bias of past symbiotic searches and suggests a hidden population of accreting-only symbiotics.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 20% population claim rests on a 6/30 rate whose numerator and denominator are drawn from different, selection-biased samples; the paper itself admits the duty cycle is unconstrained.","rationale":"The paper's discovery claims are well supported: the LCO flickering identifications pass extensive validation against check stars and aperture/systematics tests, and the Chandra spectra of SkySy 1-2 and SkySy 1-4 provide independent accretion signatures. My concern is isolated to the population-level extrapolation in Section 7.3.3. The 6/30 statistic is presented as a rate, but the six confirmed flickerers are not a random sample of the 30: four were discovered through the same sigma_delta_u/colour selection, and LCO follow-up was limited to a priority list of 11 targets. The remaining 24 objects in the denominator were largely not given the same opportunity to be counted. The paper explicitly states that the duty cycle is unconstrained, which further undermines the idea that a single sequence can classify an object as non-flickering. This is not an internal inconsistency in the detections, but it is a missing statistical link between the sample and the \"true population\" statement. The reader's CONDITIONAL verdict is exactly right: the discoveries stand, but the 20% headline should be reframed as a biased discovery yield or replaced by a number derived from a defined selection function. Hence no verdict change.","tokens_in":49975,"tokens_out":7854,"duration_ms":78229,"concrete_test":"Take all 30 objects in the Section 7.3.3 denominator and observe every one of them with a uniform LCO B-band protocol (3-5 hours, same cadence and duration as Table 2), with repeated epochs to sample the duty cycle, without pre-selecting on sigma_delta_u or colour. Recompute the confirmed flickering fraction over this fully observed, pre-registered sample. If the fraction is significantly below 20%, or if the original six flickerers are systematically brighter or have higher sigma_delta_u than the remaining 24, the population-level claim should be replaced by a discovery-yield statement with explicit selection-bias caveats.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim — \"at least 20% of the true population of symbiotics exhibit detectable optical flickering ... the majority of which do not meet the H-alpha detection thresholds\" — rests on Section 7.3.3's \"6 out of 30\". The load-bearing condition is that 6/30 is an incidence rate in a representative sample. It is not. The numerator and denominator come from different sets: the denominator is the 30 known symbiotics plus new SkySy that happen to have a sigma_delta_u measurement in the luminous-red-object sample, but the numerator is the six objects confirmed in LCO fast photometry, and Section 5.2/Table 2 show follow-up covered only 11 targets chosen by colour, sigma_delta_u, and discovery priority, not a random or complete draw from those 30. Four of the six flickerers are new SkySy selected partly because they were SkyMapper outliers; the confirmation is therefore enriched by the same selection metric being tested. The paper itself admits the duty cycle is unconstrained (Section 7.3.3), so a single 20-minute sequence, or even one LCO run, cannot certify non-flickering for the other 24. The denominator is also restricted by the J-Ks>0.85, MJ<0, isolation, and uvg-availability cuts (Sections 2.1, 2.2, 2.4), which are not the symbiotic population. Without a defined selection function, 6/30 is a discovery yield, not a population lower bound; the \"majority missed by H-alpha surveys\" corollary inherits the same bias.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a new method for discovering symbiotic stars using SkyMapper survey photometry: reconstructed uvg snapshot colours and u-band variability among the three exposures of each 20-minute filter sequence (the sigma_Delta_u metric), applied to a sample of 366,721 luminous red objects. The authors selected 234 objects for optical spectroscopy, discovered 12 new symbiotics and 10 candidates, and reported optical accretion-disc flickering in four new symbiotics and likely in the known symbiotic V1044 Cen, with hard X-ray detections in two of the flickerers. They compare their sample with recent GALAH, IR-colour, and TESS searches and conclude that at least 20% of the true symbiotic population has detectable optical flickering, most of which would be missed by H-alpha-based surveys.","tokens_in":50282,"tokens_out":6572,"duration_ms":61317,"significance":"If the method holds up, it is a valuable new route to finding accreting-only symbiotic stars, a population that is plausibly underrepresented in H-alpha and objective-prism surveys. The core discoveries are supported by independent LCO fast photometry with careful check-star and aperture validation, and by Chandra/Swift X-ray detections in two cases. The paper also makes a useful methodological contribution by showing that SkyMapper catalogue colours of pulsating giants must be reconstructed from nightly snapshot colours, and that flickering can be recognised from only three u-band exposures. The classification tree is explicit and transparent, and the LCO, Chandra, and Swift data are public. The main weakness is that the headline population fraction, 'at least 20% of the true population', is not supported by the presented 6/30 statistic because the denominator is a selection-biased subsample and the numerator is drawn from a different, targeted follow-up set.","major_comments":[{"comment":"The statement that 'at least 20% of the true population of symbiotics exhibit detectable optical flickering' is not supported by the 6/30 statistic as presented. The denominator is the 30 objects with a sigma_Delta_u measurement that survive the luminous-red-object cuts (Sections 2.1 and 2.2: (J-Ks)_0>0.85, M_J<0, isolation, and uvg availability), not a random or representative sample of the symbiotic population. The numerator is drawn from the 11 targets followed with LCO fast photometry (Table 2), which were selected by colour, sigma_Delta_u, and discovery priority (Section 4), so 6/30 is a discovery yield rather than an incidence rate. Four of the six flickerers are new SkySy that were selected partly because they were SkyMapper outliers by the same colour/variability metric being tested, further enriching the numerator. In addition, the paper itself acknowledges in Section 7.3.3 that the duty cycle of flickering is unconstrained, so a single 20-minute sequence or one LCO run cannot certify non-flickering for the other 24 objects. Without a defined selection function, the 'majority missed by H-alpha surveys' corollary inherits the same bias. Please remove the population claim or explicitly recast 6/30 as a discovery yield in a selected sample, with a clear statement that no population lower bound can be inferred from the current data.","section":"Section 7.3.3 (and abstract, Section 8 item vii)"},{"comment":"The effective 2- and 3-sigma thresholds are calibrated to the sample itself: sigma_eff=2 and 3 correspond to the 95.45% and 99.73% percentiles of sigma_Delta_u in the same luminous-red-object sample, and Section 4(v) states that sigma_Delta_u was intentionally chosen so that the known flickering symbiotic EF Aql would stand out. Consequently, the fraction of objects 'detectable through SkyMapper sigma_Delta_u' is partly a property of the calibration sample, and using these tuned thresholds as a completeness metric is circular. The independent LCO and Chandra confirmations validate the individual detections, but they do not validate the use of the tuned thresholds for population completeness. Please quantify the sensitivity of the claimed fraction to threshold choice, or apply the full selection to an independent sample, or state explicitly that the thresholds are empirical tuning parameters rather than calibrated detection limits.","section":"Section 3.2 (Eq. 1), Section 4(v), and Section 7.3.3"},{"comment":"The u-g<2.9 cutoff is introduced after inspecting the results as the 'abrupt cutoff' redward of which sigma_Delta_u>3sigma_eff is generally spurious. This boundary is not specified in the target-selection criteria of Section 4 and is derived from the same dataset used to define the search. As a result, the prescription in Section 8 item (iv) that future searches use u-g<2.9 is a post hoc selection rule rather than a validated cut. The authors should present this boundary as a retrospective diagnostic, or validate it on an independent sample, before recommending it as a search criterion. This does not affect the reality of the discovered flickerers, but it does affect the generalizability of the proposed method.","section":"Section 7.2, Fig. 14, and Section 8 item (iv)"}],"minor_comments":[{"comment":"The selection function is described only in prose and a table; a machine-readable list of the 366,721 objects with pass/fail flags for each cut would facilitate completeness corrections and re-use of the method.","section":"Table 1 and Section 2.2"},{"comment":"The sentence 'clipped up to one (two) manually-determined outlying non-flagged suspicious measurement per source' is unclear: please specify whether the one/two refers to target versus check stars, and how the manual outlier choice was made consistently across light curves.","section":"Section 5.2.1, step (v)"},{"comment":"The H-alpha |pEW| entries such as '11 (7)' and '6 (2)' need a footnote explaining what the parenthetical values represent (for example, a second epoch or an alternative continuum definition).","section":"Table 4"},{"comment":"The text describes V1044 Cen's flickering as detected 'with moderate confidence' and 'likely,' but Table 4 lists a plain 'Y' flag for optical flickering; use a 'Y?' or 'likely' flag to match the stated confidence.","section":"Section 6.2 and Table 4"},{"comment":"The statement that '5 out of 7 with SkyMapper sigma_Delta_u>2sigma_eff exhibited LCO B-band flickering' should explicitly note that the denominator is the LCO follow-up subset, not the full set of objects with sigma_Delta_u measurements, to avoid confusion with the 6/30 statistic in Section 7.3.3.","section":"Section 7.2, penultimate paragraph"}],"recommendation":"major_revision","confidential_remarks":"The core observational discoveries appear sound and are well validated with independent data, but the abstract's 'at least 20% of the true population' claim is the headline and is not supported by the presented selection-bias analysis. The paper should be acceptable after the population claim is removed or explicitly reframed as a discovery yield, and after the post hoc nature of the u-g cutoff and threshold calibration is acknowledged. No concerns about novelty or scope: the method and the new sources are of clear interest to the symbiotic-star and time-domain communities."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know. First, the SkyMapper survey method works: the sigma_delta_u flickering statistic plus reconstructed u-g/u-v colours found 12 new symbiotics, four with well-validated optical flickering and two with hard X-rays. Those are genuine discoveries, confirmed with independent LCO photometry and Chandra/Swift data, and the systematics testing on the light curves is thorough. Second, the headline population claim—'at least 20% of the true population of symbiotics flicker'—is built on a 6/30 ratio that is not a population statistic. The denominator is 30 known plus new SkySy with a sigma_delta_u measurement; the numerator is six objects confirmed by LCO follow-up. But the LCO targets were selected partly by the same sigma_delta_u and colour criteria being tested, and the follow-up was not a complete or random draw. The paper itself admits the duty cycle is unconstrained, so a 20-minute sequence or one LCO run cannot certify non-flickering for the other 24. The sample is further restricted by J-Ks>0.85, MJ<0, isolation, and uvg availability cuts. So 6/30 is a discovery yield, not a lower limit, and the corollary that most flickering symbiotics are missed by H-alpha surveys inherits the same bias.\n\nWhat is genuinely new: the sigma_delta_u statistic itself, the snapshot colour reconstruction that fixes Miras in colour-colour space, and the demonstration that flickering is detectable with only three survey exposures. The comparison to GALAH GaSS, Akras IR trees, and TESS flickering is thoughtful and places the method in context. The SkySy/SkySyC classification is conservative and clearly spelled out.\n\nSoft spots beyond the 20% claim: the effective sigma thresholds are calibrated to sample percentiles, and the u-g<2.9 cutoff is post hoc. The paper is honest about this, and it does not undermine the confirmed detections, but the method's calibration should be described as empirical rather than physical. Also the control group without sigma_delta_u is tiny (0–1 out of 4), which is fine for a demonstration but not for a rate.\n\nThis paper is for anyone working on symbiotic stars, accreting white dwarfs, or survey methods for variability plus colour outliers. It deserves a serious referee. My recommendation: major revision. Keep the discoveries and the method; reframe the 20% claim as 'we detect flickering in a substantial fraction of our targeted sample' or re-derive it with an explicit selection function and duty-cycle constraints.","headline":"Solid method and credible discoveries; the 'at least 20%' population claim is not supported as stated and needs a major rethink.","tokens_in":50909,"tokens_out":3688,"would_cite":true,"duration_ms":31544,"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":"The paper finds that at least 20% of symbiotic stars show detectable accretion-disc flickering, and most of these are missed by H-alpha surveys, making previous censuses substantially incomplete.","keywords":["symbiotic stars","accretion discs","optical flickering","SkyMapper survey","survey photometry","H-alpha surveys","white dwarf binaries","variable stars"],"falsifier":"Re-observe the 24 objects in the paper's sample that have $\\sigma_{\\Delta u}$ measurements but showed no flickering in the LCO light curves, using multi-night, multi-hour continuous photometry (for example with TESS or a dedicated ground-based network). If a large fraction of these turn out to flicker, then the 20% figure is an underestimate set by the duty cycle; if none of them flicker in extended monitoring, the 6/30 rate is a real sub-population fraction and the representativeness of the 30-object denominator becomes the decisive test.","tokens_in":49761,"feed_emoji":"⭐","tokens_out":10254,"duration_ms":85027,"temperature":0.7,"pith_summary":"This paper aims to prove that a substantial population of symbiotic stars—binaries in which a cool evolved giant accretes onto a white dwarf—has been missed by traditional H-$\\alpha$ surveys because the accretion-disc signature is drowned out by the giant's light. It builds a new detection method from two photometric fingerprints in the SkyMapper survey: a blue excess in reconstructed $u-g$ and $u-v$ colours, and rapid variability between the three $u$-band exposures within each 20-minute visit, a quantity the paper calls $\\sigma_{\\Delta u}$. Testing the method on 366,721 luminous red objects, the authors discovered 12 new symbiotics, 10 candidates, and four new optically flickering symbiotics, two of which also show hard boundary-layer X-rays. Their central inference is that at least 20% of all symbiotics exhibit detectable accretion-disc flickering, and that the majority of these would be missed by narrow-band H$\\alpha$ surveys, making previous symbiotic censuses substantially incomplete. If correct, the method offers a less biased route to finding accreting-only symbiotics and a concrete way to quantify how many are hiding.","feed_headline":"Flicker reveals a hidden population of symbiotic stars","feed_subtitle":"Using three u-band exposures per SkyMapper visit, the method finds accreting-only symbiotics that H-alpha surveys miss.","key_machinery":"The load-bearing machinery is the SkyMapper Main Survey visit, which cycles through $u,v,g,r,i,z$ filters in about 20 minutes and thus snapshots three 100-second $u$-band exposures separated by roughly 8 minutes. The paper's variability metric, $\\sigma_{\\Delta u}$, is the maximum, over pairs of exposures within one sequence, of the magnitude difference divided by the quadrature sum of photometric uncertainties (Equation 1); because the noise is systematics-dominated, it is calibrated against flat effective thresholds ($2\\sigma_{\\mathrm{eff}}=9$, $3\\sigma_{\\mathrm{eff}}=18$) that place roughly 95% and 99.7% of the sample below them. The second critical component is the reconstruction of 'nightly snapshot' $u-g$ and $u-v$ colours from the table of individual exposures rather than the clipped master catalogue, without which Mira-like pulsating giants scatter across the colour-colour diagram with unphysical colours. The paper shows that a colour cut at $u-g\\lesssim 2.9$ separates genuine flickering from spurious $\\sigma_{\\Delta u}$ detections, because flickering is only visible in $u$ when the accretion disc contributes enough excess light.","core_discovery":"The paper's central claim is that optically flickering, accreting-only symbiotic stars are not a rare side effect of the class but a sizeable, systematically undercounted subpopulation. The discovery is carried by two signatures extracted from archival SkyMapper photometry: the position of a source in the reconstructed $u-g$/$u-v$ colour-colour diagram, which isolates the hot component's short-wavelength excess, and $\\sigma_{\\Delta u}$, the maximum pairwise significance of variability among the two or three $u$-band exposures within a single 20-minute Main Survey filter sequence. Within a curated sample of 366,721 luminous red objects, the authors map a 'mostly-symbiotic zone' in the colour-colour plane where almost every object is a symbiotic, and show that flickering can be recovered from only three data points. They report 12 new symbiotics and 10 candidates; among the seven objects with $\\sigma_{\\Delta u} > 2\\sigma_{\\mathrm{eff}}$ that were followed with hours-long $B$-band light curves, five flicker, and the two observed with deep X-ray exposures show hard boundary-layer emission. From the six flickerers among the 30 symbiotics in their sample with a $\\sigma_{\\Delta u}$ measurement, the paper concludes that at least 20% of true symbiotics have optically detectable flickering, and that most of these have H$\\alpha$ pseudo-equivalent widths below the 10–50 Å thresholds used by narrow-band surveys.","pith_inferences":["Editorial inference: the 20% flickering fraction is probably a lower bound on the underlying population because it is averaged over one or two 20-minute snapshots per object; if the duty cycle of flickering is well below unity, the fraction of symbiotics that ever flicker could be much higher, while the fraction detectable in a single visit could be lower.","Editorial inference: the paper's 234-object spectroscopic survey is effectively a labelled training set in SkyMapper colour-variability space; a supervised classifier trained on it could search later data releases more efficiently, and would be a natural next step beyond the hand-tuned thresholds.","Editorial inference: the absence of $\\sigma_{\\Delta u}$ excess in the GALAH accreting-only candidates, combined with the authors' own low-amplitude findings, suggests those systems flicker at lower amplitude or in different bands; testing the SkyMapper method on a larger GALAH intersection after cross-matching colour data could unify the two discovery channels.","Editorial inference: if the majority of flickering symbiotics are accreting-only with weak H$\\alpha$, the space density of symbiotic stars may be substantially higher than current catalogs imply, with consequences for the rates of Type Ia supernova progenitors and for post-AGB binary evolution; this extrapolation goes beyond the paper's own claims but follows from its completeness argument."],"forward_implications":["Narrow-band H$\\alpha$ surveys have missed a substantial fraction of symbiotic stars: the paper estimates that 51–88 missing $u<16$ symbiotics (11–17 reported here) are recoverable from the colour-colour diagram alone with near-zero contamination.","With future all-sky $uvg$ photometry from later SkyMapper releases and MEPHISTO, the same method is expected to find an estimated 68–117 symbiotics missed by previous surveys, many of them accreting-only.","If at least 20% of symbiotics flicker, then true symbiotic populations are dominated by accreting-only systems, which would bias past statistics on the burning versus non-burning ratio and on the evolutionary state of the donors.","The $\\sigma_{\\Delta u}$ plus colour selection can directly supply targets for fast photometry and X-ray follow-up, as demonstrated by the two flickerers that turned out to emit $\\delta$-type boundary-layer X-rays.","The concordance between the SkyMapper method and IR colour selection (10 of 12 new symbiotics pass the general IR decision tree) indicates the new candidates are genuine symbiotics despite their weak Balmer lines."],"supporting_citations":[{"why":"The only previously known flickering symbiotic in the survey footprint; used to calibrate the $\\sigma_{\\Delta u}$ variability metric and as the proof-of-concept target.","marker":"Zamanov et al. 2017"},{"why":"Documents the SkyMapper Main Survey filter sequence with three $u$ exposures per 20-minute visit that makes the flickering detection possible.","marker":"Onken et al. 2019"},{"why":"Provides the Gaia DR2 distance estimates used for the $M_J < 0$ luminosity cut that defines the luminous-red-object sample.","marker":"Bailer-Jones et al. 2018"},{"why":"The catalog of previously known symbiotics used to cross-match, define completeness, and evaluate what the new method finds.","marker":"Merc et al. 2019"},{"why":"Sets the 50 Å H$\\alpha$ pseudo-equivalent width threshold for narrow-band symbiotic surveys that the new method claims to bypass.","marker":"Corradi et al. 2008"},{"why":"Provides the definition of hard ($\\delta$-type) X-rays and the boundary-layer model used to interpret the Chandra and Swift detections.","marker":"Luna et al. 2013"},{"why":"The GALAH accreting-only symbiotic sample; the comparison shows the SkyMapper method probes a different parameter regime.","marker":"Munari et al. 2021"},{"why":"The IR colour machine-learning selection; 10 of 12 new symbiotics meet its general tree, supporting real overlap in selection.","marker":"Akras et al. 2019b"},{"why":"Previous characterization of symbiotic flickering timescales that motivates using the 20-minute SkyMapper window.","marker":"Lucy et al. 2020"},{"why":"Theoretical basis for flickering timescales that distinguishes white-dwarf accretors from main-sequence accretors.","marker":"Sokoloski & Bildsten 2010"}],"fun_headline_variants":["Flicker finds hidden symbiotic stars","Accretion flicker exposes new symbiotics","SkyMapper flicker reveals missing symbiotics","Flicker spots symbiotics H-alpha surveys miss"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The 20% population fraction rests on the assumption that the 30 symbiotics with $\\sigma_{\\Delta u}$ measurements are a representative sample of the true symbiotic population, and that the unconstrained duty cycle of flickering does not bias the 6-out-of-30 detection rate.","fun_headline_variants_meta":{"raw":{"variants":["Flicker finds hidden symbiotic stars","Accretion flicker exposes new symbiotics","SkyMapper flicker reveals missing symbiotics","Flicker spots symbiotics H-alpha surveys miss"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000218,"raw_usage":{"total_tokens":1546,"prompt_tokens":1160,"completion_tokens":386,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":776,"completion_tokens_details":{"reasoning_tokens":329}},"tokens_in":776,"tokens_out":386,"duration_ms":3710,"temperature":1.0,"reasoning_tokens":329,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T04:55:59.232289+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Re-observe the 24 objects in the paper's sample that have $\\sigma_{\\Delta u}$ measurements but showed no flickering in the LCO light curves, using multi-night, multi-hour continuous photometry (for example with TESS or a dedicated ground-based network). If a large fraction of these turn out to flicker, then the 20% figure is an underestimate set by the duty cycle; if none of them flicker in extended monitoring, the 6/30 rate is a real sub-population fraction and the representativeness of the 30-object denominator becomes the decisive test.","supporting_citations":[],"review_version":1}