{"id":"338d7f8c-3e4c-4cef-aaf9-84bc2eee1c78","arxiv_id":"2501.13147","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"A five-type taxonomy of super-orbital optical variability for 111 Small Magellanic Cloud high-mass X-ray binaries connects light-curve shape, color, and orbital period to Be star disk behavior.","lead":"Using 20 years of OGLE optical data, this paper sorts 111 high-mass X-ray binaries in the Small Magellanic Cloud into five types of long-term brightness variability. The types line up with color changes and orbital periods, suggesting the variability is driven by a gas disk around the Be star growing and shrinking.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The taxonomy's Type 3/4/5 boundaries depend on an untested detrending choice: the stochastic variability metric uses wotan rspline (window=200, break_tolerance=50) but no sensitivity analysis is given, so sources could cross the 0.40/0.75 cutoffs and change the claimed correlations.","rationale":"The reader's weakest assumption identifies the same load-bearing concern: the stochastic variability metric's sensitivity to the choice of spline detrending. This is the most critical assumption because the taxonomy's boundaries between Types 3, 4, and 5 are defined directly by this metric, and the paper's physical interpretation and population-level correlations (Sect. 7.4) rely on the stability of those boundaries. The paper itself flags the sensitivity but does not quantify it. The proposed test directly addresses this by varying the detrending parameters and checking whether the types and conclusions change. I agree with the reader's CONDITIONAL verdict: the study is transparent, data-rich, and the taxonomy is a useful heuristic, but the untested detrending sensitivity prevents full acceptance. No other concern appears more load-bearing: the base-number metric is more robust (using third extrema), the period search is partially validated against X-ray periods, and the physical interpretation is clearly labeled as such. Thus the verdict should remain CONDITIONAL; the paper would become acceptable if the sensitivity analysis is added and the affected conclusions are shown to be stable.","tokens_in":41102,"tokens_out":4641,"duration_ms":50670,"concrete_test":"Recompute the stochastic variability metric for all 111 sources using wotan rspline with windows of 100, 200, 400, and 800 days and break tolerances of 25, 50, and 100, keeping all other preprocessing identical. Tabulate how many sources cross the Type 3/4 boundary at 0.40 and the Type 4/5 boundary at 0.75, and then re-run the Sect. 7.4 comparisons (spin-period distributions, fraction with 2–200 d periodogram peaks above FAL=0.001, and CMD locations) using the reassigned types. If the qualitative conclusions change or lose statistical significance for any reasonable parameter choice, the taxonomy is not robust to detrending and the physical interpretation should be treated as provisional.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that the base number and stochastic variability metric define a meaningful five-type taxonomy, with Types 1–3 interpretable as red Be-disk growth/depletion and Types 4–5 as low-coherence states. The stochastic variability metric, which sets the Type 3/4 and Type 4/5 boundaries (Table 1), is defined as the ratio of the standard deviation of a spline-detrended light curve to that of the original light curve, using wotan rspline with window=200 and break_tolerance=50 (Sects. 3 and 5). The paper explicitly acknowledges this metric is sensitive to detrending (Sect. 3: \"This metric is sensitive to the detrending approach... so we use consistent, reasonable parameters across the sample\"), but no test is reported showing how the choice of window or break tolerance affects the resulting types. This matters because the spline defines what counts as \"coherent super-orbital variability\" versus \"stochastic variability\": a more flexible spline (shorter window) will absorb more real variability, lowering the metric and pushing sources toward Type 3, while a stiffer spline will leave coherent signal in the residuals, raising the metric and pushing sources toward Type 4 or 5. Sources with super-orbital timescales near the 200-day window (e.g., SXP 756's 393-day modulation) are particularly vulnerable. Since the same detrending is used for the periodicity search, the confidence categories and the statement that no Type 1 source has a confident orbital period could also shift. If a moderate change in window moves a substantial fraction of sources across the 0.40 or 0.75 cutoffs, the reported differences between Types 4 and 5 (spin-period distributions, periodicity detection rates, CMD positions) and the physical interpretation in Sect. 7.4 would not be robust.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"Using OGLE II-IV I- and V-band light curves of 111 SMC HMXBs, the paper defines a five-type taxonomy of super-orbital optical variability based on two metrics: the 'base number' (difference between median brightness and the third-brightest/third-faintest points) and a 'stochastic variability metric' (ratio of spline-detrended to original standard deviation, computed with wotan rspline at window=200 and break_tolerance=50). Types 1-3 are interpreted as Be-disk growth/depletion states (faint blue base with flares; bright red maximum with dips; intermediate), while Types 4-5 are low-coherence states, with Type 5 possibly containing non-BeXRB interlopers. The paper also performs Lomb-Scargle period searches, reports 33 confident orbital-period candidates, validates the optical search against 24 X-ray orbital periods (75% same or harmonic), discusses CMD loops, and connects types to spin periods, colors, and H-alpha emission.","tokens_in":41402,"tokens_out":8708,"duration_ms":84961,"significance":"The paper's value is in presenting a large, homogeneous, publicly available dataset and a transparent attempt to systematize super-orbital variability of SMC HMXBs. The periodicity validation against X-ray periods and the use of external disk models (de Wit et al. 2006; Haubois et al. 2012) are strengths, as is the explicit acknowledgment that the taxonomy is a heuristic tool. If the taxonomy and its physical interpretation survive robustness checks, it would provide a useful population-level framework for connecting optical light-curve morphology to Be-disk state, orbital architecture, and NS spin. At present the central classification rests on reported but untested analysis choices, so the quantitative type-dependent correlations should be treated as provisional.","major_comments":[{"comment":"Table 1 defines Types 1-3 as having a stochastic variability metric <0.35 and Type 4 as >0.40 and <0.75, leaving the interval 0.35-0.40 formally undefined. This conflicts with Table B.3, which lists sources with metrics of 0.36 (Source 5), 0.37 (Source 35), and 0.40 (Source 77) as Type 4. Please state the intended boundary and ensure Table 1, Fig. 1, and Table B.3 are mutually consistent; the discrepancy matters because sources in this interval sit exactly at the Type 3/4 transition.","section":"Sect. 3, Table 1"},{"comment":"The stochastic variability metric, which sets the Type 3/4 and Type 4/5 boundaries, is explicitly sensitive to the spline-detrending choice (wotan rspline, window=200, break_tolerance=50; Sect. 5), but no sensitivity analysis is reported. A shorter window would absorb more real super-orbital signal and lower the metric, while a stiffer spline would leave coherent signal in the residuals and raise it. Several sources lie near the cutoffs (e.g., Source 44: metric 0.77; Source 5: 0.36; Source 77: 0.40; Source 58/SXP 756 has a 393-day modulation near the 200-day window). Please quantify how the type distribution, the confident-period fractions quoted in Sect. 7.3, and the Type 4/5 spin-period statistics in Sect. 7.4 change for plausible alternative values (e.g., window 100, 300, 400 days; break_tolerance 20, 100).","section":"Sect. 3, Sect. 5, Table 1"},{"comment":"The super-orbital periodicity search is described only as 'we checked for super-orbital periodicities,' but all period searches in Sect. 5 are otherwise performed on data detrended with a 200-day spline. If the same detrended light curves are used for the >200-day search, the spline would suppress exactly the signal being sought, making the conclusion that there are 'no true super-orbital periodicities' an artifact of the detrending. Please state explicitly whether original or detrended light curves were used for Sect. 5.5 and, if detrended, repeat the search on the original data (or with a very stiff spline), and report any peaks above the FAL.","section":"Sect. 5.5 (and Sect. 5.2)"}],"minor_comments":[{"comment":"For the 21 sources with manual OGLE-IV calibration (median set to the prior median), please state how many lie within 0.1 in base number or 0.05 in stochastic metric of the nearest type boundary, since the median adjustment directly enters the base number definition.","section":"Sect. 2"},{"comment":"The confidence criteria (range/error >9, at least one-third of rolling periodogram peaks within 5%, mean phase change ≤2 bins) are described as chosen by visual inspection; reporting the distributions of these three metrics and the number of sources passing each cut would make the 'confident' list more reproducible.","section":"Sect. 5.2"},{"comment":"The caption contains a typo: 'Savitzy-Golay' should be 'Savitzky-Golay'.","section":"Fig. 7 caption"},{"comment":"The mean spin-period comparison between Type 2 (166 s) and Type 4 (743 s) is driven by a long-period tail; please report sample sizes and a significance test (e.g., two-sample KS test) before using it as evidence that Type 4 systems are longer-timescale versions of Type 2.","section":"Sect. 7.4"},{"comment":"In the transition analysis, please state the number of Type 1-3 sources used for the 84%/54% statistics and whether the 1000-day segments are independent, so the reader can assess the uncertainty on those fractions.","section":"Sect. 3"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within the journal's scope and the dataset is valuable. The main revisions I would require are a sensitivity analysis for the stochastic metric, a clarification and possible rerun of the super-orbital periodicity search, and a fix of the Table 1 boundary inconsistency. These are feasible with the existing code and should not require new observations. I would be willing to accept after these points are addressed. No concerns about novelty disclosure or citation practice."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's the quick take: this is a solid, honest observational paper that will be the reference point for SMC HMXB optical variability for a while. It does three genuinely new things: a quantitative two-metric taxonomy of super-orbital behavior applied to 111 sources, five new confident orbital periods, and a demonstration that the previously claimed super-orbital periodicities don't hold up with a longer baseline. The period-search section is careful, especially the calibration against X-ray periods and the explicit discussion of harmonics. The CMD loop analysis and link to de Wit/Haubois disk models are thoughtful. Credit where due: the paper is transparent about its choices and flags its own caveats.\n\nThe soft spots are real but not fatal. The Type 3/4/5 boundaries rest on the stochastic variability metric, which is explicitly sensitive to the wotan rspline detrending (window=200, break_tolerance=50). The paper says it used consistent, reasonable parameters but does no sensitivity test. The stress-test worry is legitimate: a shorter window would absorb more real coherent variability and push sources toward Type 3; a stiffer spline would push them toward Type 4/5. Since the same detrending feeds the periodicity search, the confidence categories could shift too. Without a robustness check, the claimed differences between Type 4 and Type 5 (spin, periodicity detection rates, CMD position) are on shakier ground than the rest of the paper. This is an addressable gap, not a load-bearing flaw. The taxonomy is explicitly heuristic, the authors say so, and the broad Type 1 vs Type 2 vs Type 3 distinctions are robust to reasonable detrending choices. The Type 4/5 split is the fragile part.\n\nMinor complaints: 21 sources got manual OGLE-IV calibration, which is acknowledged but adds unknown systematics. Code isn't public, only available on request. The period confidence cutoffs are qualitative, though they are at least described in detail.\n\nBottom line: a serious referee should engage with this. The taxonomy will be cited, the period results are useful, and the physics discussion is sensible. I'd bring it to a reading group and I'd cite it if I worked on BeXRBs. My own verdict would be accept after minor revision, with a request for a robustness test of the detrending parameters and a version of the figures with misclassification bands if feasible.","headline":"A data-rich, transparent taxonomy of SMC HMXB super-orbital variability that is genuinely useful despite a detrending-sensitivity blind spot in the Type 3/4/5 boundaries.","tokens_in":42061,"tokens_out":1765,"would_cite":true,"duration_ms":17571,"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":"Two photometric metrics organize 111 SMC high-mass X-ray binaries into five super-orbital variability types, with Types 1–3 reflecting a red Be-star disk growing and depleting.","keywords":["Small Magellanic Cloud","high-mass X-ray binaries","Be stars","optical variability","OGLE","super-orbital variability","decretion disk","X-ray binaries"],"falsifier":"Recompute the stochastic variability metric for all 111 sources with a range of spline windows and break tolerances (e.g., windows from 100 to 400 days, tolerances from 20 to 200) and count how many sources cross the 0.40 or 0.75 cutoffs; if a substantial fraction move between Types 4 and 5, the taxonomy's boundary is an artifact of detrending. Alternatively, take a Type 1 source during a flare and a Type 2 source during a dip, obtain contemporaneous spectroscopy (H-alpha or V/R line profiles), and check whether the observed disk state matches the predicted growing-versus-depleting phase; a mismatch would refute the disk-growth interpretation.","tokens_in":40863,"feed_emoji":"🔭","tokens_out":5974,"duration_ms":49384,"temperature":0.7,"pith_summary":"The paper argues that twenty years of OGLE photometry of 111 high-mass X-ray binaries in the Small Magellanic Cloud can be organized by two numbers: a 'base number' measuring whether a light curve returns to a faint or a bright floor, and a 'stochastic variability metric' measuring how much short-term scatter remains after removing long-term trends. These two numbers define five super-orbital variability types. The paper claims that Types 1–3 are stages of a single physical picture, a red Be-star disk growing and depleting, with Type 1 sources showing a bare-star baseline with flares, Type 2 sources showing a maximal disk with dips, and Type 3 sources intermediate. Type 4 sources appear to be slower Type 2 systems, while Type 5 sources largely lack coherent variability and may include interlopers. If correct, optical light curves alone can classify disk state and connect it to colors, orbital periods, and X-ray behavior across the SMC BeXRB population.","feed_headline":"Five-type taxonomy links SMC X-ray binaries to growing disks","feed_subtitle":"Twenty years of OGLE data show flaring, dipping, and quiet sources as stages of one Be-star disk cycle.","key_machinery":"The taxonomy rests on two metrics computed from OGLE I-band light curves. The base number, $(\\text{Max}\\,I - \\text{Median}\\,I) - (\\text{Median}\\,I - \\text{Min}\\,I)$, uses the third-brightest and third-faintest points to measure whether a light curve returns to a consistent faint level (negative; flares) or a consistent bright level (positive; dips). The stochastic variability metric is the ratio of the standard deviation of the detrended light curve (spline-detrended with wotan's rspline, window 200, break tolerance 50) to that of the original data, measuring how much short-term variability remains after removing super-orbital signal. Cuts in this two-dimensional plane at base-number $-0.55$ and $0.15$ and stochastic-metric $0.40$ and $0.75$ define Types 1–5.","core_discovery":"The central claim is that the base number and stochastic variability metric define a meaningful five-type taxonomy of super-orbital optical variability for SMC HMXBs, and that Types 1–3 can generally be interpreted as a relatively red Be decretion disk growing and depleting. Type 1 light curves have a faint blue baseline consistent with a nearly bare B star, with intermittent flares corresponding to disk growth events. Type 2 light curves return to a bright red level that represents the source's maximal disk, with dips corresponding to partial depletion. Type 3 sources show coherent variability without reaching either extreme. The paper further claims that Type 4 sources are longer-timescale versions of Type 2 that may show depletion events in future data, and that Type 5 sources, especially a faint blue cluster, may largely be interlopers rather than BeXRBs.","pith_inferences":["The same two-metric plane could be computed for isolated Be stars in the SMC and LMC; comparing their distribution to the HMXB distribution would isolate the compact object's effect on the disk.","Because the metrics are survey-agnostic, OGLE, TESS, ASAS-SN, and Rubin data could be placed on the same plane, offering a data-driven way to track type transitions as light curves lengthen.","The claim that Type 4 sources become Type 2 after depletion events predicts a specific temporal sequence within individual light curves; a systematic search for such transitions in the existing OGLE data would test it.","If Type 5 sources are mostly interlopers, then the remaining Type 5 BeXRBs may represent a distinct class of persistently stable disks, potentially probing the long end of the disk-growth timescale distribution."],"forward_implications":["Type 1 sources, with their bare-star baselines and near-identical ~1000-day flare spacings, can be used as a clean sample to study disk growth events in isolation.","Type 2 bright red baselines mark each system's maximal disk; comparing H-alpha equivalent widths across Type 2 sources would directly test the truncation model.","No SMC BeXRB shows a persistent super-orbital periodicity; quasi-periodic behavior is locally stable but changes over the ~8000-day baseline, so correlations between orbital and super-orbital periods should be treated with caution.","Confident orbital periodicities are found in 25% of Type 2 and 41% of Type 3 sources but in none of the Type 1 sources, consistent with the idea that an extended disk is needed to produce an orbital signal.","The faint blue cluster of Type 5 sources, if confirmed as interlopers, would revise the SMC HMXB census."],"supporting_citations":[{"why":"Supplies the base catalog of SMC HMXBs, spin periods, and X-ray orbital periods that define the sample.","marker":"HS16"},{"why":"Provides the wotan spline detrending (rspline) used to compute the stochastic variability metric.","marker":"Hippke et al. 2019"},{"why":"Supplies the loop model of disk growth and depletion used to interpret CMD loops and the Type 1–3 disk picture.","marker":"de Wit et al. 2006"},{"why":"Earlier claim of super-orbital periodicities and color-magnitude correlations that this paper revisits with longer data.","marker":"Rajoelimanana et al. 2011"},{"why":"Review of Be star decretion disks, providing the physical basis for the red-disk interpretation.","marker":"Rivinius et al. 2013"},{"why":"Sample of SMC B stars used as the comparison for Type 1 faint blue bases.","marker":"Kourniotis et al. 2014"},{"why":"The spin-orbital period relation used to connect spin period to optical variability.","marker":"Corbet 1984"},{"why":"Describes OGLE data reduction, the source of the light curves.","marker":"Udalski et al. 2008"}],"fun_headline_variants":["Five-type taxonomy classifies SMC X-ray binaries from 20 years of OGLE data","Be-star disk growth cycle explains five optical variability types in SMC HMXBs","Two decades of optical data reveal SMC X-ray binaries' five super-orbital behaviors","OGLE survey links SMC X-ray binary flickers to growing Be-star decretion disks","New taxonomy: SMC X-ray binaries show five stages of disk growth and depletion"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The classification depends on the spline detrending used to remove super-orbital signal: if that detrending absorbs real coherent variability, sources would be misclassified across the Type 4/5 cutoffs, weakening the claimed links to spin and orbital periods.","fun_headline_variants_meta":{"raw":{"variants":["Five-type taxonomy classifies SMC X-ray binaries from 20 years of OGLE data","Be-star disk growth cycle explains five optical variability types in SMC HMXBs","Two decades of optical data reveal SMC X-ray binaries' five super-orbital behaviors","OGLE survey links SMC X-ray binary flickers to growing Be-star decretion disks","New taxonomy: SMC X-ray binaries show five stages of disk growth and depletion"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000256,"raw_usage":{"total_tokens":1536,"prompt_tokens":866,"completion_tokens":670,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":482,"completion_tokens_details":{"reasoning_tokens":558}},"tokens_in":482,"tokens_out":670,"duration_ms":7053,"temperature":1.0,"reasoning_tokens":558,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T16:24:44.216349+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recompute the stochastic variability metric for all 111 sources with a range of spline windows and break tolerances (e.g., windows from 100 to 400 days, tolerances from 20 to 200) and count how many sources cross the 0.40 or 0.75 cutoffs; if a substantial fraction move between Types 4 and 5, the taxonomy's boundary is an artifact of detrending. Alternatively, take a Type 1 source during a flare and a Type 2 source during a dip, obtain contemporaneous spectroscopy (H-alpha or V/R line profiles), and check whether the observed disk state matches the predicted growing-versus-depleting phase; a mismatch would refute the disk-growth interpretation.","supporting_citations":[{"cited_title":"F., Charles, P","cited_arxiv_id":null,"evidence_quote":"Earlier claim of super-orbital periodicities and color-magnitude correlations that this paper revisits with longer data."},{"cited_title":"Z., Soszy´nski, I., et al","cited_arxiv_id":null,"evidence_quote":"Sample of SMC B stars used as the comparison for Type 1 faint blue bases."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The spin-orbital period relation used to connect spin period to optical variability."},{"cited_title":"K., Soszynski, I., & Poleski, R","cited_arxiv_id":null,"evidence_quote":"Describes OGLE data reduction, the source of the light curves."}],"review_version":1}