{"id":"ff1b6d21-47a3-4f8d-bc2e-35b708fb1802","arxiv_id":"2412.04386","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"When only M3 and later red supergiants are compared with pre-explosion images of Type II supernovae, the luminosity distributions agree within uncertainties, so there is no statistically significant missing high-luminosity progenitor problem.","lead":"This paper re-tests the long-standing puzzle that the brightest red supergiants seem to disappear before exploding as supernovae. Using new Milky Way and Local Group star catalogues plus updated uncertainty estimates, it finds the apparent gap can be explained by measurement uncertainties and spectral-type assumptions.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Known luminosity-dependent bias in M3+ I-band bolometric corrections is not propagated into the SN progenitor CLD, so the claimed agreement may be an artifact.","rationale":"The paper's central claim reduces to a null result: uncertainties on single-band derived luminosities are too large to infer a population difference. For that null result to be meaningful, the central values entering the comparison must be unbiased. Section 2.5 demonstrates a luminosity-dependent systematic for M3+ stars in the very band (F814W) and with the very BC (BC_I=0) used for the bulk of the SN progenitor sample. Since the deviation grows with luminosity, correcting it moves the SN CLD's bright end toward fainter values, which is exactly the direction that would make a missing high-luminosity population visible. The paper's response to Fig. 4 is to cite 'significant uncertainty increase,' but inflated error bars do not undo a shift of the mean; they only reduce the significance of an already shifted comparison. Other potential weaknesses are less central. The post-hoc choice of M3+ is supported by independent multi-color detections of very red progenitors (SN2003gd, SN2004et, SN2008bk, SN2009md) and by Davies & Beasor (2018) arguments that RSGs evolve to later types near core collapse. The lack of propagated uncertainties on the RSG CLDs would only widen the comparison bands, making the no-evidence conclusion easier to reach, not harder. Therefore the bolometric-correction bias is the single load-bearing concern. The reader's CONDITIONAL verdict already flags this class of assumption, and the proposed test would either confirm or retire the concern; no verdict change is needed.","tokens_in":17425,"tokens_out":10598,"duration_ms":106167,"concrete_test":"Using the LMC M3+ sample from Section 2.5 (or any sample with both I-band and SED luminosities), fit the median offset Δlog L = log L_I − log L_SED as a function of log L_SED for M3+ stars. Apply this luminosity-dependent offset to the F814W-derived SN progenitor luminosities in Table A.2 for objects with M3+ or late-M spectral types, then recompute the Beasor et al. (2024) and Strotjohann et al. (2024) corrected CLDs. If the corrected SN CLD at log L > 5.0 falls outside the 1σ (or 3σ) band of the metallicity-weighted M3+ RSG CLD, the central claim fails; if it remains inside, the concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 2.5 and Fig. 4 show that for M3+ RSGs with known SED luminosities, the adopted BC_I=0 (Davies & Beasor 2017) yields I-band luminosities that increasingly overestimate the SED luminosity at the bright end ('slightly increased deviation above the 1-to-1 line as luminosity increases'). The SN pre-imaged progenitors in Table A.2 are predominantly F814W photometry with assumed late-M spectral types, and their luminosities are computed with these same BC_I values. A luminosity-dependent overestimate of the true luminosity shifts the bright end of the SN CLD to the right, artificially flattening it and making it appear closer to the steep M3+ RSG CLD in Fig. 7. Correcting for this documented bias would shift the faint end and especially the bright end leftward, likely increasing the deficit of high-luminosity progenitors. The paper treats the scatter seen in Fig. 4 as an additional uncertainty rather than a correction to central values; but a systematic offset is not an uncertainty, and the conclusion 'no evidence of missing high-luminosity pre-imaged progenitors' depends on the biased central values. This is the most load-bearing threat because the entire argument rests on the comparison of luminosities derived through these bolometric corrections.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper re-assesses the red supergiant (RSG) problem by comparing the luminosity distributions of large RSG samples from the Milky Way, M31, M33, LMC, and SMC with the luminosity distributions of Type II supernova pre-imaged progenitors. The authors construct cumulative luminosity distributions (CLDs) for several spectral-type subpopulations (G0+, K0+, M0+, M3+), use metallicity-weighted averages of the Local Group RSGs to mimic the metallicity distribution of the SN progenitors, and compare these to two published bias-corrected progenitor luminosity functions (Davies & Beasor 2020a; Strotjohann et al. 2024). They find that the M3+ subsample matches the steepness of the pre-imaged progenitor distribution, and conclude that, once bolometric-correction and single-band uncertainties are accounted for, there is no statistically significant evidence of missing high-luminosity progenitors; the RSG problem is largely an artifact of these uncertainties.","tokens_in":17634,"tokens_out":3913,"duration_ms":38989,"significance":"If the conclusion holds, the paper would reframe a long-standing controversy: the apparent deficit of luminous Type II progenitors would be attributed to bolometric-correction and single-band photometry uncertainties rather than to a genuine population mismatch. The analysis builds on a valuable new Milky Way RSG catalog (Healy et al. 2024), uses two independent Local Group RSG samples (Ren list and Massey list), and explicitly quantifies the discrepancies between NIR-derived, optical-derived, and SED-derived luminosities. The paper is also commendable for testing spectral-type subpopulations and for weighting the comparison by metallicity. However, the central statistical claim depends on several load-bearing modeling choices that need to be made more rigorous before the null result can be considered established.","major_comments":[{"comment":"The luminosity-dependent deviation above the 1-to-1 line for M3+ RSGs in Fig. 4 is described as a widening dispersion and a 'slightly increased deviation' with luminosity, but it is not applied as a correction to the central values of the SN progenitor luminosities. The pre-imaged progenitors in Table A.2 are mostly F814W photometry with BC_I=0 from Davies & Beasor (2017), the same correction used for the M3+ points in Fig. 4. If this deviation is a systematic bias rather than pure scatter, the bright end of the progenitor CLD is overestimated, which would flatten the apparent distribution and make it appear closer to the steep M3+ RSG CLD in Fig. 7. The conclusion that there is no missing high-luminosity progenitor population therefore depends on treating a potentially systematic offset as an uncertainty. Please correct the central values for the documented trend, or provide a test demonstrating that the deviation is scatter, and propagate the resulting systematic uncertainty into the CLD comparison.","section":"§2.5, Fig. 4, Table A.2"},{"comment":"The M3+ spectral-type cut is selected after observing that M3+ RSGs reproduce the steepness of the pre-imaged SN progenitor sample, and the same pre-imaged sample is then used to test agreement with the M3+ CLDs. This is a circular model-selection step: the agreement in Fig. 7 is partly guaranteed by the way the cut was chosen. The paper does offer independent evolutionary arguments for late spectral types (e.g., Davies & Beasor 2018; multi-color progenitors), but the quantitative significance of the 1-sigma agreement should be assessed without using the same data to both select and validate the M3+ cut. I recommend reporting significance levels for all spectral-type cuts (M0+, M1+, M2+, M3+) or performing an out-of-sample comparison, so the reader can see how much of the agreement is constructed by the cut choice.","section":"§3, §4.1, Fig. 7"},{"comment":"The text in Section 3 states 'For this, we do not consider the SMC' when determining which spectral-type subpopulation best replicates the progenitor luminosity distribution, citing the open issue of missing evolved stars in the SMC. However, the metallicity-weighted CLDs used in Fig. 7 are built using weighting factors derived from all four Local Group galaxies, including the SMC, as described in Section 3.3. This is inconsistent: if the SMC RSG population is considered unreliable for the spectral-type comparison, its inclusion in the weighted CLDs can bias the very comparison used to claim agreement. Please either justify the inclusion of the SMC with quantitative completeness and evolved-star checks, or exclude it from the weighted average and show the effect on the conclusions.","section":"§3, §3.3, Fig. 7"},{"comment":"The RSG CLDs are plotted as deterministic curves without uncertainty bands, whereas the progenitor corrections are shown with confidence intervals. The statement that the M3+ CLDs lie 'within the 1σ bounds' of the progenitor corrections is therefore not a complete statistical comparison: the uncertainty on the RSG side (due to sample completeness, distance, extinction, and NIR/I-band luminosity derivations) is not propagated into the CLDs. Without bootstrap or Monte Carlo uncertainties on the RSG CLDs, or a formal two-sample test that incorporates uncertainties on both distributions, the claim that there is 'no evidence of missing high-luminosity pre-imaged progenitors' is not fully quantified. Please provide uncertainty bands on the RSG CLDs or an equivalent statistical test.","section":"§3.3, §4.2, Fig. 7"}],"minor_comments":[{"comment":"Typographical error: 'agianst' should be 'against' in the caption of Fig. 7.","section":"§3.3"},{"comment":"In Section 4.3, 'then begin a contradiction' appears to be a typo; 'being' seems intended.","section":"§4.3"},{"comment":"The reference list contains two entries for Davies & Beasor with identical journal, volume, and page numbers (MNRAS 474, 2116) but different years (2017 and 2018); one of these is likely mis-cited and should be corrected.","section":"References"},{"comment":"The phrase 'using the ordering method' is unclear; please define the ordering method explicitly or rephrase the caption.","section":"Fig. 6 caption"}],"recommendation":"major_revision","confidential_remarks":"The manuscript contains a valuable compilation of RSG samples and a clear presentation of the systematic uncertainties in single-band luminosity derivation. The central null result is, however, sensitive to the treatment of the luminosity-dependent bias in the M3+ I-band bolometric corrections, the circular selection of the M3+ subsample, and the omission of RSG-side uncertainties from the CLD comparison. These are fixable with additional analysis rather than fatal flaws, but they are load-bearing for the main conclusion. The paper would be strengthened by making the statistical comparison symmetric and by explicitly demonstrating that the result does not depend on the M3+ selection procedure."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know. First, this is a genuinely useful synthesis: the authors combine the new Healy et al. Milky Way RSG catalog with the best Local Group samples, add recent progenitors, and make a metallicity-weighted comparison. That is the right framing, and the claim that only M3 or later RSGs reproduce the steepness of the pre-imaged SN sample is well supported by the figures. Second, the paper's headline conclusion — no missing high-luminosity progenitors — is not yet secure, because the paper itself documents a luminosity-dependent bias in the very bolometric corrections used for the pre-imaged progenitors, then declines to correct for it.\n\nThe novel contribution is the metallicity-weighted cumulative luminosity distribution built from complete Local Group samples. The two independent RSG catalogs (Ren and Massey) agreeing with each other gives some confidence, and adding SN2024ggi plus updating SN2023ixf is useful. The authors are transparent about assumptions, separate detections from upper limits, and show their comparison figures clearly. The compilation itself is solid and will be a reference for the field.\n\nThe soft spot is in Section 2.5 and Fig. 4. For M3+ RSGs, the I-band luminosity with BC_I=0 increasingly overestimates the SED luminosity at the bright end. The pre-imaged SN progenitors in Table A.2 are predominantly F814W, assume late-M spectral types, and use the same BC_I values. That means the bright end of the SN CLD is systematically inflated, which flattens it and makes it look closer to the steep M3+ RSG CLD. The paper treats this as scatter and adds it to the uncertainty budget, but if it is a systematic offset — and the trend in Fig. 4 looks systematic — the central values are wrong. A modest correction shifting bright progenitors down could bring the deficit back. This is load-bearing for the conclusion, and it is the referee's main job to press on it.\n\nTwo smaller issues. The M3+ cut is selected partly because it matches the SN sample, so the resulting agreement is not fully independent; the authors argue for the cut on physical grounds, which is plausible, but it is not a blind test. And the SMC is excluded from the spectral-type comparison but included in the metallicity-weighted average; that needs a sentence of justification. No code or data are released, which is a minor reproducibility shortcoming.\n\nWho is this for: anyone working on massive star evolution, supernova progenitors, or the RSG problem. It is a good state-of-the-field paper and deserves a serious referee. The referee should request a version where the BC_I bias is applied as a correction to the progenitor central values and the comparison redone. My verdict: accept for peer review, with heavy revision expected.","headline":"A useful Local Group reassembly of the RSG problem, but the paper's own Fig. 4 reveals a luminosity-dependent bolometric-correction bias that threatens its central null result.","tokens_in":18201,"tokens_out":3725,"would_cite":true,"duration_ms":51466,"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 claims that when Type II supernova progenitors are compared with a metallicity-weighted sample of M3-or-later red supergiants from the Local Group, the apparent shortage of high-luminosity progenitors disappears because the…","keywords":["red supergiant problem","Type II supernova progenitors","bolometric corrections","cumulative luminosity distribution","spectral type M3","Local Group red supergiants","metallicity weighting","single-band photometry"],"falsifier":"Compare a sample of M3-or-later Local Group red supergiants with both SED-derived luminosities and F814W single-band luminosities computed with the same bolometric correction prescriptions used here; if the F814W-based values show a luminosity-dependent offset exceeding about 0.3 dex at log(L/L_sun) > 5.3 relative to the SED values, then the cumulative luminosity distribution agreement between M3+ red supergiants and supernova progenitors shown in this paper would not survive, and the red supergiant problem would re-emerge.","tokens_in":17190,"feed_emoji":"🌟","tokens_out":9958,"duration_ms":86520,"temperature":0.7,"pith_summary":"The red supergiant problem is the apparent shortage of very luminous red supergiant stars among the progenitors detected in pre-explosion images of Type II supernovae, compared with the bright end of the observed red supergiant luminosity function. This paper reassesses that problem using large, nearly complete red supergiant samples from the Milky Way, M31, M33, the Large Magellanic Cloud, and the Small Magellanic Cloud, with luminosities derived from single-band near-infrared and optical photometry. The authors find that only red supergiants of spectral type M3 or later reproduce the steepness of the Type II pre-imaged progenitor luminosity distribution. When the comparison is made with a metallicity-weighted cumulative luminosity distribution of these M3 or later red supergiants, the pre-imaged progenitors agree within uncertainties, so there is no statistically significant evidence of missing high-luminosity progenitors. A sympathetic reading is that the classic red supergiant problem is largely an artifact of bolometric-correction and single-band luminosity uncertainties rather than a real deficit of massive progenitors.","feed_headline":"Missing bright supernova progenitors vanish with M3+ comparison","feed_subtitle":"Metallicity-weighted Local Group RSG samples show pre-explosion SN images agree within uncertainties","key_machinery":"The central machinery is the cumulative luminosity distribution (CLD) of red supergiant populations, split by spectral-type threshold and constructed separately for each Local Group galaxy, then combined into a metallicity-weighted average that matches the host-galaxy metallicity distribution of the supernova progenitors. Luminosities come from single-band photometry (2MASS Ks for the Milky Way, near-infrared for Local Group samples, and F814W optical for the progenitor comparison) converted with bolometric corrections from Davies and Beasor (2017), Neugent et al. (2020), and Beasor et al. (2024). The comparison is quantified by overlaying the Monte Carlo confidence regions of two corrected progenitor luminosity functions (Beasor et al. 2024 and Strotjohann et al. 2024) on the red supergiant CLDs, with agreement judged by how much of the CLD lies within the 1 sigma, 2 sigma, and 3 sigma bands.","core_discovery":"The central claim is that the red supergiant problem is not a significant population mismatch. Using complete samples of red supergiants in the Local Group plus a new Milky Way catalog, the paper constructs cumulative luminosity distributions for spectral-type cuts G0+, K0+, M0+, and M3+, and weights the galaxy samples by the metallicity distribution of the Type II supernova host galaxies. The pre-imaged progenitor luminosity functions, corrected either with the Beasor et al. (2024) bolometric-correction uncertainties or with the Strotjohann et al. (2024) sensitivity and measurement bias correction, fall within 3 sigma (and mostly within 1 sigma) of the M3+ distributions above log(L/L_sun) approximately 5.0, while the M0+ distributions disagree beyond 3 sigma over much of the range. The paper concludes that only M3-or-later red supergiants represent the immediate pre-supernova state, and that the large uncertainties in single-band F814W derived luminosities (systematic shifts of roughly 0.1 to 0.3 dex with dispersions of roughly 0.2 to 0.5 dex) prevent any meaningful claim of missing high-luminosity progenitors.","pith_inferences":["If these results hold, the persistence of the red supergiant problem in the literature owes as much to the choice of comparison population and the treatment of statistical errors as to any genuine astrophysics; re-analyses of older progenitor samples with an M3+ cut and full covariance propagation could remove the discrepancy without new data.","The paper stops short of claiming the high-luminosity end of the red supergiant luminosity function is metallicity independent, but its Table 1 shows similar maximum luminosities across M31, the Milky Way, M33, and the LMC; that pattern, if confirmed with completeness corrections, would strengthen the empirical case against strong metallicity dependence of the Humphreys-Davidson limit.","A testable extension is to collect pre-explosion multi-color or near-infrared photometry for a larger sample of Type II progenitors; the paper predicts confirmed M0 to M2 progenitors should be rare, so future detections of several early-M progenitors with well-measured luminosities would challenge the M3+ picture."],"forward_implications":["Future pre-imaged supernova progenitor analyses should adopt M3-or-later red supergiants as the reference population, because comparing against all red supergiants or M0+ stars creates an apparent deficit that is not statistically meaningful.","The typical spectral type of a Type II progenitor at core collapse is M3 or later, consistent with stars evolving redward as they approach explosion.","Single-band F814W luminosity estimates are not reliable enough to discriminate population differences; multi-band or near-infrared pre-imaging, or full SED fitting, is needed to test the red supergiant problem cleanly.","Metallicity weighting of comparison samples matters, because unweighted comparisons can produce apparent low-luminosity discrepancies that disappear when host-galaxy metallicities are matched."],"supporting_citations":[{"why":"This reference defined the red supergiant problem and the original claim of missing high-luminosity Type II progenitors that this paper re-examines.","marker":"Smartt et al. (2009)"},{"why":"This reference provides the near-infrared bolometric corrections used to convert single-band progenitor photometry to luminosity.","marker":"Davies & Beasor (2017)"},{"why":"This reference quantifies the systematic underestimate and dispersion of optical F814W single-band luminosities, which drives the main uncertainty budget for progenitor luminosities.","marker":"Beasor et al. (2024)"},{"why":"This reference supplies the survey-sensitivity and measurement-error bias-corrected progenitor luminosity function used as one of the two comparison corrections.","marker":"Strotjohann et al. (2024)"},{"why":"This reference builds the new Milky Way red supergiant catalog that anchors the solar-metallicity cumulative luminosity distribution.","marker":"Healy et al. (2024)"},{"why":"This reference provides the Local Group red supergiant samples with MARCS-model temperatures and bolometric luminosities, forming the Massey list.","marker":"Massey et al. (2023)"},{"why":"This reference supplies the M31 and M33 red supergiant samples that form the Ren list used for the completeness-weighted metallicity comparison.","marker":"Ren et al. (2021)"},{"why":"This reference supplies the Large Magellanic Cloud red supergiant sample used for the low-metallicity end of the metallicity-weighted distribution.","marker":"Yang et al. (2021)"},{"why":"This reference supplies the Small Magellanic Cloud red supergiant sample used for the lowest-metallicity end, though the SMC is excluded from the final spectral-type comparison.","marker":"Yang et al. (2023)"},{"why":"This reference supplies the effective-temperature and bolometric-correction formulas used to derive luminosities for the Ren-list Local Group samples.","marker":"Neugent et al. (2020)"}],"fun_headline_variants":["Red supergiant problem fades when comparing M3+ stars","No missing bright progenitors: uncertainties dominate","M3+ red supergiants explain supernova pre-images","RSG problem resolved: uncertainties mask any deficit","Bright supernova progenitors not missing after all"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The result hinges on the bolometric corrections used to turn single-band pre-explosion magnitudes into luminosities being unbiased, with their quoted uncertainties realistic; if those corrections shift luminosities by a few tenths of a dex as a function of luminosity, the M3+ agreement could disappear and the missing high-luminosity progenitors would reappear.","fun_headline_variants_meta":{"raw":{"variants":["Red supergiant problem fades when comparing M3+ stars","No missing bright progenitors: uncertainties dominate","M3+ red supergiants explain supernova pre-images","RSG problem resolved: uncertainties mask any deficit","Bright supernova progenitors not missing after all"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000603,"raw_usage":{"total_tokens":2832,"prompt_tokens":979,"completion_tokens":1853,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":595,"completion_tokens_details":{"reasoning_tokens":1776}},"tokens_in":595,"tokens_out":1853,"duration_ms":14211,"temperature":1.0,"reasoning_tokens":1776,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T21:24:14.754666+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compare a sample of M3-or-later Local Group red supergiants with both SED-derived luminosities and F814W single-band luminosities computed with the same bolometric correction prescriptions used here; if the F814W-based values show a luminosity-dependent offset exceeding about 0.3 dex at log(L/L_sun) > 5.3 relative to the SED values, then the cumulative luminosity distribution agreement between M3+ red supergiants and supernova progenitors shown in this paper would not survive, and the red supergiant problem would re-emerge.","supporting_citations":[],"review_version":1}