{"id":"038734ce-e62d-48e6-9cce-064748ffdced","arxiv_id":"2608.10388","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"For 26 Fermi-LAT novae, the optical three-magnitude decay time is the most common time window that maximizes gamma-ray detection significance.","lead":"This paper compares Fermi gamma-ray and optical lightcurves for 26 novae and finds that the time a nova takes to fade by three magnitudes in visible light is often the best time window for detecting its gamma-ray emission. The result gives observers a simple optical guide for scheduling gamma-ray analyses and supports the idea that optical and gamma-ray light come from the same shock-heated material.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The t3 preference is not tested against a null hypothesis; t*_gamma is the maximum over a searched bin grid, so the Fig. 2 peak at N~3 may be a selection artifact, not a physical correlation.","rationale":"The reader's weakest assumption is the trials correction for t*_gamma. This is a real issue but not the single most load-bearing one: for V679 Car, TS=61.97 corresponds to ~7.9 sigma, and even with a trials factor of 50 the detection remains >5 sigma, so the >5 sigma claim survives. The population-level t3 claim, however, is not protected by high TS. Weak sources have flat TS curves, making t*_gamma nearly arbitrary, and the paper provides no null test to show that the histogram peak at N≈3 differs from what would arise from the bin grid alone. The paper's own statement in §3 that weak sources support 'a range of comparably significant t_gamma values' underscores this. The t3 preference is the central claim of the abstract; if it is an artifact, the paper's main contribution reduces to a useful tool and a catalog. The proposed permutation test would settle this directly. I therefore agree with the CONDITIONAL verdict but identify the missing null test as the load-bearing point, making my agreement with the reader only partial.","tokens_in":12981,"tokens_out":9874,"duration_ms":111629,"concrete_test":"Shuffle the t3 values among the 26 sources (or, more conservatively, draw t*_gamma uniformly from the same logarithmic grid) and recompute the Fig. 2 histogram and the RMS distance to the 1:1 line in Fig. 3. Repeat 10^4 times; if the observed peak at N≈3 and the RMS distance are not in the extreme tail (p<0.05) of the permutation/null distribution, the claim that t3 is the favored bin is not supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim, that t3 is the favored gamma-ray integration bin (Abstract; §4.2; Fig. 2), rests entirely on t*_gamma, defined in §2.4 as the bin that maximizes TS over a logarithmically sampled grid from 0.5 to 1500 days. The authors acknowledge in §3 that weakly detected novae have broad TS distributions, yet they treat the single maximizing bin as a precise measurement. No null hypothesis is constructed for Fig. 2: if t*_gamma were unrelated to optical decay, the histogram of magnitude drop would be set by the grid spacing and the sources' lightcurve shapes, and it could peak near N=3 by chance. The RMS comparison in §4.2 is also unquantified. Thus the apparent t3 preference could be a selection artifact. The same selection mechanism affects the population-level t*_gamma values, though for an individual high-TS source like V679 Car (TS=61.97, Table 1) the trials penalty is likely mild.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper analyzes all 26 Fermi-LAT-detected novae on Koji Mukai's list for the period August 2008 through June 2024. For each nova, the authors perform a binned likelihood analysis over a logarithmically spaced grid of integration times t_gamma and define t*_gamma as the bin that maximizes the Test Statistic. They compare t*_gamma with optical decay times t_N measured from AAVSO V-band light curves using a new open-source machine-learning tool (nova-times), and interpret the resulting distribution as showing that t_3 is the favored gamma-ray integration window. They also report V679 Car as a >5 sigma gamma-ray nova, propose associations for three coincident 4FGL sources, and perform cross-correlation analyses of optical and gamma-ray light curves for sources with TS>30.","tokens_in":13149,"tokens_out":4875,"duration_ms":58421,"significance":"If the claimed t_3 correspondence is real, the paper would provide a physically motivated optical prior for choosing gamma-ray integration windows in future Fermi-LAT nova analyses, directly tying gamma-ray detectability to the shock-reprocessing scenario. The confirmation of V679 Car as a >5 sigma source would add a new gamma-ray nova to the sample. The paper also ships the open-source nova-times tool and a full table of t_N measurements, which are useful community resources. The main caveat is that the central statistical claims are currently supported mainly by visual trends and unquantified comparisons; the significance of the result depends on whether the t_3 preference survives a proper null-hypothesis treatment and trials correction.","major_comments":[{"comment":"The quantity t*_gamma is defined as the bin that maximizes TS over a logarithmically sampled grid from 0.5 to 1500 days, yet the quoted significances, including the V679 Car claim of >5 sigma (Table 1, TS=61.97), are derived from sqrt(TS) or sqrt(Delta TS) with no correction for this maximization. Because Section 3 itself notes that weakly detected novae have broad TS distributions, the single maximizing bin is not a robust measurement; please provide either a trials-corrected significance, a null-hypothesis simulation of the scan, or an explicit argument that the number of effectively independent bins is small enough that the multiplicity penalty is negligible.","section":"§2.4, §3"},{"comment":"The central claim that t_3 is the favored gamma-ray integration window rests on a histogram that peaks near N approximately 3, but no null model, p-value, or uncertainty on the bin counts is given, and t_3 was selected only after comparing t_2, t_3, and t_4. The RMS comparisons in §4.2 are quoted without errors or a statistical test. Please provide a quantitative test, such as a permutation of the t*_gamma values against the optical t_N values or a bootstrap distribution of the histogram peak, and report correlation coefficients with uncertainties for Figs. 3 and 4.","section":"§4.2, Fig. 2"},{"comment":"The t_N measurements are reported without uncertainties even though the text states that variance increases for sparse AAVSO light curves and the GBM fit has a tunable depth. Since t*_gamma is discretized on a logarithmic grid, the Fig. 3 comparison of t*_gamma with t_3 needs error bars on both axes before claims such as 'many sources lie on the 1:1 line' and the t_2/t_3/t_4 RMS ranking can be evaluated.","section":"Table 2, §4.1"}],"minor_comments":[{"comment":"The column header 'Positional Optimal t*_gamma Offset (deg) ROI TS a (days)' is ambiguous; the units for t*_gamma should be labeled clearly in the header or caption.","section":"Table 1"},{"comment":"The color-bar label 'sqrt(Delta TS) ~ sigma' is informal; since Delta TS is not guaranteed to follow the Wilks distribution for an on-peak scan, please either define the mapping precisely or relabel the color scale.","section":"Fig. 1 caption"},{"comment":"V906 Car and V959 Mon are excluded from Fig. 2 and the §4.3 analysis, but this exclusion is only explained in §2.4; please state these exclusions explicitly in the relevant figure captions and main text.","section":"§2.4, Fig. 2"},{"comment":"The nova-times tool assumes that the observations sample the peak brightness, while the text also notes that the peak may be missed; please state how this assumption affects the measured t_N values and whether any sources are particularly affected.","section":"§4.1"},{"comment":"The sentence referring to 'the undetected sources' in the context of Fig. 3 should be reworded, since the sources under discussion are weakly detected rather than undetected.","section":"§4.2"},{"comment":"The Li et al. (2017) reference appears twice in the reference list and should be deduplicated.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within scope for the journal and the central idea is interesting, but the statistical support for the t_3 preference is currently too weak to justify the abstract claim as stated. The required changes are feasible: a trials correction or null simulation for the bin scan, error bars on t_N values, and a quantitative comparison of t_2, t_3, and t_4. I do not see grounds for rejection, but the paper should not be accepted until these points are addressed."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe paper is worth a look, but the headline claim comes with a large asterisk. The systematic re-analysis of all 26 Fermi-LAT novae with a logarithmic grid of time bins is genuinely useful: it shows that t*_gamma varies widely across the population, contrary to the common 15-day window used in earlier work. The open-source nova-times tool for measuring t_N from AAVSO lightcurves is a practical contribution, and the off-peak analysis of coincident 4FGL sources is thoughtful—the three proposed associations (V1324 Sco, V5855 Sgr, V549 Vel) deserve follow-up. The V679 Car detection at TS≈62 is likely real even after a modest trials penalty.\n\nThe soft spot is the central claim that t3 is the favored bin. t*_gamma is defined as the maximum TS over a grid from 0.5 to 1500 days; that maximum is a selection effect waiting to happen, and for weakly detected novae the TS curve is broad anyway. The paper presents a histogram peaking near N=3 and a 1:1 plot, but gives no correlation coefficient, no null hypothesis, and no trials correction. The choice of t3 after comparing t2/t3/t4 adds another layer of post hoc selection. This is the load-bearing claim, and as written it is not quantitatively supported.\n\nThere is also a statistical misstep in interpreting sqrt(Delta TS) as sigma. Wilks' theorem applies to nested models fit to the same data; different time bins use different photon sets, so the TS difference is not a chi-square variable. The colored regions in Figures 1 and 3 are descriptive, not significance contours.\n\nThe t_N measurements come without error bars, which is a problem given irregular sampling and the acknowledged risk of missing peak brightness. The authors note these issues but don't propagate them into the analysis.\n\nThat said, the paper is honest and the flaws are fixable. For novae/transient researchers and Fermi-LAT analysts, the systematic dataset and the tool are worth having in the literature. The statistical issues can be addressed with a null simulation and a trials correction. I'd send it to a referee with those requests, not desk reject it.","headline":"A useful systematic re-analysis whose headline t3 claim needs a null test and trials correction before it is established.","tokens_in":13727,"tokens_out":6017,"would_cite":true,"duration_ms":65020,"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":"A nova's optical t3 decay time is the integration window that maximizes its Fermi-LAT gamma-ray significance, and V679 Car is a >5 sigma gamma-ray nova.","keywords":["gamma-ray novae","Fermi-LAT","optical decay time","t3 (three-magnitude decay)","shock-powered emission","time window optimization","V679 Car","lightcurve cross-correlation"],"falsifier":"Repeat the bin scan on null data — for example, on off-peak time intervals or scrambled photon arrival times for each source — and record the distribution of the maximum Test Statistic over the many bins tested. If this null distribution yields 'optimal' bins as significant as those seen for the real eruptions, or if applying that trials factor drops V679 Car below 5 sigma, the central claim fails.","tokens_in":12758,"feed_emoji":"🌠","tokens_out":10225,"duration_ms":92069,"temperature":0.7,"pith_summary":"Using all 26 Fermi-LAT-detected novae known through June 2024, this work asks whether a nova's optical lightcurve can tell us which time window will best reveal its gamma-ray emission. Scanning integration bins from 0.5 to 1500 days, the authors find that the bin that maximizes gamma-ray detection significance usually corresponds to the optical t3 decay time — the time for the V-band brightness to drop three magnitudes — although the population shows considerable spread. This matters because a cheap optical measurement could set the gamma-ray analysis window for future and archival novae, avoiding blind scans. The study also promotes V679 Car from a marginal detection to a >5 sigma gamma-ray nova, and its cross-correlation analysis finds near-zero optical/gamma-ray lag for most systems, consistent with both bands arising from the same shock-powered emission.","feed_headline":"Nova fade by 3 magnitudes marks best gamma-ray detection window","feed_subtitle":"Across 26 Fermi-LAT novae, the three-magnitude optical decay time matches the gamma-ray bin with top significance.","key_machinery":"The central object is t*_gamma, the gamma-ray integration window that maximizes the likelihood Test Statistic TS = 2 ln(L/L0) in a logarithmic scan of bin sizes from 0.5 to 1500 days. It is paired with t_N, the optical decay time measured by the paper's machine-learning tool (nova-times) — a gradient-boosted fit to irregular V-band lightcurves that returns the time for the brightness to fall N magnitudes from maximum. The argument is carried by comparing t*_gamma with the optical magnitude drop measured over that same window (peaking near N ≈ 3) and by plotting t_gamma against t3 across the population, with the 1:1 line as the reference for agreement.","core_discovery":"The paper's central claim is that a nova's optical t3 — the time to decay three magnitudes in the V band — is the favored analysis bin for Fermi-LAT gamma-ray detection. For each of 26 novae, the authors run a binned maximum-likelihood analysis in logarithmically spaced time windows from 0.5 to 1500 days and define t*_gamma as the window with the largest Test Statistic (TS = 2 ln(L/L0)). They then measure optical decay times t_N with a gradient-boosted machine-learning tool applied to V-band lightcurves, and compare t*_gamma with the optical magnitude drop over the same window. The resulting distribution peaks at roughly three magnitudes of decay, with first and third quartiles near two and four, and a plot of t_gamma against t3 shows many sources near the 1:1 line. Interpreting sqrt(ΔTS) between windows as sigma, the paper reports V679 Car as a >5 sigma gamma-ray nova, and cross-correlation of optical and gamma-ray lightcurves for 18 sources finds zero-lag alignment for most, with ~1-day lags for RS Oph, V1723 Sco, V357 Mus, and V5856 Sgr.","pith_inferences":["If the t3 correspondence survives a trials correction for the number of scanned bins, it gives observers a cheap optical prior for gamma-ray timing that could be applied to archival Fermi-LAT data to search for fainter or previously missed novae.","The large spread around t3 (quartiles from ~2 to ~4 magnitudes) suggests that a single universal gamma-ray window is not physical; population studies should weight each system's window by its own optical decay rather than adopt a common bin.","The nova-times tool's gradient-boosted decay-time measurement from sparse lightcurves could be carried over to other transient classes, where decay timescales similarly select observing strategies.","A direct test: for a nova discovered purely optically with no gamma-ray trigger, predict the optimal gamma-ray window from its t3 and check that this bin alone yields a significant detection in archival LAT data."],"forward_implications":["Future Fermi-LAT analyses of newly discovered novae can set their gamma-ray integration window from optical monitoring alone, choosing the optical t3 decay time and skipping blind bin scans.","The three 4FGL sources coincident with V1324 Sco, V5855 Sgr, and V549 Vel should be removed from the ROI background model during eruption windows, improving the measured flux of these novae.","V679 Car joins the confirmed gamma-ray nova population as a >5 sigma source, strengthening the case that ordinary classical novae are regularly GeV emitters.","The near-zero optical/gamma-ray lags measured for most sources, with ~1-day lags for RS Oph, V1723 Sco, V357 Mus, and V5856 Sgr, support the shock-reprocessing scenario in which optical emission is absorbed and re-radiated shock power."],"supporting_citations":[{"why":"supplies the binned maximum-likelihood method and Test Statistic definition used for all significance measurements.","marker":"J. R. Mattox et al. 1996"},{"why":"provides the asymptotic result that the paper uses to interpret sqrt(ΔTS) between time bins as a sigma scale.","marker":"S. S. Wilks 1938"},{"why":"previous population analysis that reported a common ~15-day gamma-ray window; the t*_gamma values here are compared with it.","marker":"A. Franckowiak et al. 2018"},{"why":"established novae as a GeV source class and is the basis for the PLEC spectral model adopted in the fits.","marker":"M. Ackermann et al. 2014"},{"why":"shock-reprocessing model in which optical luminosity is absorbed and re-radiated shock power, motivating the expected optical/gamma-ray correspondence.","marker":"B. D. Metzger et al. 2015"},{"why":"earlier correlation study of optical and gamma-ray nova lightcurves that this work extends with cross-correlation.","marker":"E. Aydi et al. 2020"}],"fun_headline_variants":["Nova’s three-magnitude fade sets best gamma-ray detection window","Three-magnitude fade times match optimal gamma-ray bins for novae","Fermi-LAT novae: t3 fade time predicts gamma-ray detection bin","V679 Car confirmed as gamma-ray nova via cross-correlation","Nova fade by 3 magnitudes ties optical to gamma-ray emission"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The analysis treats the time bin that maximizes the test statistic in a wide scan of bin sizes as a robust measurement, without correcting for the number of bins tested, so the claimed >5 sigma significance of V679 Car and the t3 correspondence could be inflated by a selection effect.","fun_headline_variants_meta":{"raw":{"variants":["Nova’s three-magnitude fade sets best gamma-ray detection window","Three-magnitude fade times match optimal gamma-ray bins for novae","Fermi-LAT novae: t3 fade time predicts gamma-ray detection bin","V679 Car confirmed as gamma-ray nova via cross-correlation","Nova fade by 3 magnitudes ties optical to gamma-ray emission"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00043,"raw_usage":{"total_tokens":2239,"prompt_tokens":1031,"completion_tokens":1208,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":647,"completion_tokens_details":{"reasoning_tokens":1111}},"tokens_in":647,"tokens_out":1208,"duration_ms":11963,"temperature":1.0,"reasoning_tokens":1111,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T00:43:41.979662+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Repeat the bin scan on null data — for example, on off-peak time intervals or scrambled photon arrival times for each source — and record the distribution of the maximum Test Statistic over the many bins tested. If this null distribution yields 'optimal' bins as significant as those seen for the real eruptions, or if applying that trials factor drops V679 Car below 5 sigma, the central claim fails.","supporting_citations":[],"review_version":1}