{"id":"fdf93360-d1c5-4bc2-a7aa-968d3f4ed1a5","arxiv_id":"2412.12600","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"A simulation study shows that a traceless tensor, Bayesian fitting approach recovers galaxy-count dipole, quadrupole, and octupole signals on masked skies once source counts exceed about 500,000.","lead":"This paper tests a Bayesian method for measuring the large-scale dipole pattern in galaxy counts, including cases where half the sky is masked. It finds that roughly 500,000 or more sources can recover the pattern, depending on how much sky is covered.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No-mode-coupling claim is established only for signals generated inside the fitted model family; unmodeled angular structure under a mask could still bias the recovered dipole, so the central claim is conditional on model completeness.","rationale":"The reader's weakest assumption identifies the same load-bearing point: the no-mode-coupling result is conditional on real galaxy count maps being close to Poisson realizations of a low-order multipole model. I agree with that assessment. The paper is careful and internally consistent for what it simulates, and the credible-interval and Bayes-factor results are informative, but the central claim in Proposition 4 and the abstract is broader than the simulations support. A single model-misspecification stress test, adding unmodeled angular structure to the injected sky before fitting the lower-order model, would settle whether the mask-induced leakage that affects spherical-harmonic decompositions can also affect this parametric approach in realistic settings. If the test passes, the claim is substantially strengthened; if it fails, the paper's headline should be weakened to \"robust when the model family contains the true angular structure.\" Because the paper is otherwise well-executed and the reader already placed the verdict at CONDITIONAL, I would leave the verdict unchanged rather than accept or reject on the basis of this concern.","tokens_in":27615,"tokens_out":4578,"duration_ms":46033,"concrete_test":"Use the paper's own generator (§3.1.5, eq. 24) to build S3-type catalogues with D=0.007 and Q=0.014 plus an unmodeled ℓ=4 traceless tensor term with amplitude comparable to the ℓ=3 term (say 0.03) and a random orientation. Apply the same g_mask=30 Galactic mask and N≈5×10^6 as in Fig. 10, then fit only model M3 (dipole+quadrupole) over 50 realizations. Record the posterior median of D and the angular separation of the recovered dipole direction from the input dipole. If the median D shifts by more than the 1σ credible interval, or by more than 10% of D=0.007, the no-leakage claim fails under unmodeled angular structure. A second variant should add a declination-dependent systematic map similar to the CatWISE2020 ecliptic bias described in §2 rather than a pure ℓ=4 mode; if both variants are clean, the concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Proposition 4 is an in-sample recovery result. In §3.2.1 the catalogues are Poisson draws with rates λ_i given exactly by (6), (27), and (28); in §3.3.1 the likelihood (30) is evaluated with the same functional forms. When the true signal lies in the model family, the mask only restricts the domain, so the absence of crosstalk between the fitted ℓ=1,2,3 modes is expected. The load-bearing step is generalizing this to real catalogues, where unmodeled angular structure is present. The paper itself notes two sources: multi-component radio sources break the independent-Poisson assumption (§3.2.1), and CatWISE2020 has an ecliptic bias and declination-dependent systematics with higher-order residuals (§2, citing Abghari et al. 2024). If a real map contains an ℓ=4 or higher component, or a smooth declination-dependent systematic, the traceless templates are orthogonal to it only over the full sphere; with f_sky=0.5 the mask can project part of that unmodeled structure onto the dipole template. Then the fitted D and direction could be biased even though the model is correct for every mode it explicitly includes. Proposition 4 therefore does not yet establish that mode coupling is \"not a concern\" for actual galaxy surveys; it establishes that the approach is unbiased for Poisson realizations of the fitted model family on masked skies. This is a model-completeness condition, not an internal inconsistency.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a Bayesian method for fitting source-count multipoles (dipole, quadrupole, octupole) using traceless symmetric tensor templates evaluated directly on pixel counts, rather than spherical harmonic decomposition. The authors generate synthetic Poisson catalogues with injected multipole signals, apply Galactic-plane masks of varying severity and a range of source counts, and use nested sampling to recover amplitudes and directions, with Bayes factors for model comparison. They report that the method recovers injected parameters even at f_sky = 0.5, concluding that mode coupling on masked skies is not a concern for their approach. They also derive source-count thresholds for dipole inference on small and discontinuous footprints and demonstrate sensitivity to prior choice. The paper is structured as a simulation study with propositions summarizing the main findings.","tokens_in":27886,"tokens_out":7955,"duration_ms":69273,"significance":"If the central no-crosstalk claim holds, the method offers a principled way to separate a cosmological dipole from higher-order angular structure in masked galaxy surveys, directly relevant to the current dipole-tension debate. The quantitative source-count thresholds (e.g., N ≈ 500,000 for strong support at D = 0.007) and the KL-divergence-based survey-design framework are useful, concrete deliverables for planning future dipole measurements. The paper is also commendable for explicitly demonstrating prior sensitivity and for using a proper Poisson likelihood, even though the constant terms are omitted. The main limitation is that the key robustness claim is demonstrated only for signals drawn from the same parametric family that is fitted, so the significance for real catalogues is conditional on model completeness; the paper acknowledges this in places but does not test it.","major_comments":[{"comment":"The claim that 'mode coupling on masked skies is not a concern for our approach' is established only for Poisson realizations of the fitted model family. The paper itself notes in the §3.2.1 footnote that independent source positions are broken for multi-component radio sources, and in §2 it describes CatWISE2020 as having an ecliptic bias and higher-order residuals. Because the traceless tensor templates are orthogonal only over the full sphere, unmodeled angular structure (e.g., an ℓ ≥ 4 component or a smooth declination-dependent systematic) can project onto the ℓ = 1 template over the unmasked region and bias the inferred dipole. The current experiments do not include such unmodeled modes, so Proposition 4 overstates the robustness. Please either add simulations with unmodeled angular structure and demonstrate that the dipole is still unbiased, or reformulate Proposition 4 as valid within the assumed model family.","section":"Section 5.1, Proposition 4; Abstract"},{"comment":"The octupole crosstalk test consists of a single UltraNest run with N ≈ 98 million and f_sky = 0.5, and Proposition 4's 'combinations of these underlying multipole signals' is never tested with dipole, quadrupole, and octupole present simultaneously. A single high-source-count run is weak evidence for the general no-crosstalk claim, especially because the proposed application to real catalogues involves much lower source counts. Additional verification with repeated runs at survey-relevant source counts, and with simultaneous dipole+quadrupole+octupole injection, is needed to support the proposition as stated.","section":"Section 4.4"},{"comment":"The direction prior is written as 'θ ∼ cos−1(1−2u) for u ∼ U(0, 0.1)'. As written, u ∈ [0, 0.1] restricts θ to [0, arccos(0.8)] ≈ [0°, 36.9°], which is not uniform over the sphere and would bias every direction inference. The stated goal of uniform coverage over the sphere requires u ∼ U(0, 1). Given the sensible direction posteriors reported later, this is presumably a typographical error, but the formula must be corrected because it is load-bearing for all directional results.","section":"Section 3.3.1, priors"}],"minor_comments":[{"comment":"The caption labels the middle-right column as 'Dipole and quadrupole sample S4'; this should be Sample S3, since S4 is the dipole and octupole sample.","section":"Figure 1 caption"},{"comment":"The text states that samples are generated with values of N up to 10,000,000, but Section 4.4 uses N ≈ 98 million for Sample S4; please clarify that S4 uses a larger source-count regime.","section":"Section 3.2.2"},{"comment":"The fitted function for the KL divergence is typeset ambiguously as 'D_KL(N,r°) = A log10 N + (r°)B − N C + D'; please clarify the intended functional form and report the fitted constants.","section":"Section 4.5.1"},{"comment":"The expression 'θ ∼ cos−1(1−2u)' is missing a closing parenthesis; it should read 'cos−1(1−2u)' with the parenthesis closed after 'u'.","section":"Section 3.3.1"},{"comment":"The data availability statement says data 'will be made available with a reasonable request' but does not mention releasing the analysis code; for a methods-oriented paper, archiving the code would substantially aid reproducibility.","section":"Data Availability"}],"recommendation":"major_revision","confidential_remarks":"The paper is a carefully executed simulation study, and the statistical framework is sound, but the headline claim of robustness to masked skies exceeds the evidence presented. The central no-crosstalk claim rests on in-model-family simulations, and the octupole result is a single high-N run. The fix is achievable within the paper's scope: either add targeted tests with unmodeled angular structure and repeated runs at realistic source counts, or temper the proposition and abstract accordingly. I would not reject on these grounds, but the load-bearing claims need revision before publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is a competent, useful methods paper, but the headline claim that mode coupling is \"not a concern\" outruns the evidence. What the paper actually establishes is that the Bayesian traceless-tensor pipeline recovers injected dipole, quadrupole, and octupole parameters from Poisson realizations of the same parametric model family, even with heavy masks. That is worth having. The source-count thresholds (about 500,000 for dipole support, about 6.6 million for an 80-degree slice at KL=5.5) are concrete numbers people planning surveys will want. The small-footprint and non-contiguous mask tests, including the MALS-like scattered geometry, are practically relevant. Section 4.6 on prior sensitivity is a genuinely good cautionary note, and the paper is honest about the Poisson independence assumption being broken for real radio catalogs.\n\nThe soft spots are real but not fatal. Proposition 4 rests on one high-source-count octupole run (98 million sources, one mask) and one dipole+quadrupole configuration. No test includes angular structure outside the fitted model, such as an l=4 term or a declination-dependent systematic, even though the paper itself cites these as features of CatWISE2020. With f_sky=0.5, the mask can project unmodeled structure onto the dipole template; the claim \"mode coupling is not a concern\" should be conditional on model completeness. The KL-divergence thresholds also depend on an interpolation fit whose constants are not reported, and on the chosen prior widths and injected amplitudes. These are ordinary refinements, not reasons to reject. The math is standard and correctly applied; the citation pattern is fair, with appropriate credit to Pirani/Maxwell and the CMB literature.\n\nWho is this for? Observers planning radio or optical dipole surveys, and anyone revisiting Secrest et al. with higher-order multipoles. It deserves peer review. The referee should press for code and data release, repeated octupole runs, and an injection test with unmodeled structure. I would bring it to reading group, and I would cite it if I were working in this area.","headline":"Useful survey-planning numbers from a sound in-sample Bayesian pipeline, but the no-mode-coupling claim is conditional on model completeness.","tokens_in":28462,"tokens_out":2334,"would_cite":true,"duration_ms":21998,"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":"A Bayesian fit of traceless-symmetric-tensor multipole templates recovers dipole, quadrupole, and octupole parameters from galaxy count maps even when half the sky is masked, with no mode coupling between harmonics.","keywords":["cosmic dipole","dipole tension","kinematic dipole","multipole inference","Bayesian inference","masked sky","galaxy number counts","traceless symmetric tensor"],"falsifier":"Take the paper's pipeline and apply it to a real catalogue with known systematics, such as CatWISE2020: subtract the best-fit monopole+dipole+quadrupole template and compute the angular power spectrum of the residuals. Significant residual power at $\\ell \\geq 4$, or a dipole amplitude that shifts outside its credible interval when an octupole is added to the fit, would show that unmodeled angular structure does leak into the inferred dipole on real skies. A purely synthetic version of the same test is to inject an unmodeled higher-order or axis-mismatched multipole into the Poisson samples and check whether the dipole posterior moves.","tokens_in":27355,"feed_emoji":"🔭","tokens_out":18102,"duration_ms":136485,"temperature":0.7,"pith_summary":"The paper sets out to determine whether the anomalously large cosmic dipole reported in galaxy surveys could be a by-product of analyzing masked skies. It fits dipole, quadrupole, and octupole signals directly as traceless-symmetric-tensor templates in a Bayesian framework, rather than decomposing the observed map into spherical harmonics. Across synthetic Poisson-sampled catalogues, the method recovers the injected amplitudes and directions even with half the sky masked, and the Bayes factors show no cross-talk between the fitted modes. The paper also maps the source counts needed for a credible dipole measurement: roughly half a million sources for an all-sky-like sample, and low tens of millions for small patches covering 12-41% of the celestial sphere. If the result carries over to real catalogues, the elevated dipole amplitudes seen in radio and infrared surveys would not be a mask-induced harmonic-leakage artifact, and the 'dipole tension' would have to be explained by other systematics or by new physics.","feed_headline":"Cosmic multipoles untangled despite half the sky being masked","feed_subtitle":"With no harmonic crosstalk on masked skies, the method tests if the anomalous galaxy dipole is real or an artifact.","key_machinery":"The central object is the traceless symmetric tensor multipole: for order $\\ell$, take the outer product of $\\ell$ unit vectors, symmetrize over all index permutations, and subtract the traces, leaving a tensor whose full contraction with a pixel's unit direction vector gives the signal at that pixel, scaled by an amplitude ($D$, $Q$, or $\\mathcal{O}$). Expected cell counts are the monopole density times (1 + signal), and observed counts are Poisson draws at that rate. Bayesian inference over the amplitudes and unit-vector directions, with nested sampling for the marginal likelihood, yields posterior distributions and log Bayes factors comparing monopole, dipole, dipole+quadrupole, and dipole+octupole models, and the Kullback-Leibler divergence between posterior and prior measures how much information the data carries. Because the templates are fitted directly rather than projected onto a complete-sky harmonic basis, the orthogonality failure that produces mode mixing on masked skies never enters the calculation.","core_discovery":"On its own terms, the paper's central discovery is that multipole inference on the celestial sphere does not need the full-sky orthonormality of spherical harmonics. Each multipole is constructed as a traceless symmetric tensor built from unit vectors — the dipole as $d_j\\hat{p}_j$, the quadrupole as $Q_{jk}\\hat{p}_j\\hat{p}_k$, the octupole as $O_{jkl}\\hat{p}_j\\hat{p}_k\\hat{p}_l$ — and these templates are fitted to Poisson-sampled pixel counts through a Bayesian likelihood, with nested sampling supplying the evidence for model comparison. For samples with a Galactic-plane mask leaving $f_{\\rm sky}=0.5$, the paper reports that the injected dipole, quadrupole, and octupole parameters are recovered accurately, with no mode coupling between $\\ell=1$, $\\ell=2$, and $\\ell=3$; even discontinuous masks resembling the 391 pointings of a real telescope survey pose no added difficulty. It further establishes that the information content of a survey is set mainly by total source count and the angular radius of visible sky, that a dipole of the CMB-implied amplitude $D=0.007$ needs about $N\\approx5\\times10^5$ sources for strong model support, and that small patches placed near the dipole equator can deliver a $\\geq4\\sigma$ exclusion of a $2\\times$ CMB-amplitude dipole.","pith_inferences":["Extension the paper leaves implicit: its no-crosstalk claim is demonstrated only for skies that are exactly Poisson realizations of the fitted low-order models; injecting an unmodeled mode (a higher $\\ell$, or a quadrupole whose axis is not in the fitted set) into the same pipeline is a direct test of whether unmodeled real-world structure — such as CatWISE2020's ecliptic systematics — would reint","The prior-sensitivity analysis implies that part of the scatter across published dipole amplitudes may come from estimator choice: a flat-Cartesian-component prior, which the paper shows biases amplitudes upward, is effectively what least-squares template fits impose (the paper tests such a fit on a pure monopole sample and recovers a spurious $D\\approx0.0035$), pushing low-information samples tow","The near-linear growth of per-source information with slice radius suggests a design rule for future surveys: for a fixed source budget, angular breadth buys more dipole constraining power than depth, so wide shallow coverage should outperform deep narrow pointing for cosmic-dipole science.","Because the iso-information contours (e.g., $D_{\\rm KL}=5.5$) are calibrated against the paper's specific priors and amplitudes, the quoted source-count requirements should be re-derived, not reused, when an analysis adopts different priors or targets different dipole amplitudes."],"forward_implications":["The CatWISE2020 dipole measurement that triggered the tension can be revisited: fitting a simultaneous dipole+quadrupole model separates the kinematic dipole from the strong ecliptic-bias quadrupole instead of letting higher-order power leak into the $\\ell=1$ estimate.","Small-footprint surveys are viable: a 40-degree-radius slice needs about 42 million sources and an 80-degree slice about 6.6 million sources to reach the paper's $D_{\\rm KL}=5.5$ threshold, at which the dipole is constrained to $\\geq4\\sigma$ significance against a $2\\times$ CMB-amplitude dipole.","There is a practical detection floor: near $N\\approx5\\times10^5$ sources, the Bayes factor for a dipole over a monopole crosses 'overwhelming' support at the CMB-implied amplitude, so smaller samples cannot claim either a detection or a null result for the cosmic dipole.","Discontinuous surveys are not handicapped: a scattered mask built from 391 individual pointings of the MALS telescope recovers the dipole with the same KL divergence as a continuous slice of equal angular breadth, and the paper estimates $N\\approx1.2$ million sources in that geometry would suffice for a roughly $3\\sigma$ discrimination of a $2\\times$ CMB dipole from the CMB expectation.","The framework generalizes to arbitrary multipole order, so future wide surveys such as SKA, Euclid, and LSST can be analyzed with a unified low-order multipole model rather than a truncated spherical-harmonic expansion."],"supporting_citations":[{"why":"Derives the expected kinematic dipole amplitude in source counts, $D=(2+x(1+\\alpha))\\beta$, which the paper's synthetic samples are built to recover.","marker":"Ellis & Baldwin (1984)"},{"why":"Fixes the CMB kinematic dipole ($v=369.82\\pm0.11$ km/s at $(l,b)=(264.021^\\circ,48.253^\\circ)$) that sets the fiducial amplitude $D=0.007$.","marker":"Planck Collaboration et al. (2020)"},{"why":"The canonical spherical-harmonic dipole analysis of NVSS whose frequentist estimator the paper argues against.","marker":"Blake & Wall (2002)"},{"why":"Reports the $5\\sigma$ dipole tension in CatWISE2020, the measurement the paper's method is designed to revisit.","marker":"Secrest et al. (2021)"},{"why":"Contends that mode coupling contaminates the CatWISE2020 dipole estimate on masked skies, the rival claim Proposition 4 directly addresses.","marker":"Abghari et al. (2024)"},{"why":"The source of the Bayesian likelihood, priors, and nested-sampling procedure adopted unchanged here.","marker":"Mittal et al. (2024)"},{"why":"Supplies the likelihood form and prior conventions reused in this work, and the estimate that local clustering shifts the NVSS and RACS dipoles by 10-15%.","marker":"Oayda et al. (2024)"},{"why":"Provides the 391-pointing MALS survey geometry used as the discontinuous-mask template and the comparison point for source-count requirements.","marker":"Wagenveld et al. (2024)"}],"fun_headline_variants":["Masked skies no longer break cosmic multipole measurements","500k galaxies suffice to probe the cosmic dipole","Multipole fitting on masked skies without crosstalk","Bayesian multipoles robust to masks, small skies"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that a real galaxy count map is essentially a Poisson realization of a smooth low-order multipole model, with no angular structure beyond the dipole, quadrupole, and octupole being fitted; if a real catalogue carries extra structure (ecliptic-plane systematics, source clustering, multi-component radio sources), the claimed absence of cross-talk between fitted modes is not guaranteed.","fun_headline_variants_meta":{"raw":{"variants":["Masked skies no longer break cosmic multipole measurements","500k galaxies suffice to probe the cosmic dipole","Multipole fitting on masked skies without crosstalk","Bayesian multipoles robust to masks, small skies"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000536,"raw_usage":{"total_tokens":2617,"prompt_tokens":1027,"completion_tokens":1590,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":643,"completion_tokens_details":{"reasoning_tokens":1539}},"tokens_in":643,"tokens_out":1590,"duration_ms":11198,"temperature":1.0,"reasoning_tokens":1539,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T13:55:41.081712+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the paper's pipeline and apply it to a real catalogue with known systematics, such as CatWISE2020: subtract the best-fit monopole+dipole+quadrupole template and compute the angular power spectrum of the residuals. Significant residual power at $\\ell \\geq 4$, or a dipole amplitude that shifts outside its credible interval when an octupole is added to the fit, would show that unmodeled angular structure does leak into the inferred dipole on real skies. A purely synthetic version of the same test is to inject an unmodeled higher-order or axis-mismatched multipole into the Poisson samples and check whether the dipole posterior moves.","supporting_citations":[],"review_version":1}