{"id":"6527a671-45a3-419c-81be-9c6e2d1c110d","arxiv_id":"2506.23029","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"high","formal_verification":"none","parameter_count":6,"one_line_summary":"A joint analysis of ACT DR6, DESI DR2 BAO and DESY5 supernovae finds a 3.4σ preference for dark matter with an evolving equation of state, but the significance is sensitive to model choice and data selection.","lead":"Using ACT's cosmic microwave background, DESI's galaxy surveys, and DESY5 supernovae, this paper reports a 3.4 sigma hint that dark matter's pressure changes over cosmic time. The claim is that dark matter behaves less like cold matter and more like a fluid whose equation of state scales with the expansion factor.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 3.4 sigma DDM detection is undercut by the paper's own SDDM result: the same ACT+DESI+DESY5 data, restricted to the claimed proportional EoS, give omega_dm = 0.006 +/- 0.015, inconsistent with the DDM best-fit lying on that line.","rationale":"The paper's central claim is a 3.4 sigma detection of a time-varying dark matter EoS from ACT+DESI+DESY5, plus an independent confirmation of the proportional relation omega_dma = -omega_dm0. For either claim to hold, the ADS likelihood must actually prefer a nonzero DM EoS. The strongest check is the paper's own SDDM model, which imposes the claimed relation and has a single parameter. Table I reports omega_dm = 0.006 +/- 0.015 for ADS, i.e., no preference. This is not a minor discrepancy: the DDM best fit from the same data is omega_dm0 = -0.084, omega_dma = 0.088, which lies almost exactly on the proportional line with omega_dm ~ -0.084. A well-behaved posterior restricted to that line would therefore prefer omega_dm ~ -0.08, not 0.006. The disagreement indicates either a strongly non-Gaussian or degenerate posterior that makes the reported DDM means and errors misleading, or an inconsistency between the DDM and SDDM implementations. Either way, the 3.4 sigma headline cannot be taken at face value. This conclusion is reinforced by the arithmetic (0.088/0.030 = 2.9, not 3.4), by the drop to 2.98 sigma when Planck is added, and by the drop to 1.5 sigma when DESY1 is included; the latter is excluded rather than explained. The paper's own warning that CMB+DESI constraints are biased through r_d and H_0 identifies a plausible mechanism for a spurious positive omega_dma. I therefore agree with the reader's REJECT verdict: the central claim is internally inconsistent. My emphasis differs slightly from the reader's formal weakest_assumption: I see the SDDM null result as the most direct falsifier, while the r_d/H_0 bias is a supporting systematic concern. The LSS forecasts are illustrative and do not rescue the detection. The issue is in the reported numbers and model comparison, not in the authors' intent.","tokens_in":20010,"tokens_out":10530,"duration_ms":111937,"concrete_test":"Re-run the ACT+DESI+DESY5 analysis in both the DDM and SDDM models, then project the DDM chains onto the line omega_dma = -omega_dm0 and compare the resulting one-dimensional posterior for omega_dm with the SDDM posterior. If the projected DDM posterior peaks near omega_dm ~ -0.08 while the SDDM posterior gives 0.006 +/- 0.015, the 3.4 sigma claim is a projection/volume artifact. As a secondary check, recompute the significance of omega_dma = 0.088 +/- 0.030; if it is 2.9 sigma rather than 3.4 sigma, the headline significance is overstated.","verdict_should_be":"UNCHANGED","load_bearing_attack":"For the central claim to hold, the ACT+DESI+DESY5 likelihood must genuinely prefer a nonzero DM equation of state along the direction reported for DDM. The paper's own single-parameter SDDM model is exactly this model with the claimed linear relation omega_dma = -omega_dm0 imposed. Table I reports omega_dm = 0.006 +/- 0.015 for the same ADS data, i.e., no preference for a time-varying DM EoS. If the two-parameter DDM posterior truly had a mode on the proportional line at omega_dm ~ -0.08, the restricted SDDM posterior would peak near that value; instead it is consistent with zero. This internal inconsistency means the quoted 3.40 sigma from omega_dma = 0.088 +/- 0.030 (which is also arithmetically 2.9 sigma, not 3.4 sigma) is not a stable detection but an artifact of the added parameter omega_dm0 and of how the two-parameter posterior is marginalized. The paper's own discussion that CMB+DESI combinations are biased through r_d and H_0 provides a plausible mechanism for a spurious positive omega_dma, but the SDDM null is sufficient by itself to contradict the claim that ACT independently confirms the proportional EoS. The exclusion of DESY1 after it lowers the significance to 1.5 sigma, and the larger drop when Planck is added, further show that the headline significance depends on dataset choice rather than on a robust signal.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces a CPL-like parameterization for the dark matter equation of state, ωdm(a)=ωdm0+ωdma(1−a), and constrains it with ACT DR6 CMB anisotropy and lensing, DESI DR2 BAO, and DESY5 supernova data. It claims a 3.4σ detection of nonzero ωdma and states that ACT independently confirms the linear relation ωdma=−ωdm0, so that the DM EoS is proportional to the scale factor. The paper also constrains a one-parameter 'proportional' DDM model and a model with dynamical DE, and it presents forecasts for LSS observables. The central claim is that dark matter is not cold and pressureless over cosmic time.","tokens_in":20348,"tokens_out":8614,"duration_ms":87927,"significance":"If the 3.4σ detection were real, the paper would provide an important observational challenge to the standard CDM paradigm and would motivate new particle and modified-gravity scenarios. The work expands the parameter space explored in the authors' earlier study and uses an independent ACT DR6 likelihood, which is a useful cross-check. However, the evidence as presented is not robust: the quoted 1σ error gives only 2.9σ, the one-parameter SDDM version of the same model shows no preference for evolving DM from the same data (ωdm=0.006±0.015 in Table I), and adding DESY1 reduces the significance to 1.5σ. These internal checks, together with the authors' own warning about rd/H0 bias in CMB+DESI combinations, mean the central claim is not sustained.","major_comments":[{"comment":"The 3.4σ number in the abstract is not supported by the quoted errors. Table I lists ωdma=0.088±0.030 for ADS, which is 2.9σ from zero, and ωdm0=−0.084±0.029, also 2.9σ. If the 3.4σ significance is computed from a joint marginalized statistic or a profile likelihood, the paper should state the statistic and its definition; as written, the reader cannot reproduce the headline significance.","section":"DDM evidence induced by ACT; Table I"},{"comment":"The SDDM result in Table I is internally inconsistent with the claim that ACT confirms the proportional relation. SDDM is the same model restricted to ωdma=−ωdm0, so its single parameter should track the DDM best fit along that line. Yet for ADS, SDDM gives ωdm=0.006±0.015, i.e., no preference for a non-zero DM EoS, whereas the DDM best fit lies at ωdm0=−0.084 and ωdma=0.088. This means the apparent DDM signal is not a robust preference along the claimed proportionality direction; it emerges only when the second EoS parameter is added. The paper does not reconcile this contradiction.","section":"Table I and SDDM discussion"},{"comment":"The dataset-dependence of the result is not described accurately. Adding DESY1 changes ωdma from 0.088±0.030 to 0.039±0.026, a 1.5σ effect, and adding Planck to ADS lowers it to 0.063±0.025, below the 3σ threshold. The paper then excludes DESY1 on the grounds that it introduces inconsistencies. Because the exclusion is motivated by the effect it has on the result, the quoted 3.4σ should be treated as a selection-dependent finding rather than a robust detection.","section":"DDM versus LSS"},{"comment":"The paper itself states that the combination of CMB and DESI gives biased constraints because the CMB-derived rd induces a larger H0, and it invokes this bias to explain the anomalous positive ωdm in SDDM for the AD data. Since the same mechanism could plausibly bias ωdma upward, the paper should verify with a concrete test, such as marginalizing over rd or comparing with a free sound-horizon calibration, that the ADS preference for ωdma>0 is not a propagation of this known bias. Without such a test, the unbiasedness assumption behind the headline claim is unsupported.","section":"Discussions and conclusions"},{"comment":"The 'confirmation' of the linear relation ωdma=−ωdm0 is a post-hoc reading of the posterior degeneracy. The one-dimensional constraints in Fig. 2 and the SDDM row in Table I show that the data do not independently prefer this line. The paper should report a model comparison, such as Δχ² or Bayesian evidence, for DDM, SDDM, and ΛCDM before treating the relation as a confirmed physical property.","section":"DDM evidence induced by ACT; Fig. 1"}],"minor_comments":[{"comment":"The sentence 'we propose a conjecture that DM obeys the directly proportional DM EoS ωdma = ωdma emerged on cosmic scales' is not grammatical and should read ωdm(a)=ωdm a; as printed, it is meaningless.","section":"Discussions and conclusions"},{"comment":"The authors refer to their previous DDM work as 'Ref. [1]', but in the bibliography Ref. [1] is Penzias and Wilson (1965); the previous DDM paper is Ref. [69]. This makes it difficult to follow the line of reasoning.","section":"Introduction and Supplementary A"},{"comment":"No statement of code or data availability is given; since the analysis uses a custom modification of CAMB (DDMCAMB), a public release or a clear statement of availability is needed for reproducibility.","section":"Data and methodology"},{"comment":"The paper does not define how the 2σ upper limits in Table I are computed, whether marginal or profile; adding a footnote would remove ambiguity.","section":"Table I"}],"recommendation":"reject","confidential_remarks":"The internal contradiction between the DDM and SDDM rows of Table I is, in my view, sufficient for rejection: the same data, restricted to the claimed linear relation, show no preference for a time-varying DM EoS. Should the authors resubmit with the central claim narrowed to a 2.9σ hint and with a full model-comparison statistic, I would be willing to reconsider."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a competent ACT DR6 application of the DDM framework from Wang's earlier paper, but the headline 3.4σ detection is not supported by the paper's own numbers. It deserves a serious referee, but it should not be cited as evidence for dynamical dark matter in its current form.\n\nWhat is genuinely new: constraints on the two-parameter DM EoS from ACT DR6 plus DESI DR2 BAO and DESY5 SNe, a modified-CAMB implementation, and a broad LSS phenomenology supplement. The ACT-only posterior does point in the same direction as the earlier Planck-based result, and the paper is transparent about running SDDM, DDME, and constant-ω_dm0 variants. The methods are standard, the data are public, and the supplementary forecast curves are useful for people working on generalized dark matter.\n\nThe soft spots are real and load-bearing. First, the central inconsistency: the two-parameter DDM fit gives ω_dma = 0.088 ± 0.030, which is arithmetic ~2.9σ, not 3.4σ. More importantly, the same ACT+DESI+DESY5 data analyzed with the single-parameter proportional model—exactly the model the paper claims ACT confirms—gives ω_dm = 0.006 ± 0.015, consistent with zero. If the two-parameter posterior really had a mode on the proportional line at ω_dm ≈ −0.08, the restricted SDDM posterior would peak near there. It does not. That is an internal contradiction, and it undercuts the claim that ACT independently confirms the linear relation. Second, the significance is dataset-dependent: adding Planck drops it to about 3σ, adding DESY1 drops it to 1.5σ, and DESY1 is then excluded because of “inconsistencies.” That is not the behavior of a robust signal. Third, “independent of Planck” is overstated: the ACT-only runs use the Sroll2 Planck low-ℓ prior, which should be flagged more prominently. Fourth, there is a sign typo in the main text: APDS ω_dma is quoted as −0.061 ± 0.025 while Table I lists +0.063 ± 0.025. Finally, the LSS section is an illustrative parameter scan, not a likelihood test; the paper does say this, but the abstract's “thoroughly studied” oversells it.\n\nWho this is for: cosmologists working on generalized dark matter or dark-sector parametrizations will find the constraints and LSS plots useful as a reference, not as a detection. A serious referee should demand that the authors reconcile the DDM best-fit with the SDDM null, correct the significance arithmetic and the sign typo, and either remove the 3.4σ claim or explain why the restricted model is not the relevant test. I would not cite it as evidence, but I would send it to review because the question is important and the analysis is mostly reproducible.","headline":"Careful ACT DR6 constraint work on a CPL-like dark-matter EoS, but the headline 3.4σ detection is contradicted by the paper's own single-parameter SDDM null and by dataset-dependent significance shifts.","tokens_in":20985,"tokens_out":2373,"would_cite":false,"duration_ms":25949,"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":"Combining ACT DR6, DESI DR2, and DESY5 data, the paper reports 3.4σ evidence that dark matter has an evolving equation of state, with pressure proportional to the cosmic scale factor.","keywords":["dynamical dark matter","equation of state","ACT DR6","DESI DR2","DESY5","cosmic microwave background","baryon acoustic oscillations","large-scale structure"],"falsifier":"Re-run the ACT+DESI+DESY5 fit with the comoving sound horizon treated as a free parameter or calibrated by a local distance ladder; if ωdma drops to within about 1σ of zero, the 3.4σ is a calibration artefact. A second check is to compute the model-selection evidence for DDM against ΛCDM, since the single-parameter proportional model already returns ωdm=0.006±0.015.","tokens_in":19688,"feed_emoji":"🌌","tokens_out":8890,"duration_ms":89730,"temperature":0.7,"pith_summary":"The paper tries to establish that dark matter is not the pressureless, time-independent fluid assumed by standard cosmology, but has an equation of state that changes with the expansion of the universe. Using ACT DR6 cosmic microwave background data together with DESI DR2 baryon acoustic oscillations and DESY5 supernovae, it finds a 3.4σ preference for a nonzero evolution parameter ωdma=0.088±0.030. It also claims that ACT alone, without Planck, confirms the linear relation ωdma=−ωdm0, so that the dark-matter equation of state is directly proportional to the scale factor. If this is right, the dark sector contains an evolving dark-matter component, and structure growth, cosmic acceleration, and the fate of the universe would all need to be reconsidered.","feed_headline":"Dark matter's pressure may evolve with cosmic time: 3.4σ hint","feed_subtitle":"Combining ACT, DESI, and supernova data, a 3.4-sigma sign that dark matter is not a cold, pressureless fluid.","key_machinery":"The central object is the CPL-style dark-matter equation of state ωdm(a)=ωdm0+ωdma(1−a), inserted into the Friedmann equations through ΩDM(a)=Ωdm $a^{{-3(1+ωdm0+ωdma)}}$ $e^{{3ωdma(a−1)}}$. The paper develops DDMCAMB, a modified version of the CAMB Boltzmann code, to compute background and perturbation evolution for arbitrary dark-matter equation-of-state models, and fits three models—full DDM, single-parameter proportional SDDM, and DDME with a CPL dark-energy component—to ACT DR6, DESI DR2, DESY5, and Planck data with MCMC. The load-bearing identity is the linear relation ωdma=−ωdm0, which reduces the two free parameters to one and makes the dark-matter pressure track the scale factor directly.","core_discovery":"The central discovery is that a dark-matter equation of state of the form ωdm(a)=ωdm0+ωdma(1−a) is preferred over pressureless cold dark matter at the 3.4σ level when ACT DR6 CMB data are combined with DESI DR2 BAO and DESY5 supernovae. The best-fit values ωdma=0.088±0.030 and ωdm0=−0.084±0.029 are consistent with the linear relation ωdma=−ωdm0, implying that the dark-matter equation of state is proportional to the scale factor a. ACT data alone independently support that same relation, matching the earlier Planck-based result. The paper also finds that adding Planck lowers the significance to 2.98σ and adding DESY1 large-scale-structure data lowers it to 1.50σ, while the proportional relation remains. The broader claim is that the dark sector is likely dynamical dark matter plus dynamical dark energy rather than cold dark matter plus a cosmological constant.","pith_inferences":["Beyond the paper: a 3.4σ parameter significance in a two-parameter extension is not automatically model-selection evidence; the paper's own single-parameter SDDM fit shows no preference (ωdm=0.006±0.015), so a Bayesian comparison of DDM against ΛCDM could be considerably weaker.","Beyond the paper: the rd-induced Hubble-constant bias the authors flag implies a clean test—re-fit with rd calibrated from the local distance ladder or left free—and if ωdma then returns to zero, the headline signal is a distance-scale artefact.","Beyond the paper: the negative today's pressure implied by ωdm0≈−0.08 predicts suppressed small-scale structure growth (lower S8) relative to ΛCDM, a signature that DESI and next-generation lensing surveys should be able to check within a few years."],"forward_implications":["Dark matter would not be cold and pressureless: its equation of state evolves with the scale factor, with negative pressure today if ωdm0<0.","Because the relation ωdma=−ωdm0 is confirmed by ACT independently of Planck, the dark-matter equation of state can be described by a single parameter proportional to a.","The same data combinations show roughly 2σ evidence for coexisting dynamical dark matter and dynamical dark energy, supporting the replacement of cold dark matter plus a cosmological constant by an evolving dark sector.","The DDM parameter space allowed by the tight ACT+DESI+DESY5 constraints produces measurable changes in the matter power spectrum, CMB lensing potential, velocity spectra, and lensing-galaxy correlations, so future surveys can test the model.","Adding Planck or DESY1 lowers the significance to about 3σ or 1.5σ, so the strength of the evidence depends on which datasets are combined."],"supporting_citations":[{"why":"Supplies the ACT DR6 power spectra and ACT-lite likelihood that provide the CMB data for the headline 3.4σ result.","marker":"[74]"},{"why":"Provides the DESI DR2 BAO measurements, including galaxy, quasar, and Lyman-alpha samples, that anchor the late-time expansion history.","marker":"[64]"},{"why":"Supplies the DESY5 type Ia supernova sample that adds low-redshift distance constraints to the combined fit.","marker":"[77]"},{"why":"Introduces the CPL parametrization whose functional form the paper borrows for the dark-matter equation of state.","marker":"[72, 73]"},{"why":"Provides the CAMB Boltzmann code that the paper modifies into DDMCAMB to compute spectra for dark-matter equation-of-state models.","marker":"[80]"},{"why":"Makes the case for the ACT-lite likelihood and low-ell Sroll2 prior used in the ACT-only analyses.","marker":"[78]"}],"fun_headline_variants":["Dynamical dark matter gets 3.4σ boost from ACT+DESI+SN","Dark matter's pressure may vary: 3.4σ evidence","3.4σ hint that dark matter is not pressureless","Dark matter equation of state evolves: 3.4σ","ACT DESI supernovae: 3.4σ for dynamical dark matter"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The 3.4σ result rests on the assumption that the ACT+DESI+DESY5 combination measures the dark-matter evolution parameter without a systematic bias; the paper itself notes that combining CMB and DESI data produces a larger Hubble constant through the CMB's sound-horizon calibration, and if that bias mimics the signal, the evidence would vanish.","fun_headline_variants_meta":{"raw":{"variants":["Dynamical dark matter gets 3.4σ boost from ACT+DESI+SN","Dark matter's pressure may vary: 3.4σ evidence","3.4σ hint that dark matter is not pressureless","Dark matter equation of state evolves: 3.4σ","ACT DESI supernovae: 3.4σ for dynamical dark matter"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000242,"raw_usage":{"total_tokens":1506,"prompt_tokens":906,"completion_tokens":600,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":522,"completion_tokens_details":{"reasoning_tokens":502}},"tokens_in":522,"tokens_out":600,"duration_ms":58016,"temperature":1.0,"reasoning_tokens":502,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T21:53:08.358586+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Re-run the ACT+DESI+DESY5 fit with the comoving sound horizon treated as a free parameter or calibrated by a local distance ladder; if ωdma drops to within about 1σ of zero, the 3.4σ is a calibration artefact. A second check is to compute the model-selection evidence for DDM against ΛCDM, since the single-parameter proportional model already returns ωdm=0.006±0.015.","supporting_citations":[{"cited_title":"Accelerating universes with scaling dark matter,","cited_arxiv_id":null,"evidence_quote":"Supplies the ACT DR6 power spectra and ACT-lite likelihood that provide the CMB data for the headline 3.4σ result."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the DESI DR2 BAO measurements, including galaxy, quasar, and Lyman-alpha samples, that anchor the late-time expansion history."},{"cited_title":"Planck 2018 results. V. CMB power spectra and likelihoods,","cited_arxiv_id":null,"evidence_quote":"Supplies the DESY5 type Ia supernova sample that adds low-redshift distance constraints to the combined fit."},{"cited_title":"The Atacama Cosmology Tele- scope: A Measurement of the DR6 CMB Lensing Power Spectrum and Its Implications for Structure Growth,","cited_arxiv_id":null,"evidence_quote":"Makes the case for the ACT-lite likelihood and low-ell Sroll2 prior used in the ACT-only analyses."}],"review_version":1}