{"id":"87e4f7e1-5af9-4f68-a3f5-28546605c71c","arxiv_id":"2601.05324","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"A global U-spin fit finds a severe Standard-Model tension in the longitudinal polarisation of Bs→K*0K*0 once the naive-factorisation helicity hierarchy is imposed, and simple b→s new-physics operators only partially relieve it.","lead":"Using U-spin symmetry to relate eight B→V1V2 decay channels, the authors find that the Standard Model cannot simultaneously explain the measured polarisation fractions once the expected helicity hierarchy is imposed, with Bs→K*0K*0 the most discrepant mode. The paper quantifies the tension, shows that 30% U-spin breaking does not help, and finds that simple new-physics operators improve but do not fully cure the fit.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Central SM 'puzzle' is driven by the imposed helicity hierarchy; the paper's own free-amplitude fit shows data are SM-compatible once the hierarchy is relaxed.","rationale":"I read the manuscript with the central argument being: measuring f_L in U-spin-related B→V1V2 modes and imposing the naive-factorisation hierarchy yields a poor SM fit; hence either large nonfactorisable SM effects or NP are needed. The most load-bearing premise is the hierarchy, since the paper's own unconstrained fit (Table III) shows the same data are perfectly consistent with SM U-spin amplitudes when A−/A0 is free. Therefore the 'SM failure' reduces to the assumed hierarchy. The paper does not independently validate the hierarchy beyond citing Ref. [22]; it explicitly allows nonfactorisable contributions as a possible SM resolution. Thus the central claim is conditional. This matches the reader's weakest assumption. I do not see a stronger concern: the 7σ vs 3.5σ mismatch in the abstract is a reporting issue, and the NP fits remain poor (χ²/dof ~30/10), so the paper does not claim a full NP solution. The correct verdict remains CONDITIONAL, contingent on a controlled calculation of the helicity ratios.","tokens_in":18976,"tokens_out":7836,"duration_ms":86289,"concrete_test":"Compute the helicity amplitude ratio |A−/A0| for B_s→K*0K*0 (and B_d→K*0K*0) using QCD factorization at NLO including 1/m_b and nonfactorisable corrections (as in Ref. [22]), with the hard-collinear scale varied by a factor of 2 and form-factor inputs varied within their errors. If the predicted ratio is 0.3 with uncertainty ≲0.2, the data's extracted ~1.7 is a genuine SM discrepancy and the central claim stands. If the uncertainty is O(1) or the central value shifts toward unity, the 'puzzle' is an artifact of the unjustified hierarchy.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim—that the SM is disfavoured by the LPF data—is not intrinsic to the data or U-spin, but is produced by the a priori imposition of the naive-factorisation helicity hierarchy (Section II, Eq. (7) and the paragraph 'We restore the hierarchy... uniformly take A−/A0=0.3, A+=0'). This is demonstrated inside the paper: with the same U-spin decomposition and only A+=0, the SM fit to the same 17 observables is excellent (χ²/dof=3.4/5; Table III), and the best-fit |A−/A0| ratios for ΔS=1 modes are 1.03±0.19, 1.72±0.54, 1.28±0.40 (Eq. (19))—far from 0.3. Thus the 'discrepancy' is a corollary of assuming factorisation dominance, not a conclusion forced by data. The paper acknowledges that 'within SM, the only effective resolution would be to allow for large nonfactorisable contributions', but this is exactly the assumption being made. No independent evidence is provided that nonfactorisable terms are subdominant in penguin-dominated b→s VV decays; indeed, the measured f_L(B_s→φφ)=0.379±0.008 lies far below the hierarchy expectation, suggesting that power corrections are systematically large in b→s VV modes. Unless the hierarchy is separately validated (e.g., by a first-principles calculation with controlled uncertainties), the central claim does not follow.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a U-spin analysis of eight B→V1V2 channels, using 17 measured observables (branching ratios, CP asymmetries, polarisation fractions) to fit U-spin reduced matrix elements. With only A_+=0, the SM fit is good (χ²/dof=3.4/5), but the best-fit transverse amplitudes violate the expected helicity hierarchy, especially for ΔS=1 modes. Imposing A_-/A_0=0.3 and A_+=0 produces χ²/dof=370.5/11, driven mainly by f_L(Bs→K*0K̄*0). The authors show that allowing up to 30% U-spin breaking does not cure this, and that flavour-specific NP operators in b→s can improve the fit but do not make it good. They also predict LPFs for the unobserved modes Bs→K*+K*- and Bs→K*-ρ+.","tokens_in":19337,"tokens_out":7236,"duration_ms":78442,"significance":"If the helicity-hierarchy assumption were independently established, the paper would document a serious challenge to QCDF/naive factorisation, with a concrete observable ratio at about 3.5σ. The U-spin decomposition is clean, the fits are transparent, and the NP survey is a useful bottom-up study. The paper also offers falsifiable predictions for unobserved modes, notably Bs→K*+K*-, whose f_L differs strongly among NP scenarios. However, as the paper itself demonstrates, the central 'puzzle' is conditional on the imposed hierarchy; without independent evidence that nonfactorisable contributions are subdominant, the claim that the SM is disfavoured is not established.","major_comments":[{"comment":"The SM fit with 12 RMEs and only A_+=0 gives χ²/dof=3.4/5 (p=0.64), and Eq. (19) yields |A_-/A_0| for the ΔS=1 modes as 1.03±0.19, 1.72±0.54, and 1.28±0.40, far from the imposed 0.3. Thus the data alone do not disfavour the SM; the catastrophic χ²/dof=370.5/11 in Table IV is entirely produced by the external assumption A_-/A_0=0.3, A_+=0. The paper does not provide an independent estimate of the size of nonfactorisable contributions to these penguin modes; it assumes they are subdominant before Eq. (7). The abstract and conclusions state that the SM yields a poor fit, which overstates the result. Please reframe the analysis as a test of the factorisation/helicity-hierarchy assumption, and either derive the hierarchy from a controlled calculation with uncertainties or present the direct f_L ratio in Eqs. (16)–(17) as the central observable.","section":"§II.B, Table III and Eq. (19)"},{"comment":"The treatment of U-spin breaking by randomising RME ratios within [0.7,1.3] is not described as a statistical procedure. It is not stated how many samples are generated, whether the sampled ratios are then fixed and fitted, or how the resulting distribution of χ² is used. Since the robustness to U-spin breaking is part of the argument that the SM cannot explain the data, please specify the procedure exactly and report the best χ² and p-value obtained after the randomisation. The 30% range also appears arbitrary; a profile over nuisance parameters representing U-spin breaking would be more defensible.","section":"§II.B, Eq. (20)"},{"comment":"The NP fits are constructed so that NP acts only on ΔS=1 modes through operators specifically chosen to enhance the transverse helicity amplitude, so the improvement over the SM hierarchy fit is partly built in. More importantly, the SM part in these fits still assumes exact U-spin and the 0.3 hierarchy. Comparing Table VI (χ²/dof≈3–7) with the relaxed-hierarchy SM fit in Table III (χ²/dof=3.4/5) shows that the data are better described by relaxing the hierarchy within the SM than by any of the NP scenarios considered. The paper should state this explicitly. In addition, the prediction for f_L(Bs→K*+K*-) changes from 0.50±0.04 in Table I to 0.26–0.79 in Table VII depending on the NP model; these are model-dependent predictions, not independent measurements, and the text should make that clear.","section":"§III, Table VI"},{"comment":"The abstract says 'the Standard Model (SM) yields only a very poor agreement with the data' and that the individual disagreement for Bs→K*0K̄*0 is 'more than 7σ'. Both statements are made without the qualifier that they follow only after imposing the helicity hierarchy A_-/A_0=0.3, A_+=0. Given the paper's own Table III, these statements are misleading as written. The abstract and conclusions should present the result as a conditional statement: if the factorisation hierarchy is imposed, the SM fit is extremely poor; without that assumption, the SM is compatible with the data. Also, the phrase 'We restore the hierarchy' in §II.B should be 'We impose the hierarchy', since it is an external assumption.","section":"Abstract and Conclusions"}],"minor_comments":[{"comment":"The text says the hierarchy is approximately A_0:A_-:A_+ ≃ 1:0.27:0.01 from naive factorisation, but then uniformly sets A_-/A_0=0.3 and A_+=0. Please clarify whether 0.3 is the numerical form-factor result or an order-of-magnitude rounding, and state the corresponding f_L values implied by the hierarchy.","section":"§II.B after Eq. (7)"},{"comment":"The entries marked ♠ are predictions, not measurements, but they are displayed in the same table as the experimental data. It would be clearer to separate predictions from data, or at least state in the caption that the ♠ rows are model-dependent outputs of the fit, not inputs.","section":"Table I"},{"comment":"There are minor language issues, e.g. 'consider, instead, to a NP effective Hamiltonian' should read 'consider instead a NP effective Hamiltonian'. Also, the definitions of c_i are given only by examples; a sentence stating that c_i are chiral projection coefficients would help readability.","section":"§III, Eq. (22)"}],"recommendation":"major_revision","confidential_remarks":"The paper contains a useful global U-spin fit and a sharp, falsifiable comparison of f_L ratios, but its headline claim is conditional on an assumption that the authors do not independently validate. I think the manuscript can be made publishable if the authors reframe the conclusion as a test of the factorisation/helicity-hierarchy assumption and quantify the uncertainty on that assumption. As it stands, the abstract overstates the case, so I recommend major revision rather than acceptance."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The useful thing in this paper is the explicit demonstration that the B→V1V2 LPF data, treated U-spin-symmetrically with free helicity amplitudes, are perfectly SM-compatible (χ²/dof = 3.4/5). The apparent “puzzle” appears only once you impose |A−/A0| = 0.3 and call that the SM expectation. That is the central result, and the paper deserves credit for showing it so cleanly, including the robustness to ±30% U-spin breaking and the systematic NP operator scan, which is a reasonable piece of work even though no NP scenario actually leaves you with a good fit.\n\nThe soft spots are real but mostly in the interpretation. First, the abstract quotes “more than 7σ” while the body quotes about 3.5σ for the same ratio; that inconsistency should have been caught. Second, the entire tension rests on one precise measurement, f_L(Bs→K*0K*0), and on the imposed hierarchy ratio. The paper itself admits that large nonfactorisable contributions would dissolve the puzzle, so the claim is conditional on an assumption that is plausible but unquantified. Third, the NP fits never reach a good χ² (best about 29/10), so the paper cannot claim evidence for NP; at best it shows certain operators help, which is what you'd expect when you add an amplitude that breaks the hierarchy.\n\nWhat is genuinely new: the global U-spin fit over eight modes, the explicit randomisation of U-spin breaking, and the demonstration that moderate breaking cannot rescue the hierarchy-constrained fit. The treatment of the two unobserved modes as predictions is a nice touch. The paper is written honestly—it reports the good free fit and does not hide the fact that the hierarchy is an input.\n\nWho should read it? Anyone working on non-leptonic B decays, and anyone who needs a case study in how the choice of hadronic framework determines whether a “puzzle” exists. I would not cite its central tension as evidence for NP, but I would cite the global fit as a quantitative U-spin benchmark.\n\nPeer review: yes, send it. The referee should insist on a corrected abstract, a clearer separation between “SM without hierarchy” and “SM with hierarchy”, and a more careful discussion of the single-mode dependence. The analysis is solid enough to warrant a serious referee despite the caveats.","headline":"A thorough U-spin fit that reframes the B→V1V2 polarisation puzzle as a test of the helicity hierarchy assumption rather than a data-driven SM failure.","tokens_in":19936,"tokens_out":1649,"would_cite":false,"duration_ms":20774,"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":"The paper establishes that, with the naive factorisation helicity hierarchy imposed, the Standard Model cannot fit B→V1V2 polarisation data, mainly because f_L(Bs→K*0K*0) sits 3.5σ below its U-spin expectation.","keywords":["B meson decays","U-spin symmetry","longitudinal polarisation fractions","helicity amplitudes","heavy-quark factorisation","b→s penguins","new physics in B decays","QCD factorisation"],"falsifier":"A larger, high-precision angular analysis of Bs→K*0K*0 (and ideally Bd→K*0K*0 in the same experiment) that measures f_L and the transverse amplitude ratio |A−/A0| directly: if f_L moves toward 0.6–0.7 and |A−/A0| near 0.3, the tension vanishes; if |A−/A0| stays ≳1 while f_L≈0.24, the hierarchy breakdown is confirmed and the paper's NP or large-nonfactorisable conclusion stands.","tokens_in":18778,"feed_emoji":"⚛️","tokens_out":6112,"duration_ms":62117,"temperature":0.7,"pith_summary":"The paper tries to show that a set of eight B decays to two light vector mesons, connected pairwise by U-spin symmetry (d↔s), cannot be described by the Standard Model once one accepts the heavy-quark-factorisation hierarchy between helicity amplitudes (A0:A−:A+ ≈ 1:0.3:0). With that hierarchy imposed, a global fit to 17 measurements gives χ²/dof = 370.5/11, and even allowing 30% U-spin breaking barely improves it. The culprit is the longitudinal polarisation fraction in ΔS=1 decays, especially Bs→K*0K*0, whose value 0.24±0.04 is only about a third of its U-spin partner Bd→K*0K*0 (0.74±0.05), corresponding to a 3.5σ discrepancy. If the hierarchy is dropped and amplitudes are left free, the fit is excellent, but the extracted amplitudes violate the hierarchy, meaning the Standard Model would need large nonfactorisable contributions. The paper then shows that flavour-specific new physics in b→s transitions with non-hierarchical Lorentz structures relieves, but does not fully cure, the tension.","feed_headline":"Helicity hierarchy fails in B_s→K*0K*0 by 3.5σ","feed_subtitle":"Longitudinal fraction is 0.24 vs 0.74 in its U-spin partner; only big nonfactorisable effects or new physics can fix it.","key_machinery":"The argument runs on two ingredients put together: U-spin SU(2)U, which relates ΔS=0 and ΔS=1 amplitudes by interchanging d and s quarks and expresses all eight decay amplitudes through 12 reduced matrix elements (RMEs) A_{nD}^{p,h}; and the naive-factorisation heavy-quark hierarchy A0:A−:A+ ≈ 1:ΛQCD/mb:(ΛQCD/mb)^2 ≈ 1:0.3:0, which is used to cut the 12 RMEs to 6. The U-spin RME decomposition gives the fit its structure; the hierarchy is what makes the fit predictive and, as the data show, wrong for ΔS=1. NP is introduced via effective b→s operators of scalar, vector, and tensor type, whose helicity dependence is computed under factorisation and added to the SM amplitudes.","core_discovery":"Within the U-spin framework, the paper finds a dichotomy: exact U-spin with free helicity amplitudes gives a good fit (χ²/dof = 3.4/5), but the best-fit |A−/A0| ratios for ΔS=1 modes are roughly 1–1.7, violating the expected 0.3 hierarchy; imposing the hierarchy uniformly makes the fit fail at χ²/dof = 370.5/11, even with 30% U-spin breaking. The authors conclude that within the SM, only large nonfactorisable contributions to all these amplitudes could resolve the puzzle, and that otherwise one needs new physics in b→s that does not respect the hierarchy. Fits with scalar, vector, or tensor NP operators improve the description but remain poor, with best χ²/dof ≈ 29/10; the NP amplitudes that","pith_inferences":["If nonfactorisable contributions really are large enough to break the hierarchy, the same mechanism could contaminate other penguin-dominated b→s observables; the paper's NP brackets should then be read as bounds on short-distance effects rather than signals.","The U-spin framework treats the hierarchy as a theory input rather than a fitted quantity. A cleaner test would be to fit A−/A0 per mode as a pull parameter; the paper's own free-amplitude fit already indicates a mode-dependent breakdown.","The predicted spread in f_L(Bs→K*+K*−) is essentially a map of Lorentz structures: near 0.3 favours scalar or tensor NP, near 0.8 favours vector (V−A)⊗(V+A) NP; this is a testable short-term discriminator.","Should f_L(Bs→K*0K*0) shift upward with more data, the puzzle would collapse into a statistical fluctuation; the paper's strongest phenomenological legacy would then be the systematic U-spin framework itself."],"forward_implications":["If the hierarchy is a genuine SM prediction, the SM cannot account for the measured B→V1V2 polarisations; abandoning naive factorisation for penguin-dominated b→s decays is the only in-SM way out.","Permitting U-spin breaking up to 30% does not rescue the fit, so the puzzle is not a moderate SU(3)-breaking effect.","New physics restricted to b→s with non-hierarchical helicity amplitudes improves the fit for several scalar/vector/tensor operators, but none of the simple single-operator scenarios yields a fully satisfactory χ².","The unpolarised branching fractions and CP asymmetries remain consistent with data once NP is added, because the NP mainly feeds transverse amplitudes.","A measurement of f_L(Bs→K*+K*−) is predicted to lie anywhere from about 0.26 to about 0.79 depending on the NP Lorentz structure, providing a direct discriminator."],"fun_headline_variants":["U-spin hierarchy fails: B_s→K*K* polarisation off by 7σ","B_s→K*0K*0 polarisation breaks U-spin at 7σ","Helicity hierarchy impossible under U-spin; NP needed","U-spin fit collapses when hierarchy imposed; >7σ for B_s","Polarisation puzzle: only nonfactorisable or NP can save U-spin"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The load-bearing premise is that |A−/A0|=0.3 and A+=0 is the correct Standard Model expectation for every mode—if nonfactorisable corrections are large enough to break this hierarchy, the poor SM fit and the inferred need for new physics both disappear.","fun_headline_variants_meta":{"raw":{"variants":["U-spin hierarchy fails: B_s→K*K* polarisation off by 7σ","B_s→K*0K*0 polarisation breaks U-spin at 7σ","Helicity hierarchy impossible under U-spin; NP needed","U-spin fit collapses when hierarchy imposed; >7σ for B_s","Polarisation puzzle: only nonfactorisable or NP can save U-spin"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000248,"raw_usage":{"total_tokens":1498,"prompt_tokens":977,"completion_tokens":521,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":721,"completion_tokens_details":{"reasoning_tokens":414}},"tokens_in":721,"tokens_out":521,"duration_ms":5749,"temperature":1.0,"reasoning_tokens":414,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T11:40:45.836932+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A larger, high-precision angular analysis of Bs→K*0K*0 (and ideally Bd→K*0K*0 in the same experiment) that measures f_L and the transverse amplitude ratio |A−/A0| directly: if f_L moves toward 0.6–0.7 and |A−/A0| near 0.3, the tension vanishes; if |A−/A0| stays ≳1 while f_L≈0.24, the hierarchy breakdown is confirmed and the paper's NP or large-nonfactorisable conclusion stands.","supporting_citations":[],"review_version":1}