{"id":"bc19d383-4e38-40d6-9014-60b63ab7fe92","arxiv_id":"1908.01727","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":3.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"COMPASS combines 2015 and partial 2018 Drell-Yan data to measure five transverse-spin azimuthal asymmetries, finding a Sivers asymmetry consistent with the QCD sign-change prediction at about one standard deviation.","lead":"This conference paper reports updated COMPASS measurements of transverse-spin asymmetries in Drell-Yan production using part of the 2018 data combined with the 2015 results. The data are consistent with the predicted sign change of the Sivers function between semi-inclusive deep inelastic scattering and Drell-Yan, though at low statistical significance.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Twist-3 F2_UT is assumed zero in footnote 2 with no kinematic bound; if non-negligible at COMPASS's 4.3-8.5 GeV scale, the quoted A_T^{sin phi_S} is not a clean twist-2 Sivers observable.","rationale":"The reader identified the same weakest assumption, and I find it genuinely load-bearing. The leading-twist interpretation is not an inessential detail: it is the bridge from the measured number to the QCD sign-change prediction. The one-sigma significance of the central value is openly acknowledged and is not itself a flaw; a preliminary result can legitimately be reported as consistent with a prediction at modest significance. The more serious problem is that even with perfect statistics, the extracted observable would not be the twist-2 Sivers asymmetry unless F2_UT is negligible. Since the hard scale is only 4-8 GeV, the usual power-suppression argument is not automatically safe, and the manuscript provides no numerical check. The suggested re-analysis with a modelled F2_UT would settle this directly. I do not see an internal inconsistency or a reason to reject; the appropriate reaction is to keep the conditional verdict and request the missing twist-3 estimate. Therefore no change to the reader's verdict.","tokens_in":7856,"tokens_out":8770,"duration_ms":100189,"concrete_test":"Re-analyze the combined 2015 plus 50% of 2018 dimuon sample with a Monte Carlo or analytic model that switches on F2_UT using a twist-3 parametrization, for example a Wandzura-Wilczek-type relation or the model inputs used in Refs. [14-16], and repeat the unbinned maximum-likelihood extraction with the same binning. Compare the resulting integrated A_T^{sin phi_S} with the F2_UT = 0 result; if the central value shifts by an amount comparable to its statistical uncertainty, or if the sign flips, the assumption is load-bearing and a bound or corrected extraction must be published before claiming consistency with the Sivers sign-change prediction. If the shift is negligible, the twist-3 concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing assumption is the one in footnote 2: 'the F2_UT structure function is assumed to be zero [3].' The measured A_T^{sin phi_S} in Eq. 1.1 is interpreted as the leading-twist Sivers asymmetry, and the central claim that its positive sign is consistent with the QCD sign change depends on that identification. F2_UT is the twist-3 piece that enters the same transverse-spin azimuthal modulation; if it is not zero, the amplitude extracted under Eq. 1.1 is a mixture of twist-2 and twist-3 terms. At M_mumu = 4.3-8.5 GeV, the suppression M/Q is only about 0.1-0.2, which is not negligible relative to a measured asymmetry of order 0.05 and its statistical uncertainty. The paper gives no numerical estimate, model calculation, or kinematic argument showing that this contribution is small in the x_N, x_pi, q_T region considered. If F2_UT is not small, the extracted sign and amplitude cannot be directly compared with the LO twist-2 Sivers function extracted from SIDIS, so the sign-change test is contaminated. This does not invalidate the measurement, but it makes the central interpretation conditional on an unquantified assumption.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This proceedings paper reports preliminary COMPASS measurements of the five target transverse-spin asymmetries in pion-nucleon Drell-Yan production at 190 GeV/c, combining the published 2015 data with about 50% of the 2018 data. The analysis uses an extended unbinned maximum likelihood fit in the dimuon mass range 4.3 < M_mumu < 8.5 GeV/c^2 and q_T > 0.4 GeV/c, with a total background estimated below 4%. Kinematic dependences in x_N, x_pi, x_F, q_T, and M_mumu are presented together with the integrated values. The central physics statement is that the integrated Sivers asymmetry A_T^{sin phi_S} is above zero at about one standard deviation of the total uncertainty, and that this positive sign and amplitude are consistent with the predicted QCD sign change of the Sivers function between SIDIS and Drell-Yan. The other four TSAs are also reported; one is approximately two standard deviations below zero, one is compatible with zero, and the two subleading-twist asymmetries are compatible with zero.","tokens_in":8129,"tokens_out":6320,"duration_ms":69687,"significance":"If the interpretation holds, these data provide one of the only direct experimental probes of the predicted universality properties of TMD parton distributions, and the Sivers sign-change test is a central prediction of QCD factorization. The paper has several strengths: it uses a standard unbinned maximum-likelihood estimator with simultaneous extraction of all five TSAs, it cross-checks 2015 and 2018 kinematic distributions, it estimates background at the few-percent level, and it reports point-to-point systematic uncertainties as bands. However, the statistical power is currently limited: the Sivers asymmetry is only about one standard deviation above zero, and the extraction relies on an unquantified assumption about a twist-3 structure function. The result is therefore best viewed as a preliminary consistency check rather than a precision test of the sign-change prediction.","major_comments":[{"comment":"The interpretation of the measured A_T^{sin phi_S} as the leading-twist Sivers asymmetry depends on the assumption that the twist-3 structure function F2_UT is zero, stated in footnote 2. At the COMPASS hard scale, 4.3 < M_mumu < 8.5 GeV/c^2, twist-3 admixtures need not be negligible compared with the measured asymmetry size, and the paper provides no numerical estimate, model calculation, or kinematic argument showing that F2_UT is small in the x_N, x_pi, q_T region considered. Because the sign-change comparison is the central claim of the paper, I ask that the authors either quantify the expected size of the twist-3 contamination or explicitly state in the conclusions that the sign-change interpretation holds only under this assumption.","section":"Section 1, footnote 2 and Eq. (1.1)"},{"comment":"The paper states that the positive sign and amplitude of the integrated Sivers asymmetry are consistent with the predicted sign change, but it does not quote the numerical value of A_T^{sin phi_S} or its statistical and systematic uncertainties, nor does it show the range expected from the SIDIS-based Sivers extraction. With the integrated value only about one standard deviation above zero, the evidence is weak; the text should report the central value and uncertainty and frame the statement as a consistency check rather than a confirmation of the sign-change prediction.","section":"Section 3, integrated results"},{"comment":"The point-to-point systematic uncertainties are stated to be about 0.7 times the statistical uncertainties, but no decomposition or method is described. Since the key Sivers result is significant at only about one standard deviation of the total uncertainty, the way the systematic contribution is evaluated directly affects the central claim; please provide at least a summary of the sources entering this factor.","section":"Section 2, systematics"}],"minor_comments":[{"comment":"The running title in the full text reads \"COMP ASS\" due to a spacing error; it should read \"COMPASS\".","section":"Title/header"},{"comment":"The abstract mentions spin-(in)dependent azimuthal asymmetries, but this paper reports only transverse-spin asymmetries; the wording could be tightened to avoid implying spin-independent results are included.","section":"Abstract"},{"comment":"The quantities F1_U and F2_U are used to define sigma_hat_U, but the symbols are not defined directly after the equation; a one-sentence definition would improve readability.","section":"Equation (1.1)"},{"comment":"The assumption that F2_UT is zero is attributed to Ref. [3]; a physics reference or a short justification would be more informative than a citation to the COMPASS-II proposal.","section":"Footnote 2"},{"comment":"The asymmetry A_T^{sin(2 phi_CS - phi_S)} is reported as being below zero with a significance of about two standard deviations; because five asymmetries are extracted simultaneously, the discussion would benefit from a note on whether the significance accounts for the multiple-comparison/trial factor.","section":"Section 3"},{"comment":"The integrated values shown in the rightmost panel of Fig. 4 are difficult to read without numerical values; a small table or numeric labels would help the reader assess the amplitudes and uncertainties.","section":"Figures"}],"recommendation":"major_revision","confidential_remarks":"This is a conference proceedings contribution, so the requested changes can be brief: a quantitative discussion of the twist-3 assumption, numerical values for the integrated asymmetries, and a more cautious wording of the sign-change conclusion. The underlying analysis appears sound as a preliminary measurement, but the central physics claim is currently conditional on an unquantified assumption, which is why I recommend major revision rather than minor revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe one thing to know: this is a conference-proceedings status report from COMPASS, updating the published 2015 Drell-Yan TSA results with the first ~50% of the 2018 data. The new combined average Sivers asymmetry A_T^sin phi_S is positive at about one standard deviation. That is a hint, not a measurement, and the paper mostly says so, but the phrase 'consistent with the predicted sign change' is easy to overread.\n\nWhat the paper does well: the analysis is standard and careful — unbinned maximum likelihood fit to all five TSAs, background below 4%, point-to-point systematics quoted at 0.7 times the statistical error, and a nice check that 2015 and 2018 kinematic distributions overlap. The 2015 data were already published in PRL 119, 112002, so the genuinely new content is the partial 2018 statistics and the combined numbers.\n\nThe soft spots, in order of seriousness. First, the central Sivers asymmetry is only ~1σ above zero, so this is a weak consistency check, not a test of the sign-change prediction. Second, and more substantive, the extraction identifies A_T^sin phi_S with the twist-2 Sivers term and assumes F2_UT = 0 in footnote 2, with no quantitative justification. The stress-test note is right: at M_mumu = 4.3–8.5 GeV, M/Q is 0.1–0.2, so a twist-3 admixture is not obviously negligible compared with an asymmetry of order 0.05. The assumption is standard and appears in the published PRL, but for a paper whose point is the sign-change comparison, a sentence showing the expected effect is small would materially strengthen it. Third, the 'first and only currently available data' sentence is true in a narrow sense but will date quickly.\n\nBottom line: an honest, methodologically sound incremental update. Not groundbreaking, and the F2_UT caveat should be addressed if this ever appears in a full journal paper. As a proceedings article, it's a useful service to the spin community. I'd send it to a serious referee if submitted to a journal, but I wouldn't cite it instead of the PRL.","headline":"Preliminary COMPASS DY update: Sivers asymmetry positive at ~1σ, with an unquantified twist-3 F2_UT assumption as the main caveat.","tokens_in":8634,"tokens_out":4253,"would_cite":false,"duration_ms":42180,"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":"Drell-Yan data point to the Sivers sign QCD predicts","keywords":["Drell-Yan process","Sivers asymmetry","transverse spin","TMD parton distribution functions","azimuthal asymmetries","COMPASS","sign change","transversity"],"falsifier":"A concrete test: re-extract $A_T^{\\sin\\phi_S}$ without assuming $F_2^{UT}=0$, for instance by including a model estimate of $F_2^{UT}$ or by studying the asymmetry at lower $q_T$ where higher-twist effects grow; if the sign ceases to be positive or the amplitude departs from the SIDIS-opposite expectation, the claim that the data are consistent with the Sivers sign change would be contradicted.","tokens_in":7662,"feed_emoji":"⚛️","tokens_out":11345,"duration_ms":91346,"temperature":0.7,"pith_summary":"This paper presents the first measurements of transverse-spin-dependent azimuthal asymmetries in Drell-Yan lepton-pair production from a transversely polarized nucleon target, combining COMPASS 2015 data with about half of the 2018 data. Its central result is that the average Sivers asymmetry is positive, at roughly one standard deviation of the total uncertainty, and that the sign and size are consistent with the QCD prediction that the Sivers function changes sign between semi-inclusive deep-inelastic scattering (SIDIS) and Drell-Yan. The paper also finds a two-standard-deviation negative value for the related transversity-sensitive asymmetry, in agreement with model calculations, while the remaining asymmetries are consistent with zero. Because the COMPASS SIDIS measurements were taken at the same hard scale, this comparison tests the predicted sign change with reduced theoretical uncertainty from TMD evolution. If confirmed with the full 2018 statistics, the result would be direct experimental evidence for the process-dependent universality of transverse-momentum-dependent parton distributions.","feed_headline":"Drell-Yan data point to the Sivers sign QCD predicts","feed_subtitle":"COMPASS 2015 plus half of 2018 data put the asymmetry on the predicted sign-change side of the Sivers function.","key_machinery":"The central object is the Sivers transverse-momentum-dependent parton distribution function, $f_{1T}^{\\perp}$: a function describing how the intrinsic transverse momentum of quarks inside a transversely polarized nucleon correlates with the nucleon spin. The load-bearing identity is the QCD prediction of its process dependence: the same naive-time-reversal-odd distribution is expected to appear with opposite signs in SIDIS and in Drell-Yan, a consequence of the gauge-link (Wilson line) structure of TMD operators. The measurement isolates the effect by extracting the $\\sin\\phi_S$ azimuthal modulation in the Drell-Yan cross section off a transversely polarized target, using an unbinned maximum-likelihood fit of five target-spin-dependent asymmetries simultaneously, and by choosing a hard-scale range matched to earlier COMPASS SIDIS measurements.","core_discovery":"The paper claims that the combined 2015 plus partial 2018 Drell-Yan data from a $\\pi^-$ beam on a transversely polarized NH$_3$ target yield a positive average Sivers asymmetry $A_T^{\\sin\\phi_S}$, at about one standard deviation of the total uncertainty, whose sign and magnitude match the predicted change of sign of the quark Sivers function between SIDIS and Drell-Yan. It further reports that the TSA $A_T^{\\sin(2\\phi_{CS}-\\phi_S)}$ is negative with about two standard-deviation significance, consistent with model calculations and relevant to the universality of nucleon transversity, and that the pretzelosity-related asymmetry $A_T^{\\sin(2\\phi_{CS}+\\phi_S)}$ and the two subleading-twist asymmetries are compatible with zero. The author presents these as the first and only currently available data on transverse-spin-dependent azimuthal asymmetries in the Drell-Yan process, and notes that analysis of the remaining 2018 data is ongoing and should improve statistical precision.","pith_inferences":["Beyond the paper: extracting the $\\cos 2\\phi_{CS}$ Boer-Mulders asymmetry from the same dataset would test a second naive-T-odd sign-flip prediction; a sign reversal relative to SIDIS would extend the universality test beyond the Sivers function.","Beyond the paper: because the 2018 sample is only about half-processed, the full dataset plausibly shrinks statistical uncertainties by roughly $\\sqrt{2}$, which could move the Sivers asymmetry from about one to about two standard deviations; this is an extrapolation, not a claim in the paper.","Beyond the paper: binning the asymmetry in $q_T$ and $M_{\\mu\\mu}$ while estimating $F_2^{UT}$ would directly test the assumption on which the sign-change comparison rests, since a $q_T$-dependent rise of the $\\sin\\phi_S$ amplitude would signal twist-3 contamination."],"forward_implications":["If the sign change is real, the Sivers function extracted from COMPASS Drell-Yan data should be the negative of the SIDIS-extracted function at the same hard scale, giving a direct experimental test of QCD gauge invariance.","The negative two-standard-deviation signal in $A_T^{\\sin(2\\phi_{CS}-\\phi_S)}$ provides a new constraint on the nucleon transversity TMD PDF when compared with model calculations.","Comparing SIDIS and Drell-Yan data at matched hard scale sidesteps much of the TMD-evolution uncertainty that complicates comparisons with the much higher-scale STAR W/Z measurements.","The pretzelosity-related asymmetry and the two subleading-twist asymmetries being compatible with zero supports the leading-twist interpretation of the measured TSAs."],"supporting_citations":[{"why":"Supplies the QCD prediction that the Sivers function reverses sign between SIDIS and Drell-Yan, the prediction this measurement is designed to test.","marker":"[12]"},{"why":"Published COMPASS 2015 Drell-Yan TSA measurement to which this letter adds about 50% of the 2018 data for the combined result.","marker":"[18]"},{"why":"COMPASS SIDIS Sivers extraction at the hard scale of Drell-Yan, enabling the same-scale comparison that reduces TMD evolution uncertainty.","marker":"[9]"},{"why":"COMPASS SIDIS Sivers measurement, one of the SIDIS inputs from which the SIDIS-side Sivers function is extracted.","marker":"[6]"},{"why":"HERMES observation of the naive-T-odd Sivers effect in SIDIS, the canonical SIDIS-side measurement.","marker":"[8]"},{"why":"STAR W/Z production asymmetry measurement, the higher-scale Drell-Yan-type Sivers test whose evolution connection to fixed-target data is discussed.","marker":"[17]"},{"why":"Model calculation used to compare the measured $A_T^{\\sin(2\\phi_{CS}-\\phi_S)}$ magnitude for transversity universality.","marker":"[19]"}],"fun_headline_variants":["COMPASS Drell-Yan sees Sivers sign flip predicted by QCD","First transverse-spin Drell-Yan asymmetry matches Sivers sign change","COMPASS Drell-Yan data favor Sivers sign change","Positive Sivers asymmetry in Drell-Yan supports QCD prediction"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The extraction assumes that the subleading background structure function $F_2^{UT}$ is zero; if a nonzero $F_2^{UT}$ contributes measurably at the COMPASS kinematics, the measured $\\sin\\phi_S$ asymmetry would contain an extra twist-3 piece and would no longer provide a clean test of the leading-order Sivers sign change.","fun_headline_variants_meta":{"raw":{"variants":["COMPASS Drell-Yan sees Sivers sign flip predicted by QCD","First transverse-spin Drell-Yan asymmetry matches Sivers sign change","COMPASS Drell-Yan data favor Sivers sign change","Positive Sivers asymmetry in Drell-Yan supports QCD prediction"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000288,"raw_usage":{"total_tokens":1690,"prompt_tokens":946,"completion_tokens":744,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":562,"completion_tokens_details":{"reasoning_tokens":667}},"tokens_in":562,"tokens_out":744,"duration_ms":7341,"temperature":1.0,"reasoning_tokens":667,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T15:04:05.033335+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A concrete test: re-extract $A_T^{\\sin\\phi_S}$ without assuming $F_2^{UT}=0$, for instance by including a model estimate of $F_2^{UT}$ or by studying the asymmetry at lower $q_T$ where higher-twist effects grow; if the sign ceases to be positive or the amplitude departs from the SIDIS-opposite expectation, the claim that the data are consistent with the Sivers sign change would be contradicted.","supporting_citations":[{"cited_title":"Sivers asymmetry extracted in SIDIS at the hard scale of the Drell-Yan process at COMPASS","cited_arxiv_id":"1609.07374","evidence_quote":"COMPASS SIDIS Sivers extraction at the hard scale of Drell-Yan, enabling the same-scale comparison that reduces TMD evolution uncertainty."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Model calculation used to compare the measured $A_T^{\\sin(2\\phi_{CS}-\\phi_S)}$ magnitude for transversity universality."}],"review_version":1}