{"id":"2bf322ca-ed01-45e6-a826-5545b1298b0f","arxiv_id":"2607.09477","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":8.0,"correctness_risk":"low","formal_verification":"none","parameter_count":1,"one_line_summary":"Even fully separable quantum states yield advantageous quantum correlated equilibria in suitably constructed Bayesian games that outperform all classical correlated equilibria.","lead":"Separable quantum states can produce Bayesian-game equilibria that beat every classically correlated equilibrium. This removes the need for entanglement and brings quantum game advantage closer to experiment.","discovery_kind":"extension","skeptic_critique":{"model":"grok-4.5","headline":"No significant objection identified","rationale":"The paper’s analytic argument for pseudo-telepathy games is self-contained and the quantitative control of the guessing game is proved rather than merely assumed. The numerical LP results for CHSH, magic square and GHZ supply independent corroboration that a finite cutoff Λ already forces classical social welfare down to β(G). Because the reader already identified the same technical hinge and correctly judged it secure, no adjustment of the ACCEPT verdict is warranted.","tokens_in":22946,"tokens_out":384,"duration_ms":4423,"concrete_test":"Independently re-derive the lower bound (6) of Lemma 2 from the positive-semidefiniteness of W and the definition of total variation; confirm that the constant |X|/(|X|-1)/∥p∥^{2} is tight for the uniform distribution used in the CHSH and magic-square examples.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim of Theorem 3 (and multi-player analogues) holds under the stated hypotheses. The construction of eG_Λ from a quantum pseudo-telepathy game G, the separable cq-state ω of Eq. (9), and the equilibrium verification for both players are complete. Lemma 2 and the quantitative bound (6) correctly force any classical correlated equilibrium either to keep Alice’s type distribution within total-variation ε of the prior (hence EV ≤ β(G)+ε) or to produce a strictly positive W-payoff that Alice can annihilate by reverting to p, destroying equilibrium once Λ is large. The CHSH case is handled by direct verification of Alice’s four local deviations (Fig. 4) plus LP, confirming the same gap. The reader’s weakest-assumption concern is therefore already discharged by the paper’s own proofs; no further load-bearing gap remains.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The paper constructs Bayesian games of incomplete information from nonlocal (especially pseudo-telepathy) games such that a fully separable classical-quantum advice state yields a quantum correlated equilibrium of social welfare 1, while every classically correlated equilibrium has social welfare at most β(G)+ε once the mechanism parameter Λ is large enough. The construction equips the original winning predicate V with a carefully chosen zero-sum guessing game W that forces Alice’s type distribution to stay near the prior at classical equilibrium; the separable state ω of Eq. (9) together with Bob’s original POVMs then forms a quantum correlated equilibrium (Theorems 3–6). Explicit verification for CHSH (Fig. 4) and linear-programming comparisons for Magic Square, GHZ and CHSH confirm a strict social-welfare gap. The same construction also separates classically correlated equilibria from the larger set of local communication equilibria, refining Forges’ hierarchy (Appendix A).","tokens_in":23159,"tokens_out":900,"duration_ms":25626,"significance":"The central claim is new and load-bearing for quantum game theory: quantum advantage in competitive Bayesian games need not rely on entanglement. Because the advice states are fully separable (and in the CHSH/GHZ cases extremely simple), the result materially lowers the experimental barrier relative to all prior examples that required highly entangled states. The analytic proofs for every pseudo-telepathy game, the explicit mechanism-design matrix W, the direct four-case verification for CHSH, and the linear-programming confirmation of the predicted gap constitute reproducible, falsifiable evidence. The separation Corr(eG_Λ) ⊊ BI(eG_Λ) ∩ LO is an independent conceptual contribution to the taxonomy of correlated equilibria. These strengths make the manuscript a clear advance for both foundations and near-term quantum advantage.","major_comments":[],"minor_comments":[{"comment":"Section IV.A: the Magic-Square linear program is reported only for a limited sample of Λ values because of the 20 736-dimensional strategy space. A short remark on the computational limitation (or a reduced-symmetry formulation) would help readers assess how complete the numerical support is for that example.","section":null},{"comment":"The multi-player Theorems 5–6 are stated as “analogues” with proofs omitted. A one-paragraph sketch confirming that the reduction to the two-player case (grouping the classical players) preserves both the equilibrium conditions and the quantitative total-variation bound would remove any residual ambiguity.","section":null},{"comment":"Figure captions (Figs. 1–3, 5–6) rely heavily on the label convention introduced at the start of Section IV. Adding a one-line legend inside each panel (or expanding the captions) would make the plots self-contained.","section":null},{"comment":"Appendix A introduces eleven notions of equilibrium and leaves several inclusions open. Flagging which separations are already witnessed by the present construction versus which remain open would sharpen the contribution of the appendix.","section":null},{"comment":"Notation density is high (e.g., eG_Λ, ω, W, eta(G), \tau). A short table of symbols early in Section II would improve readability without changing content.","section":null},{"comment":"Page 11, Remark after Theorem 3: the phrase “price of privacy, or indeed the price of knowing too much” is evocative but informal; a single clarifying sentence tying it back to the total-variation argument would keep the tone consistent with the rest of the paper.","section":null}],"recommendation":"accept","confidential_remarks":"The manuscript is technically solid and the novelty claim is well supported. The future arXiv date (10 July 2026) is an artifact of the source and does not affect the science. Fit for a quantum-information or foundations journal is excellent; the practical-realisation angle may also interest a broader quantum-technology audience."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"The real news is that quantum advantage in competitive Bayesian games no longer needs entanglement. They take any quantum pseudo-telepathy game G, build a modified Bayesian game eG_Λ by adding a carefully chosen zero-sum guessing matrix W, and equip the players with a fully separable cq-state ω that is just Alice’s classical sample of her type plus Bob’s post-measurement state. Theorems 3–6 prove that this is a quantum correlated equilibrium of social welfare 1, while every classical correlated equilibrium is forced down to social welfare at most β(G)+ε once Λ is large enough. The same construction also yields the first example of a local communication equilibrium that is not a classical correlated equilibrium.\n\nWhat they do well is keep the argument constructive and checkable. Lemma 2 and the total-variation bound (6) give quantitative control of the guessing game; the CHSH case is verified by hand on the four BB84 post-measurement states (Fig. 4) plus linear programming; Magic Square and GHZ get the same LP treatment. The appendix that expands Forges’ hierarchy to eleven notions is useful housekeeping and correctly places their new separation. Citations are honest: they locate the result against La Mura, Pappa, Abbott, Auletta et al., and the classical literature.\n\nSoft spots are real but secondary. The games are engineered; W is a designed fine rather than a natural economic feature, and free parameter Λ must be large. The numerical evidence is only for three small games and a limited sample of Λ. None of this breaks the existence claim or the separation Corr ⊊ BI ∩ LO. The stress-test note is right that the reader’s “weakest assumption” is already discharged by the proofs.\n\nThis is for people who care about quantum foundations, algorithmic game theory, or multi-agent systems that use correlated advice. It is solid enough for a serious referee and worth citing if you work on quantum resources beyond entanglement or on the taxonomy of Bayesian equilibria. I would engage with it.","headline":"First clean separation of quantum advantage in Bayesian games from entanglement, via separable cq-states and a mechanism-design construction that also splits local communication equilibria from classical correlated ones.","tokens_in":23703,"tokens_out":504,"would_cite":true,"duration_ms":11979,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["03.65.Ud","03.67.Mn","02.50.Le"],"model":"grok-4.5","headline":"Separable quantum states can produce Bayesian-game equilibria better than every classical correlation.","keywords":["Bayesian games","quantum correlated equilibrium","separable states","pseudo-telepathy","social welfare","no-signalling","CHSH","magic square"],"falsifier":"For the concrete CHSH- or magic-square-based games, solve the linear program over classically correlated equilibria for large Λ and check whether any equilibrium still exceeds social welfare β(G)+ε.","tokens_in":23877,"feed_emoji":"⚛️","tokens_out":583,"duration_ms":6183,"temperature":0.7,"pith_summary":"The paper shows that quantum advantage in competitive Bayesian games does not require entanglement. Starting from any nonlocal game that admits a perfect quantum (or no-signalling) strategy, the authors build a related Bayesian game whose players receive a fully separable advice state. That state, together with the original measurements, forms a quantum correlated equilibrium whose average payoff equals 1, while every classically correlated equilibrium is forced down near the classical bound of the original game once a penalty parameter is large enough. The construction works because a carefully chosen zero-sum guessing game keeps each player’s type distribution honest; the separable state can still encode the nonlocal winning strategy without leaking the type information that classical shared randomness inevitably reveals. The result therefore places quantum advantage within reach of experimentally simpler resources and simultaneously separates several previously conflated notions of correlated equilibrium.","feed_headline":"Separable states beat every classical correlation in Bayesian games","feed_subtitle":"Quantum advantage no longer needs entanglement; a simple cq-state already lifts social welfare.","key_machinery":"The modified Bayesian game eG_Λ whose payoffs combine the original nonlocal winning predicate V with a scaled zero-sum guessing matrix W; the separable cq-state ω that records Alice’s local measurement outcomes while leaving Bob the post-measurement assemblage.","core_discovery":"For any quantum pseudo-telepathy game G there exists a Bayesian game eG_Λ and a fully separable advice state ω such that (ω together with Bob’s original measurements) is a quantum correlated equilibrium of social welfare 1, while every classically correlated equilibrium has social welfare at most β(G)+ε once Λ is larger than an explicit threshold.","pith_inferences":[],"forward_implications":[],"fun_headline_variants":["Separable states beat every classical correlation in Bayesian games","Unentangled quantum advice tops all classical equilibria in Bayesian games","Separable states create superior Bayesian equilibria beyond classical reach","Quantum game advantage without entanglement via separable states","Bayesian games: separable quantum states lift social welfare past classical"],"cache_read_input_tokens":15360,"weakest_assumption_plain":"The zero-sum guessing matrix must force every classical equilibrium to keep Alice’s type distribution extremely close to the prior; if a classical strategy could keep the guessing payoff near zero while still correlating the hidden variable with her type, the social-welfare gap would disappear.","fun_headline_variants_meta":{"raw":{"variants":["Separable states beat every classical correlation in Bayesian games","Unentangled quantum advice tops all classical equilibria in Bayesian games","Separable states create superior Bayesian equilibria beyond classical reach","Quantum game advantage without entanglement via separable states","Bayesian games: separable quantum states lift social welfare past classical"]},"model":"grok-4.5","effort":"low","cost_usd":0.004132,"raw_usage":{"total_tokens":1158,"prompt_tokens":656,"num_sources_used":0,"completion_tokens":80,"cost_in_usd_ticks":41320000,"prompt_tokens_details":{"text_tokens":656,"audio_tokens":0,"image_tokens":0,"cached_tokens":128},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":422,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":656,"tokens_out":80,"duration_ms":5365,"temperature":1.0,"reasoning_tokens":422,"cache_read_input_tokens":128,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-13T02:43:50.230504+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"For the concrete CHSH- or magic-square-based games, solve the linear program over classically correlated equilibria for large Λ and check whether any equilibrium still exceeds social welfare β(G)+ε.","supporting_citations":[],"review_version":1}