{"id":"026a9cbd-1458-4104-9a4c-5e04deccc910","arxiv_id":"2412.15615","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The advantage of a state and an incompatible measurement set in subchannel discrimination games equals (1 + robustness of state)(1 + robustness of measurement set), and similarly for weight in exclusion games.","lead":"This paper introduces new guessing games where a player uses both a quantum state and a set of measurements to win. The authors prove the maximum advantage in such games equals a simple product of two resource measures, one for the state and one for the measurement set.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Result 1's equality is a supremum, not an attained maximum: the J→∞ limit in Appendix B falls outside the finite-outcome game definition, so Eq. (8) should be stated with a supremum.","rationale":"The reader's ACCEPT is well-founded: the proofs are detailed, the conic-programming duals are standard, and the constructibility checks (CP, trace-nonincreasing, free-set closure) are explicit. I agree with the reader that the free-set closure is the key hypothesis and that the GPT positivity caveat is acknowledged. My partial disagreement is with the weight given to the J→∞ issue: I see it as a genuine gap in the statement of Result 1, not just a notational convenience, because Section III defines games with finite outcome sets and Eq. (8) claims a 'max'. The proof establishes a supremum. This is easily repairable and does not affect the paper's central message—indeed the same 'max/sup' convention appears in the cited single-object results—so the verdict remains ACCEPT. The proposed test is a symbolic recalculation of the finite-J ratio and a check for an alternative finite saturating construction.","tokens_in":30553,"tokens_out":23175,"duration_ms":201767,"concrete_test":"Compute the finite-J ratio from Eqs. (B5)–(B11): show it equals [1+R_F(ρ)][1+R_F(MA|X)]/(1+1/(αJ)), strictly below the RHS for all finite J. Then check whether any finite-outcome construction in the same framework can saturate Eq. (8); if none can, replace 'max' by 'supremum' in Result 1 and verify that the 'any resourceful pair is useful' conclusion still follows from the finite-J lower bound. This is a one-page symbolic calculation, not a numerical search.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The core concern is that Eq. (8) in Result 1 is not attained by any finite game. In Appendix B, the constructed instrument Ψ^{(ρ,MA|X,pX,J)} has l+J outcomes, and the fully-free-player success probability is bounded above by α + 1/J, while the resourceful player achieves at least α[1+R_F(ρ)][1+R_F(MA|X)]. The ratio is therefore [1+R_F(ρ)][1+R_F(MA|X)]/(1+1/(αJ)), which is strictly less than the RHS for every finite J. Taking J→∞ is not a legitimate game under Section III's definition (B = {1,...,m} finite), so the 'max' over games in Eq. (8) is a supremum. This does not destroy the qualitative claim that any resourceful pair is useful (choosing J large enough gives a finite ratio > 1), but it means the exact multiplicative quantification is only approached, not achieved. A second fragile point is the closure of the free POVM-set family under classical pre/post-processing, used in the achievability step at Eq. (B3) to ensure the coarse-grained \\tilde{N} is free; without it the bound ΣTr[\\tilde{N}Z^M]≤1 can fail. This is explicit in the theorem's hypothesis and holds for measurement incompatibility, so it is a stated assumption rather than a hidden flaw.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript introduces multi-object subchannel discrimination and exclusion games with prior information, in which a player simultaneously uses a quantum state and a POVM set. The central claim is that the maximal over-games ratio of the success probability against the best fully free pair equals [1+R_F(ρ)][1+R_F(M_{A|X})] for discrimination games (Result 1, Eq. (8)), and that the minimal over-games ratio of the error probability equals [1-W_F(ρ)][1-W_F(M_{A|X})] for exclusion games (Result 2, Eq. (17)). These results are stated for arbitrary state resources and arbitrary POVM-set resources closed under classical pre- and post-processing, with measurement incompatibility as a special case, and Result 3 extends both statements to general probabilistic theories. The proofs rely on conic-programming duality, with upper bounds derived from the primal problems and achievability from explicit games built from dual optimal witnesses.","tokens_in":30804,"tokens_out":13741,"duration_ms":129323,"significance":"If the stated equalities hold, the paper provides a clean multiplicative operational interpretation of generalized robustness and weight for pairs consisting of a state and a measurement set, unifying and generalizing the single-object results of Refs. [41-43] and the state-measurement pair results of Ref. [47]. The extension to GPTs, modulo the acknowledged restriction to positive maps, broadens the scope and is likely to be of interest to the resource-theory community. The appendix proofs are detailed and largely checkable, with explicit dual witnesses and game constructions, and the results include the qualitative consequence that every resourceful pair is useful in some finite discrimination or exclusion game. The technical corrections identified below concern the precise formulation of the discrimination result, not the main proof strategy.","major_comments":[{"comment":"The equality in Eq. (8) is stated as a maximum over games, but the proof only establishes a supremum. The constructed instrument Ψ^{(ρ,M_{A|X},p_X,J)}_{B|Y} has l+J outcomes, and for every finite J the derivation gives a free-player success probability ≤ α + 1/J (Appendix B, around Eqs. (B6)-(B9)) and a resourceful-player success probability ≥ α[1+R_F(ρ)][1+R_F(M_{A|X})] (Eq. (B10)). The resulting finite-J ratio is therefore strictly below the right-hand side by the factor (1+1/(αJ))^{-1}. The proof then takes J→∞, which is outside the class of games defined in Section III because the outcome set B={1,...,m} is required to be finite. Eq. (8) should be restated with a supremum over finite games, or the game definition should be extended to countably infinite outcome sets with a precise limiting argument. The same issue affects the discrimination part of Result 3, Eq. (25), and the corresponding proof in Appendix D. The qualitative claim that a resourceful pair is useful survives, since for any fixed resourceful pair the ratio exceeds 1 for sufficiently large finite J, but the exact multiplicative quantification is only approached, not attained.","section":"Section IV (Result 1, Eq. (8)) and Appendix B; also Result 3, Eq. (25) and Appendix D"}],"minor_comments":[{"comment":"The notation for the coarse-graining operation is confusing: the definition uses parameters K and N, while the application requires coarse-graining a POVM with l+J outcomes to one with l outcomes. Please rewrite the definition directly in terms of l and J to match the use in the proof.","section":"Appendix B, Eq. (B3)"},{"comment":"The upper-bound chain contains malformed expressions, in particular the term 'Ma|x \\Ñ^*_{a|x}' and the subsequent 'max_{≈} \\Ñ_{B|Y} ≺ \\Ñ^*_{A|X}'. The intended step is to replace M_{a|x} by the free POVM set N^*_{A|X} from the robustness decomposition; please rewrite this chain with unambiguous notation.","section":"Appendix B, Eq. (B2)"},{"comment":"The formal statements of Results 1 and 2 (and of Result 3 for the GPT case) do not include the hypothesis that the free family of POVM sets is closed under classical pre- and post-processing. This condition appears in the text before Eq. (7) and is used essentially in the achievability proofs (Appendix B, Eq. (B3); Appendix C, Eq. (C6)). Please add it explicitly to the theorem statements.","section":"Results 1-3 statements"},{"comment":"There is a typo 'Slatter's condition' that should read 'Slater's condition', and 'refereed' in Section II should be 'referred'.","section":"Appendix A and Section II"},{"comment":"The GPT formulation uses positive maps rather than completely positive maps as (sub)channels. The text acknowledges this caveat, but the statement of Result 3 should repeat it, since the quantum Results 1 and 2 are formulated with CP instruments.","section":"Section V and Result 3"},{"comment":"The summation in Eq. (3) includes an index µ that is not defined or used in the rest of the expression; the corresponding sum in Appendix A, Eq. (A1), correctly uses z. Please remove the stray index or define it.","section":"Eq. (3)"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is technically sound in its main ideas, and the max-versus-supremum issue is local and fixable by restating the discrimination results with a supremum or by extending the game definition to infinite outcome sets. The notation in the Appendix B upper-bound proof should also be cleaned up. I see no grounds for concern about novelty or overlap beyond the cited literature."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Punchline: the paper is a correct, honest extension of single-object resource discrimination/exclusion games to POVM sets, and the main theorem is a supremum rather than an attained maximum. The math checks out up to that caveat.\n\nWhat's new: the multi-object games where a player uses a state plus a POVM set, with simulability as the free operation, and the product formulae [1+R_F(rho)][1+R_F(M)] for discrimination and [1-W_F(rho)][1-W_F(M)] for exclusion. The GPT extension is a nice bonus. The proof uses standard conic duality and the recovery of [41-43] and [47] is worked out explicitly.\n\nSoft spots: (1) The J->infinity limit in Appendix B: the game has l+J outcomes for finite J, but the limiting game is not finite and hence not a game under the definition in Section III. Eq. (8) should be stated with a sup. This is a minor mathematical imprecision, not a fatal flaw. (2) The GPT part uses positive maps rather than completely positive ones. The authors flag this, and it is a known boundary issue in GPTs; the result stands in their framework. (3) The closure of the free POVM-set family under classical pre/post-processing is an explicit hypothesis, used at Eq. (B3). For measurement incompatibility it holds; for other resources it is an assumption, but the paper says so.\n\nRecommendation: worth peer review. A serious referee should check the dual programs and the achievability calculations, but the structure is solid. The authors should patch the max-to-sup issue and add a sentence clarifying the finite-J approximation. I would probably cite this in a resource theory paper.","headline":"Solid extension of single-object resource games to POVM-set pairs; main caveat is that Eq. (8) is a supremum, not an attained max.","tokens_in":31370,"tokens_out":2452,"would_cite":true,"duration_ms":21355,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81P15"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that a player using both a quantum state and a set of measurements obtains an advantage in discrimination and exclusion games that factorises exactly into the resource quantifiers of the two objects.","keywords":["measurement incompatibility","multi-object operational tasks","subchannel discrimination","exclusion games","generalised robustness","weight of resource","general probabilistic theories","joint measurability"],"falsifier":"Compute the left and right sides of Eq. (8) for a specific pair $(\\rho, M_{A|X})$ and a finite set of subchannel games by semidefinite programming: if any game yields a ratio larger than $[1+R_F(\\rho)][1+R_F(M_{A|X})]$, the upper-bound theorem is false. The same check can be run with a free set that is not closed under simulability to test whether the closure assumption is genuinely required.","tokens_in":1661,"feed_emoji":"⚛️","tokens_out":3455,"duration_ms":73800,"temperature":0.7,"pith_summary":"The paper claims that the usefulness of a quantum state and a set of measurements used together can be measured exactly. It introduces multi-object subchannel discrimination and exclusion games with prior information, in which one player controls both a state and a POVM set. The central result is that the best possible advantage over all fully free pairs factorises: in discrimination games the maximum success-probability ratio equals the product of the generalised robustnesses of the two objects, while in exclusion games the minimum error-probability ratio equals the product of their weights of resource. Because measurement incompatibility is a resource of POVM sets closed under classical pre- and post-processing, the result applies directly to incompatible measurements: every incompatible set, together with any state, becomes useful for some such game. The same statements hold in general probabilistic theories, so the quantification is not an artifact of the quantum formalism.","feed_headline":"Incompatible measurements' power is a product rule","feed_subtitle":"Pair a resourceful quantum state with an incompatible measurement set and the advantage multiplies—exactly as robustness predicts.","key_machinery":"The carrying object is the multi-object quantum subchannel discrimination and exclusion game with prior information, in which the referee applies an instrument to the player's state, sends the post-measurement state and the instrument label to the player, and the player then uses a POVM set with classical pre- and post-processing to guess the subchannel outcome (discrimination) or to avoid it (exclusion). The proof machinery is conic-programming duality: the generalised robustness and the weight of resource admit dual characterisations as optimisations over positive operators, and the achievability arguments use the optimal dual witnesses to define instrument sets that fully free players cannot simulate well. Closure of the free POVM-set family under simulability is what lets coarse-grained versions of free POVM sets remain free, a step essential to the upper bounds.","core_discovery":"Result 1 states that for any state $\\rho$ and POVM set $M_{A|X}$, the maximum over all subchannel discrimination games with prior information of the ratio between the success probability of the player using $(\\rho, M_{A|X})$ and the best fully free pair equals $[1+R_F(\\rho)][1+R_F(M_{A|X})]$. Result 2 states the exclusion analogue: the minimum error-probability ratio equals $[1-W_F(\\rho)][1-W_F(M_{A|X})]$, with the weight of resource replacing the generalised robustness. Result 3 extends both formulas to general probabilistic theories, with states, measurement sets, instruments, and resource quantifiers rephrased in GPT terms. The proofs construct explicit games that saturate the bounds: for discrimination, dual feasible operators $Z_\\rho$ and $\\{Z_{a|x}\\}$ are used to build an instrument set with $J$ extra subchannels and the limit $J\\to\\infty$ is taken; for exclusion, operators $Y_\\rho$ and $\\{Y_{a|x}\\}$ build a game with one extra subchannel. The free set of POVM sets is assumed closed under classical pre- and post-processing, so coarse-graining a free POVM set yields a free POVM set; under this closure, the resource quantifiers coincide exactly with the operational advantage.","pith_inferences":["Because the two factors multiply, the pair's advantage decomposes into independent contributions from each object; a natural composite monotone would be $R_F(\\rho)+R_F(M_{A|X})+R_F(\\rho)R_F(M_{A|X})$.","The discrimination proof requires infinitely many extra subchannels ($J\\to\\infty$) while the exclusion proof needs only one; a finite-$J$ analysis would give explicit rates of approach and potentially experimentally accessible approximate versions of the result.","Interpreting these games as betting or expected-utility tasks could extend the multiplicative robustness and weight formulas to decision-theoretic settings, linking resource quantifiers to utility.","Testing the bound in a measurement-set resource theory that is not closed under classical post-processing, such as POVMs with a fixed noise parameter that coarse-graining can reduce, would pinpoint exactly where the closure assumption binds."],"forward_implications":["Every partially or fully resourceful pair wins some subchannel discrimination game with a ratio exactly $[1+R_F(\\rho)][1+R_F(M_{A|X})]$ over all fully free pairs.","If only the measurement set is resourceful, the advantage reduces to $[1+R_F(M_{A|X})]$, recovering the earlier single-object incompatibility bound.","In exclusion games, the same pair beats all fully free pairs by the factor $[1-W_F(\\rho)][1-W_F(M_{A|X})]$ in error probability.","The formulas hold for any state resource and any measurement-set resource closed under classical pre- and post-processing, so incompatibility is one instance among many.","The same multiplicative quantification holds in any general probabilistic theory satisfying the no-restriction hypothesis, making the result theory-independent."],"supporting_citations":[{"why":"Establishes the single-object result for measurement incompatibility in state discrimination that Result 1 recovers when the state is free.","marker":"[41]"},{"why":"Provides a concurrent single-object characterisation of incompatibility advantage that Result 1 generalises to pairs.","marker":"[42]"},{"why":"Gives the ensemble form of the single-object discrimination bound that appears as a special case of Eq. (8).","marker":"[43]"},{"why":"Introduces multi-object tasks for a state and a single measurement; Result 1 reduces to its Eq. (12) when the POVM set is trivial.","marker":"[47]"},{"why":"Defines simulability of POVM sets, the classical pre- and post-processing closure used by the free-set assumption and the proofs.","marker":"[58]"},{"why":"Introduces quantum state discrimination with prior information, the template for the subchannel games studied here.","marker":"[61]"},{"why":"Provides the single-object GPT robustness characterisation that Result 3 extends to the multi-object regime.","marker":"[67]"}],"fun_headline_variants":["State and measurement resources multiply","Multi-object games show resource multiplication","Incompatibility plus state: product advantage","Robustness and weight: product formulas"],"cache_read_input_tokens":33536,"weakest_assumption_plain":"The free family of POVM sets must be closed under classical pre- and post-processing, so that any classical relabelling or merging of outcomes of a free measurement set is again free; the GPT extension additionally works with positive maps rather than completely positive ones.","fun_headline_variants_meta":{"raw":{"variants":["State and measurement resources multiply","Multi-object games show resource multiplication","Incompatibility plus state: product advantage","Robustness and weight: product formulas"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000246,"raw_usage":{"total_tokens":1562,"prompt_tokens":993,"completion_tokens":569,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":609,"completion_tokens_details":{"reasoning_tokens":519}},"tokens_in":609,"tokens_out":569,"duration_ms":5672,"temperature":1.0,"reasoning_tokens":519,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T11:14:50.617562+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the left and right sides of Eq. (8) for a specific pair $(\\rho, M_{A|X})$ and a finite set of subchannel games by semidefinite programming: if any game yields a ratio larger than $[1+R_F(\\rho)][1+R_F(M_{A|X})]$, the upper-bound theorem is false. The same check can be run with a free set that is not closed under simulability to test whether the closure assumption is genuinely required.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the ensemble form of the single-object discrimination bound that appears as a special case of Eq. (8)."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces multi-object tasks for a state and a single measurement; Result 1 reduces to its Eq. (12) when the POVM set is trivial."},{"cited_title":"Guerini, J","cited_arxiv_id":null,"evidence_quote":"Defines simulability of POVM sets, the classical pre- and post-processing closure used by the free-set assumption and the proofs."},{"cited_title":"Carmeli, T","cited_arxiv_id":null,"evidence_quote":"Introduces quantum state discrimination with prior information, the template for the subchannel games studied here."}],"review_version":1}