{"id":"22f3ecb0-2269-4900-be90-763b52756620","arxiv_id":"2608.00124","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For faithful local deterministic models of n-partite GHZ–Mermin correlations, the exact minimum surrendered measurement independence for n=3–13 is R/[2(R+1)] with R=2^floor((n-1)/2), conjectured to hold for all n.","lead":"This paper works out exactly how much 'measurement dependence' — hidden-variable dependence on measurement choices — is needed for a classical local model to perfectly fake the correlations of many-party GHZ-Mermin quantum states. The minimum follows a simple formula, never exceeds one half, and is confirmed exactly for up to 13 parties.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Exactness for n≥5 depends on an unstated machine-verified Mermin-operator identity and code-checked frustration-ansatz overlaps; the hand proofs stop at n=4.","rationale":"The reader's weakest assumption—that exactness for n≥5 hangs on a machine-verified Mermin-operator identity that is neither stated nor human-proven—is the same load-bearing concern I identify. I also stress that the frustration-ansatz overlap computation is a second code-dependent ingredient, but that falls under the same category. The paper's internal mathematics is largely sound: Theorem 1 and Theorem 2 are carefully argued, Theorem 6's proof is valid and even elegant, the n=3 and n=4 hand proofs are concrete, and the ceiling theorem is straightforward. The problem is that the exact staircase through n=13 is not self-contained: it relies on external code for both sides of the squeeze. Since the paper explicitly says code is in the repository, this is not a rejection; but without the identity being stated or the code independently reproduced, a conditional verdict is appropriate. I see no basis to move to ACCEPT or REJECT. The reader's CONDITIONAL verdict should stand unchanged.","tokens_in":12998,"tokens_out":16881,"duration_ms":189692,"concrete_test":"Independently implement an exact integer-arithmetic verification (without reusing the deposited scripts) that for each n=3–13: (a) enumerates all 2^{n+1} classes, (b) computes m(S′) for the full Mermin set and, when n is even, for an embedded Mermin (n−1)-subset, (c) constructs the frustration-ansatz densities from the ground states of H, and (d) computes the maximum pairwise overlap. For n≥11 use bitset representations to keep this feasible. Check equality with R/[2(R+1)] and F_min·s=1/4. Any deviation invalidates or shifts the claimed exact value; a full pass confirms the machine-certified part of the paper.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The lower-bound half of the squeeze (Theorem 6) is proven, but for n=5–13 it is only evaluated using maximum satisfiability counts m(S′) that the text says follow from an 'exhaustively machine-verified Mermin-operator identity' (Sec. V, Sec. XI). That identity is never stated or proved in human-readable form. The upper-bound half uses the frustration ansatz of Sec. VIII, whose pairwise overlaps are said to be 'computed by pure integer counting'; no closed formula or proof for these overlaps appears for general n. Thus the exactness of F_min = R/[2(R+1)] for every n≥5 rests on two external computational claims: the m-counts entering the lower bound and the ansatz-overlap values entering the upper bound. If either is wrong, the claimed coincidence can fail and the exact floors could shift. This is not an accusation of error; it is a precise missing link. The reduction theorem, ceiling theorem, lower-bound theorem, and hand proofs at n=3,4 are not affected, but the central exactness claim is, because an unverified piece of software is the only bridge between a general bound and a claimed exact value.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the minimum fraction of measurement independence that must be surrendered by a local deterministic model reproducing the n-partite GHZ–Mermin correlations, under the additional faithfulness requirement that all proper-subset marginals vanish. The main results are: a reduction theorem (Theorem 1) showing that the faithfulness constraints are cost-free, a lossless symmetrization (Theorem 2), hand proofs for the tripartite and four-partite floors F(3)=F(4)=1/3 (Theorems 3 and 4), a universal ceiling F_min(n)≤1/2 (Theorem 5), and a general overlap lower bound (Theorem 6). The paper claims exact values F_min(n)=R/[2(R+1)], R=2^{floor((n-1)/2)}, for every n=3,...,13, certified by an integer-arithmetic squeeze between the Theorem 6 lower bound and an explicit frustration-ansatz construction, and it gives a physical characterization of the optimal densities as uniform measures on contextual ground states of the classicalized stabilizer Hamiltonian.","tokens_in":13324,"tokens_out":10641,"duration_ms":112345,"significance":"If the exactness claims hold, this is a substantial extension of Hall's measurement-dependence pricing to faithful multipartite GHZ–Mermin correlations. The paper's rigorous contributions are real: the reduction theorem is elegant and practically important, the universal ceiling is a clean general result, Theorem 6 is a nontrivial general lower bound, and the hand proofs at n=3 and n=4 are complete. The exact integer-arithmetic certification philosophy and the availability of code are also strengths. The central caveat is that the exact values for n≥5 are not established in the text itself: they depend on two external computational claims—the satisfiability counts entering Theorem 6 and the frustration-ansatz overlap values—neither of which is stated in human-readable form.","major_comments":[{"comment":"The exact values for n≥5 are not verifiable from the manuscript. The lower bound of Theorem 6 is evaluated using satisfiability counts m(S') that are said to follow from 'an exhaustively machine-verified Mermin-operator identity' (Sec. V), and Sec. XI repeats that the counts follow from this identity, but the identity is never stated. For odd n the required count is m=2^{n-1}(R+1)/(2R); for even n the text invokes an embedded Mermin (n−1) sub-scenario without stating the corresponding construction. Without a statement of the identity and either a human proof or a formal machine-checkable certificate, the lower half of the squeeze cannot be checked. This is load-bearing: if m is wrong, the claimed equality F_min=R/[2(R+1)] can fail.","section":"Secs. V and XI; Conjecture 1"},{"comment":"The upper-bound half of the squeeze, the frustration ansatz, is asserted to evaluate to R/[2(R+1)] and to be optimal for every computed n, with pairwise overlaps 'computed by pure integer counting' (Sec. XI). No closed-form expression or proof for these overlaps is given for general n. A referee cannot verify the upper bound from the manuscript. Please provide a general derivation of the ground-state counts and intersection counts entering the ansatz evaluation, or state precisely which steps are machine-checked and how the deposited code certifies them.","section":"Sec. VIII A/B and Sec. XI"},{"comment":"The claim that 'complete proofs of all theorems are given in the main text and appendices' is misleading for the central exactness result. Theorems 1, 2, 5, and 6 are proved in the text, and Theorems 3 and 4 are hand-proved, but the values for n=5,...,13 are not theorems in the text: they are computational results whose certification depends on the unstated identity and on unstated overlap computations. The paper should state, as a theorem or a precisely described computational proposition, what is proved and what is certified by code, and should make the machine-verifiable artifact explicit and self-contained.","section":"Introduction and Sec. V"}],"minor_comments":[{"comment":"The quoted symmetry-reduced variable count for n=9, '3,673', does not match the orbit count implied by Theorem 2, which gives 2∑_{k even}(k+1)(n−k+1)=220 variables for n=9. Please clarify what is being counted.","section":"Sec. IV"},{"comment":"The CHSH row lacks entries for s and F_min·s. Consider using explicit em dashes to avoid implying missing values.","section":"Table I"},{"comment":"The statement '0 of 120 orderings at n=4, exhaustive' is cryptic; please define the ordering test or provide the relevant script reference.","section":"Sec. VIII E"}],"recommendation":"major_revision","confidential_remarks":"I did not run the deposited code. My assessment is based on the manuscript text; the missing Mermin-operator identity and the unstated ansatz-overlap computations are the only obstructions to verifying the central exactness claim. If the author can supply a human-readable proof or a formal certificate for these two computational ingredients, the result would be a strong and publishable contribution. The hand-provable results and the general lower-bound theorem are sound and valuable on their own."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper solves the faithful Mermin problem Hall left open, and the headline numbers are probably correct. The reduction theorem showing the faithfulness constraints are free is clean and useful, and it explains why the tripartite faithful optimum coincides with Hall's correlator-only 1/3. The universal ceiling F ≤ 1/2 is a neat argument, and Theorem 6 is a valid general lower bound. The hand proofs at n=3 and n=4 are complete, and the out-of-sample cluster-state checks give the results a feel of being robust rather than tuned.\n\nThe soft spot is exactly where you'd expect it: everything for n≥5 rests on a squeeze between a proven lower bound and an explicit construction, but the two sides of that squeeze are evaluated using computational data that the paper does not expose. The lower bound uses maximum-satisfiability counts m(S′) that are attributed to an 'exhaustively machine-verified Mermin-operator identity' — never stated, never proved. The upper bound uses the frustration-ansatz construction, but the pairwise overlaps are said to be 'computed by pure integer counting' with no closed formula shown. So the claimed exact values for n=5..13 are only as good as the deposited code, which I could not fully verify. This is a missing link, not evidence of error. The coincidence of lower and upper bounds at all those points is persuasive, and the pattern is simple enough that it would be surprising if the code were wrong. But a referee cannot certify this without either a human-readable statement of the identity or a fully auditable certificate trail.\n\nMinor notes: some references are dated 2026 and may be hard to verify, and the discussion of the landscape being 'glassy' is more suggestive than foundational. Neither affects the main result.\n\nThe paper is written clearly and the claims are stated precisely. The hand-provable parts are solid, and the computational parts are probably correct. It deserves a serious referee — but referees should be instructed to demand either the explicit satisfiability identity or a reproducible certification script before accepting the exact n≥5 values as established.","headline":"A genuinely new solution to Hall's faithful Mermin problem, with exact floors through n=13 that are probably right, but whose n≥5 certification rests on a machine-checked identity the paper never states.","tokens_in":13727,"tokens_out":1573,"would_cite":true,"duration_ms":21705,"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":"Classical models that faithfully reproduce n-party Mermin-GHZ correlations must surrender exactly F_min(n)=R/[2(R+1)] of measurement independence for n=3 through 13, and never more than half at any size.","keywords":["measurement independence","Bell inequalities","Mermin correlations","GHZ state","local deterministic models","hidden variables","stabilizer Hamiltonian","device-independent certification"],"falsifier":"Independently recompute the maximum satisfiability of the full Mermin constraint set at n=13, which the table's s values imply is 2080; if a brute-force or independent SAT-based count differs, the proven lower bound separates from the frustration ansatz and the exact staircase fails. A cheaper check is the predicted n=14 and n=15 floors, which the law fixes at 32/65 and 64/129.","tokens_in":12896,"feed_emoji":"🎲","tokens_out":9922,"duration_ms":104131,"temperature":0.7,"pith_summary":"This paper asks how much of the free-choice assumption must be given up by any classical local model that reproduces the Mermin-GHZ correlations of n qubits exactly—every full correlator correct and every smaller marginal zero. It answers that the minimum surrendered fraction follows the closed staircase F_min(n)=R/[2(R+1)] with R=2^floor((n-1)/2), certified for n=3 through 13: 1/3, 2/5, 4/9, 8/17, 16/33, 32/65, each value repeated at the following even size. The result is a squeeze between a proven combinatorial lower bound and an explicit construction whose optimal densities are uniform on the ground states of a frustrated Hamiltonian counting violated Mermin constraints—the classical shadow of the GHZ state's stabilizer. A universal ceiling F≤1/2 shows faithful reproduction never requires total measurement dependence, however many parties are involved. A sympathetic reader would care because these floors put an exact, finite price on measurement dependence as the Bell exit, and set the settings-dependence an adversary needs to counterfeit multipartite device-independent certificates.","feed_headline":"Measurement-independence cost is exactly R/(2R+2) for Mermin states","feed_subtitle":"Faithful local models must spend that fraction of settings freedom; the ceiling stays at half for every n.","key_machinery":"The argument runs on four connected pieces. (1) The parity-class reduction (Theorem 1) maps each deterministic strategy to a pair (P,γ), collapsing the variable space to 2^{n+1} classes and deleting the marginal constraints without changing the optimum. (2) The overlap lower bound (Theorem 6) states that for any subset S' of Mermin constraints whose maximum satisfiable count is m, F≥(|S'|-m)/(|S'|-1). (3) The frustration ansatz supplies matching upper bounds: each setting's hidden-variable density is uniform over the ground states of the Hamiltonian that counts violated Mermin constraints. (4) Fourier analysis over GF(2)^{n+1} shows this landscape cannot be generated by pairwise interactions","core_discovery":"The central claim is that the exact minimum measurement dependence for faithful local deterministic models of n-partite Mermin-GHZ correlations is F_min(n)=R/[2(R+1)], R=2^floor((n-1)/2), for every n=3,...,13. Theorem 1 collapses the linear program to 2^{n+1} parity classes and proves the faithfulness constraints are free; Theorem 6 provides a general lower bound F≥(|S'|-m)/(|S'|-1) for any constraint subset S' with maximum satisfiability m; and the frustration ansatz—placing each setting's density uniformly on the least-frustrated classes of the violation-counting Hamiltonian—meets that bound on every computed scenario. The exact values are established by integer-arithmetic squeeze, indepen","pith_inferences":["Editorial extension: if the staircase law continues beyond 13 parties, the exact floors saturate toward the ceiling 1/2 without reaching it; the next predicted values are 32/65 at n=14 and 64/129 at n=15, both directly computable with the paper's symmetry-reduced program.","Editorial extension: the pairwise no-go implies that any future 'lawful' measurement-dependent mechanism reproducing these correlations must depend on settings collectively (three or more parties at once), since no two-body interaction can generate the frustration landscape; a concrete search for such mechanisms should target collective rather than pairwise dependence.","Editorial extension: the out-of-sample cluster-state cores suggest a general inheritance principle—a scenario's floor is set by its tightest embedded irreducible all-versus-nothing sub-scenario. Computing the floor for additional graph-state cores, e.g. a seven-qubit ring cluster, would test whether the F·s=1/4 invariant survives outside the families studied."],"forward_implications":["The reduction theorem makes faithfulness free: matching the Mermin correlators is enough, and the vanishing-marginal constraints can be imposed by a fiber lift without raising the minimum measurement dependence.","The universal ceiling F≤1/2 means no faithful local deterministic model of Mermin-GHZ correlations, at any number of parties, is forced to the conspiratorial limit of complete measurement dependence.","The exact floors double as counterfeiting thresholds: a classical adversary of a multipartite device-independent certificate needs at most the listed fraction of settings dependence to reproduce the Mermin violation, and at most half at any size.","The frustration ansatz characterizes the optimal hidden-variable densities explicitly as uniform measures on the ground states of the classicalized stabilizer Hamiltonian, and the Fourier analysis shows the landscape cannot be generated by pairwise interactions.","The closed staircase law is certified exactly on all computed points and, together with the ceiling, fixes the asymptotic price of measurement dependence for Mermin statistics at 1/2 from below."],"fun_headline_variants":["Faithful local models pay exactly R/(2R+2) in settings freedom","Mermin correlations: minimal measurement independence is R/(2R+2)","Exact cost of measurement independence for Mermin states: R/(2R+2)","Measurement dependence floor for Mermin: R/(2R+2) exactly","Mermin states: measurement independence priced exactly at R/(2R+2)"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The exact values for n≥5 rest on a computer-checked formula for how many of the Mermin equations one hidden state can satisfy at once; that formula is not written out for humans to audit.","fun_headline_variants_meta":{"raw":{"variants":["Faithful local models pay exactly R/(2R+2) in settings freedom","Mermin correlations: minimal measurement independence is R/(2R+2)","Exact cost of measurement independence for Mermin states: R/(2R+2)","Measurement dependence floor for Mermin: R/(2R+2) exactly","Mermin states: measurement independence priced exactly at R/(2R+2)"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000869,"raw_usage":{"total_tokens":3707,"prompt_tokens":959,"completion_tokens":2748,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":703,"completion_tokens_details":{"reasoning_tokens":2641}},"tokens_in":703,"tokens_out":2748,"duration_ms":20925,"temperature":1.0,"reasoning_tokens":2641,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-04T01:13:09.140318+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Independently recompute the maximum satisfiability of the full Mermin constraint set at n=13, which the table's s values imply is 2080; if a brute-force or independent SAT-based count differs, the proven lower bound separates from the frustration ansatz and the exact staircase fails. A cheaper check is the predicted n=14 and n=15 floors, which the law fixes at 32/65 and 64/129.","supporting_citations":[],"review_version":1}