{"id":"8e1a0cce-019f-4580-84db-21ce005a6840","arxiv_id":"2501.14612","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper classifies exactly when the Spohn curve of a 2x2 game is reducible, proves real points are dense on it in those cases under genericity assumptions, and proposes a j-invariant based equivalence notion for games.","lead":"This paper studies the geometry of dependency equilibria in two-player two-choice games, where players maximize payoffs based on each other's expected choices. It classifies when the associated algebraic curve breaks into pieces and shows that real solutions are dense on it in many cases, then proposes using elliptic curve invariants to compare games.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central denseness claim is broader than Theorem 3.7 and depends on unreproduced primary decompositions; if the Mathrepo computations are incomplete or wrong, the classification and Theorem 3.7 both fail.","rationale":"The reader's conditional verdict is well calibrated. My stress-test did not find an internal inconsistency in the algebraic conclusions, but confirms that the two main theorems are not self-contained: the classification and the component decomposition are outsourced to a Mathrepo page, and the abstract states a stronger \"in all cases\" result than the theorem's hypotheses. The most fragile point is Theorem 3.7's lifting of real smooth points from C to V_X: it relies on the elimination ideals and codimension checks in [7], and on Proposition 3.6 applied to a specific coordinate projection; neither the computation nor the genericity of the projection is demonstrated in the text. If the decomposition omits a component or misidentifies a prime, the denseness claim for V_X could fail for that component. The proposed test—an independent symbolic primary decomposition in a fresh CAS session—would settle both the completeness of the twelve-case list and the existence of real smooth points. I therefore leave the verdict conditional; no rejection is warranted because the claims are plausible, concrete, and checkable.","tokens_in":22151,"tokens_out":18510,"duration_ms":177657,"concrete_test":"Independently recompute, in a fresh Macaulay2 or Singular session and without loading the authors' Mathrepo, the symbolic primary decompositions: (i) the ideal defining reducibility of the ternary cubic f (via the map from the space of (line, conic) pairs to the seven coefficients c1...c7, intersected with the payoff parameter map), and compare its variety with the union of the twelve cases in Theorem 3.2; (ii) the minimal primes of I(V_X) over Q(a_ij, b_ij), verify codimension 2 for each of the twelve cases and that the elimination ideal of each minimal prime agrees with a minimal prime of C (or is a line through [0:0:0:1], for case (7)). If the union from (i) matches and all primes from (ii) admit real smooth points under the generic-other-entries assumptions, the concern is settled; any discrepancy means the central claims need revision.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing step is the transition from the planar cubic C to the Spohn variety V_X in Theorem 3.7, together with the reducibility classification in Theorem 3.2. Both proofs are deferred: Theorem 3.2 says \"The detailed computations can be found in [7]\" and only sketches one ideal decomposition (J5), with J3 and J4 dismissed as \"similarly\" and the long factors d_i, e_i not displayed. Theorem 3.7 contains the sentence \"One can check in all cases that every irreducible component has codimension 2\" and the elimination ideals for the twelve cases are again only in [7]. A wrong primary decomposition would change which components exist and could destroy the smooth-real-point conclusion. Independently of the computation, the theorem's hypotheses are narrower than the abstract: Theorem 3.7 assumes \"all the other entries are generic\" and covers only the twelve reducibility cases, while the abstract claims real points are dense \"in all cases\". Example 3.8 exhibits additional reducible V_X cases outside the twelve, so the completeness of the case list is itself part of the claim. Finally, the proof invokes Proposition 3.6 for \"the generic projection between affine spaces\" but then uses the specific elimination projection forgetting p22; that this particular projection is generic enough is not shown. Any of these gaps, if real, undermines the headline denseness result.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the Spohn variety of totally mixed dependency equilibria for 2x2 normal-form games, which generically is an intersection of two quadrics in P3 and hence an elliptic curve. The authors eliminate one variable to obtain a plane cubic, classify when this cubic is reducible (Theorem 3.2, twelve explicit cases), and prove that under these reducibility conditions the real points of the Spohn variety are Zariski dense, assuming the remaining payoff entries are generic (Theorem 3.7). The paper also reviews algorithms for computing j-invariants of elliptic curves given as intersections of quadrics, applies them to Spohn curves, and proposes a notion of game equivalence based on the j-invariant. The proofs of the two main theorems rely heavily on computations in the companion Mathrepo [7], and the abstract states the denseness result more broadly than Theorem 3.7 establishes.","tokens_in":22399,"tokens_out":2764,"duration_ms":27897,"significance":"If the classification and denseness results are correct, they provide a useful and nontrivial contribution to the real algebraic geometry of dependency equilibria: the reducibility classification in Theorem 3.2 is explicit and combinatorial, and the real-point denseness statement is exactly the kind of result needed to justify passing from complex Spohn varieties to real dependency equilibria. The paper also gives a clear exposition of j-invariant computations for intersections of quadrics, with reproducible examples, and introduces an equivalence notion for 2x2 games that could stimulate further work. The authors are generally careful to distinguish generic from non-generic statements, and they provide concrete game-theoretic applications such as Proposition 2.7 and Remark 2.8. The main weakness is that the central proofs are not self-contained: they defer essential ideal decompositions, elimination computations, and component checks to the companion repository, and one headline claim in the abstract is not supported by the stated theorem.","major_comments":[{"comment":"The abstract states that real points are dense on the Spohn curve 'in all cases', but Theorem 3.7 proves this only for the twelve reducibility cases of Theorem 3.2, and only under the assumption that all other payoff entries are generic. Example 3.8 explicitly describes reducible Spohn varieties outside those twelve cases (e.g. a12=a22) and states that there are 'no other known cases' rather than proving completeness. The denseness result for all cases is therefore not established by the manuscript, and the abstract should be weakened to match Theorem 3.7.","section":"Abstract and Theorem 3.7"},{"comment":"Theorem 3.2 is the classification on which the denseness proof rests, but its proof is not self-contained. The text says 'The detailed computations can be found in [7]', the ideals J3 and J4 are dismissed with 'we can proceed similarly', and the long factors d_l and e_l in the displayed decompositions are not shown. Since a missed factor or an incorrect primary decomposition would change the list of reducibility cases, the proof as printed does not allow a reader to verify the classification independently. The authors should either reproduce the relevant decompositions (at least the minimal prime lists) or provide a certificate in the appendix that is checked in the text.","section":"Section 3.1, proof of Theorem 3.2"},{"comment":"The passage from real smooth points on the planar cubic C to real smooth points on the Spohn variety VX depends on unstated computational facts: 'One can check in all cases that every irreducible component has codimension 2' and the claim that for all cases except (7) the elimination ideals of the minimal primes are exactly the minimal primes of C. These checks are again deferred to [7]. Moreover, the proof invokes Proposition 3.6 for 'the generic projection between affine spaces' but then uses the specific elimination projection that forgets p22. It is not shown that this particular projection is generic enough for Proposition 3.6 to apply to each irreducible component. This is a load-bearing step in the denseness theorem and should be addressed explicitly.","section":"Section 3.2, proof of Theorem 3.7"}],"minor_comments":[{"comment":"In the irreducible case the proof states that 'the cubic is an elliptic curve and has real points'; an irreducible real plane cubic need not be smooth, so 'elliptic curve' should be replaced by 'irreducible cubic' unless smoothness is justified, although the existence of real points is not in question.","section":"Section 3.2, Lemma 3.5"},{"comment":"There is a typo in the footnote: 'Decompostion' should be 'Decomposition'.","section":"Section 3.1, footnote 1"},{"comment":"The code comments contain garbled spacing, e.g. 'Co nt in ue d fraction r e p r e s e n t a t i o n' and 'e v a l _ c o n t f r a c', which should be cleaned for readability.","section":"Appendix A, Listing 1"},{"comment":"The phrase 'contined fraction' should read 'continued fraction'.","section":"Appendix A, final paragraph"},{"comment":"The observation that affine transformations of the payoff matrices preserve the Spohn curve is useful, but it would be clearer to state explicitly that this is a symmetry of the defining quadrics and not only of the j-invariant.","section":"Section 4.2, Example 4.8"}],"recommendation":"major_revision","confidential_remarks":"The manuscript's reliance on the companion Mathrepo [7] is unusually heavy for the two central theorems. If the authors can provide, within the paper or as a clearly reviewed supplement, the actual decompositions and component checks, the results will be verifiable; in the current form the refereed claims are not independently checkable from the text. The abstract overstatement should also be corrected. The topic fits the journal well and the paper is readable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: the paper genuinely answers the reducibility question for Spohn cubics and pushes real-point denseness into non-generic territory, but the abstract overstates the denseness result and the main proofs lean on computations you can only find in the authors' Mathrepo. I'd send it to a referee, but it needs a major revision pass.\n\nWhat's new: Theorem 3.2's classification of when the planar Spohn cubic splits (12 cases, plus the f=0 list in Remark 3.1) is a real result and directly addresses [12, Problem 4.3]. Theorem 3.7, together with Example 3.8, extends the generic denseness result in [12] to a substantial set of non-generic 2x2 games. The self-contained proofs of Proposition 2.6 (Nash implies dependency for all 2x2 games) and Proposition 2.7 (Prisoner's dilemma has infinitely many Pareto-better dependency equilibria) are clean and useful. The j-invariant equivalence in Definition 4.11 is modest but perfectly sensible, and the worked examples in Section 4 do what they claim.\n\nThe soft spots are real, though. The abstract says real points are dense 'in all cases', but Theorem 3.7 proves this only for the twelve reducibility cases under a genericity assumption on the remaining payoff entries. Example 3.8 patches two more families, but the paper itself says no other cases are 'known' where V_X is reducible while C is not — that is not a proof of completeness. More seriously, the load-bearing algebra is deferred: the decompositions of the ideals J3, J4, the elimination ideals for the twelve cases, and the codimension checks are all in [7], with only one sample decomposition (J5) shown in the text. That makes the classification and the denseness proof impossible to check without going to the repository, and if a primary decomposition is wrong, both theorems fail. There is also a genuine gap in the proof of Theorem 3.7 where Proposition 3.6's generic projection is used, but the argument specifically projects away p22; that this particular projection is generic enough is not shown. Section 4 is mostly review, which is fine for a game-theory audience, but it is not new.\n\nWho is it for: people working on dependency equilibria, or on real algebraic geometry applied to game theory, will get value. The classification is the sort of thing that should be on record. It deserves a serious referee, not a desk reject. The referee should ask for the abstract to be restated to match the theorem, for the computational steps to be either reproduced in the paper or made available as a versioned, independently runnable artifact, and for the projection step to be justified. If those are fixed, I'd be comfortable citing Theorem 3.2.","headline":"The reducibility classification is new and valuable, but the abstract oversells the denseness result and the main proofs lean on unreproduced computations.","tokens_in":22988,"tokens_out":4857,"would_cite":false,"duration_ms":76155,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14H52","14P25","91A05"],"pacs":[],"model":"deepseek-v4-flash","headline":"The reducibility of the Spohn curve of a 2x2 game is fully classified: twelve conditions, and real points stay dense.","keywords":["Spohn curve","dependency equilibrium","totally mixed equilibria","2x2 normal-form games","elliptic curve","intersection of two quadrics","j-invariant","real algebraic geometry"],"falsifier":"Take a 2x2 payoff matrix that satisfies none of the twelve conditions in Theorem 3.2 and factor the corresponding ternary cubic; if it splits, the classification is incomplete. In one of the twelve cases, vary the payoff entries assumed generic and inspect the real points of each irreducible component of $V_X$; a component with no smooth real point would disprove the real-density claim as stated.","tokens_in":21893,"feed_emoji":"🎲","tokens_out":12758,"duration_ms":111819,"temperature":0.7,"pith_summary":"Spohn curves are the algebraic models of totally mixed dependency equilibria — equilibria in which every pure strategy is used with positive probability and players condition on one another's expected choices — for 2x2 normal-form games, and generically they are elliptic curves realized as the intersection of two quadrics in projective 3-space. The paper classifies when the planar Spohn cubic is reducible: exactly when its defining polynomial is nonzero and one of twelve explicit conditions on the payoff entries holds. It then argues that real points are Zariski dense in the Spohn variety, with the written theorem proving this for the twelve reducible cases when the remaining payoff entries are generic while the abstract states the density conclusion in all cases. This matters because real points on the curve are the usable equilibria, so density means the algebraic model does not lose the real solution space. The paper also gives a step-by-step computation of the $j$-invariant of a Spohn curve and proposes defining full game equivalence by isomorphism of these curves over $\\mathbb{Q}$.","feed_headline":"Twelve payoff conditions split the Spohn curve","feed_subtitle":"Real solution points survive in every split case, so 2x2 game equilibria are reachable by real algebra.","key_machinery":"The carrying object is the Spohn cubic $C\\subset\\mathbb{P}^2$: the ternary cubic obtained by eliminating $p_{22}$ from the two quadrics $\\det M_1$ and $\\det M_2$ that define the Spohn variety $V_X\\subset\\mathbb{P}^3$. Its seven coefficients are built from payoff differences, and the reducibility proof checks candidate lines through the intersections of $C$ with three coordinate lines. For the real-density step the load-bearing mechanism is the criterion that one smooth real point in every irreducible component forces Zariski density of the real points, combined with a generic projection result that lifts such points from the plane cubic back to $V_X$. For the invariant part the machinery is the reduction of two quadrics in $\\mathbb{P}^3$ with a common rational point to a plane cubic, the Aronhold invariants $S$ and $T$, the discriminant $\\Delta=(64S^3-T^2)/1728$, and the $j$-invariant $j=64S^3/\\Delta$.","core_discovery":"The central claim is Theorem 3.2: for a 2x2 game with payoff tables $A$ and $B$, the Spohn cubic $C=V(f)$, where $f=c_1x^2y+c_2x^2z+c_3xy^2+c_4xz^2+c_5y^2z+c_6yz^2+c_7xyz$, is reducible if and only if $f$ is nonzero and one of twelve listed conditions on the payoff entries holds. The coefficients $c_i$ are explicit differences of payoff entries, so each reducibility condition has a direct reading as an equality among payoffs. In Theorem 3.7 the paper proves, for each of the twelve reducible cases and assuming the remaining payoff entries are generic, that every irreducible component of the Spohn variety $V_X$ contains a smooth real point, and therefore the real points of $V_X$ lie Zariski dense in $V_X$; the abstract states this density conclusion in all cases, while the theorem itself carries the genericity assumption. The proof decomposes the ideal of $V_X$ into minimal primes, matches components with those of the plane cubic under elimination, and lifts smooth real points through a generic projection. The paper further supplies a complete pipeline for computing the $j$-invariant of a Spohn curve from two quadrics with a common rational point, and uses it to define full equivalence of 2x2 games by isomorphism of their Spohn curves over $\\mathbb{Q}$.","pith_inferences":["Editorial inference: because the abstract claims real density in all cases while the written theorem covers the twelve reducibility cases under genericity, a natural completion is to check the residual non-generic cases where $V_X$ can be reducible while the plane cubic is not, as in the paper's Example 3.8.","Editorial inference: since planar irreducibility does not imply irreducibility of the Spohn variety, the next combinatorial target suggested by this work is a direct classification of when $V_X$ itself splits, independent of the planar model.","Editorial inference: the continued-fraction approximation in Appendix A suggests a numerical stability test: approximate real payoffs by rationals, compute the $j$-invariant, and check whether the resulting game-equivalence classes stabilize as precision increases."],"forward_implications":["For generic 2x2 games the Spohn curve is an elliptic curve, so the Aronhold algorithm computes its $j$-invariant directly from payoff data and organizes game-theoretic properties by elliptic curve invariants.","Where the real-density proof applies, every irreducible component of the Spohn variety has real points, so solving over the reals does not discard any algebraic component of the equilibrium model.","In the reducibility cases (8) through (12), under the same genericity assumption, every point of the open simplex on the Spohn variety is a dependency equilibrium, while in cases (1) through (7) some components lie on boundary hyperplanes and require separate treatment.","Every Nash equilibrium of a 2x2 game is a dependency equilibrium, and for Prisoner's Dilemma-type games infinitely many dependency equilibria Pareto dominate the unique Nash equilibrium.","Two 2x2 games are fully equivalent when their Spohn curves are isomorphic over $\\mathbb{Q}$; equal $j$-invariants are necessary but not sufficient over $\\mathbb{Q}$, so twist-related curves remain an open case."],"supporting_citations":[{"why":"Stores the detailed polynomial decompositions and component calculations on which the proofs of Theorem 3.2 and Theorem 3.7 rely, since the paper says the detailed computations can be found there.","marker":"[7]"},{"why":"Establishes the general geometry of Spohn varieties, including that 2x2 games yield elliptic curves while other finite games give rational Spohn varieties.","marker":"[11]"},{"why":"Introduces boundary dependency equilibria and the genericity results on real points that this paper extends, and its Problem 4.3 motivates the reducibility classification.","marker":"[12]"},{"why":"Supplies the theorem that a smooth real point in an irreducible real variety implies Zariski density of the real points, the bridge between smooth-point checks and density.","marker":"[9]"},{"why":"Supplies the generic projection result used in the paper's Proposition 3.6 to pull smooth real points from the plane cubic back to the Spohn variety in $\\mathbb{P}^3$.","marker":"[1]"},{"why":"Supplies the Weierstrass normal form, the $j$-invariant, and the criterion that equal $j$-invariants characterize isomorphism over the algebraic closure.","marker":"[15]"},{"why":"Gives the algorithm that reduces two quadrics in $\\mathbb{P}^3$ with a common rational point to a plane cubic, used throughout the $j$-invariant computation section.","marker":"[4]"},{"why":"Provides the computer algebra computations that the paper invokes for the ideal decompositions in the proofs of Theorem 3.2 and Theorem 3.7.","marker":"[6]"}],"fun_headline_variants":["Twelve payoff splits on Spohn curve, real points dense","Elliptic curves in game theory: 12 reducibility cases, all real","Spohn curve reducibility: 12 conditions, real solutions survive","Game theory's elliptic curves: real points dense when split","2x2 games: Spohn curve splits in 12 ways, real equilibria"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the companion repository's polynomial decompositions are correct and exhaustive, and that the real-density statement proved for the twelve reducibility cases under generic remaining payoff entries genuinely extends to every case the abstract describes as 'all cases'.","fun_headline_variants_meta":{"raw":{"variants":["Twelve payoff splits on Spohn curve, real points dense","Elliptic curves in game theory: 12 reducibility cases, all real","Spohn curve reducibility: 12 conditions, real solutions survive","Game theory's elliptic curves: real points dense when split","2x2 games: Spohn curve splits in 12 ways, real equilibria"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000287,"raw_usage":{"total_tokens":1720,"prompt_tokens":1016,"completion_tokens":704,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":632,"completion_tokens_details":{"reasoning_tokens":608}},"tokens_in":632,"tokens_out":704,"duration_ms":6871,"temperature":1.0,"reasoning_tokens":608,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T14:57:59.761156+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a 2x2 payoff matrix that satisfies none of the twelve conditions in Theorem 3.2 and factor the corresponding ternary cubic; if it splits, the classification is incomplete. In one of the twelve cases, vary the payoff entries assumed generic and inspect the real points of each irreducible component of $V_X$; a component with no smooth real point would disprove the real-density claim as stated.","supporting_citations":[{"cited_title":"Kidambi, E","cited_arxiv_id":null,"evidence_quote":"Stores the detailed polynomial decompositions and component calculations on which the proofs of Theorem 3.2 and Theorem 3.7 rely, since the paper says the detailed computations can be found there."},{"cited_title":"Portakal, B","cited_arxiv_id":null,"evidence_quote":"Establishes the general geometry of Spohn varieties, including that 2x2 games yield elliptic curves while other finite games give rational Spohn varieties."},{"cited_title":"Portakal, D","cited_arxiv_id":null,"evidence_quote":"Introduces boundary dependency equilibria and the genericity results on real points that this paper extends, and its Problem 4.3 motivates the reducibility classification."},{"cited_title":"Mangolte: Real algebraic varieties, Springer Monographs in Mathematics, Springer , 2020","cited_arxiv_id":null,"evidence_quote":"Supplies the theorem that a smooth real point in an irreducible real variety implies Zariski density of the real points, the bridge between smooth-point checks and density."},{"cited_title":"Baldi, B","cited_arxiv_id":null,"evidence_quote":"Supplies the generic projection result used in the paper's Proposition 3.6 to pull smooth real points from the plane cubic back to the Spohn variety in $\\mathbb{P}^3$."},{"cited_title":"Silverman: The Arithmetic of Elliptic Curves, Graduate Texts in Mathematics, vol","cited_arxiv_id":null,"evidence_quote":"Supplies the Weierstrass normal form, the $j$-invariant, and the criterion that equal $j$-invariants characterize isomorphism over the algebraic closure."},{"cited_title":"Cohen: Number theory: Volume I: Tools and Diophantine Equations, Graduate Texts in Mathematics, vol","cited_arxiv_id":null,"evidence_quote":"Gives the algorithm that reduces two quadrics in $\\mathbb{P}^3$ with a common rational point to a plane cubic, used throughout the $j$-invariant computation section."},{"cited_title":"Grayson, M","cited_arxiv_id":null,"evidence_quote":"Provides the computer algebra computations that the paper invokes for the ideal decompositions in the proofs of Theorem 3.2 and Theorem 3.7."}],"review_version":1}