{"id":"866412b4-f5a5-495e-a949-68856507c638","arxiv_id":"2606.01477","paper_version":3,"verdict":"UNVERDICTED","confidence":"LOW","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Under cash-additivity, normalization, concavity, strong dynamic consistency and law-invariance, the Avellaneda-Stoikov framework is the unique entropic model and Cartea-Jaimungal is its Taylor expansion with phi forced to gamma sigma squared over 2.","lead":"The paper shows that two popular inventory market making models are linked by a uniqueness theorem: under five axioms on preferences, the Avellaneda-Stoikov model is the only one allowed and the Cartea-Jaimungal model is its second-order approximation with the running penalty forced to equal half gamma times sigma squared. A smart generalist might read it to see how separate parameter calibrations in trading desks can be cross-checked for consistency rather than treated as in","discovery_kind":"unification","skeptic_critique":{"model":"grok-4.3","headline":"Regularity condition invoked to force terminal coefficient α = ½ L''(0) is the least secure link between the two frameworks","rationale":"The reader’s weakest_assumption is exactly the point at which the claimed equivalence between the two frameworks can break. The axiom-based uniqueness to the entropic functional appears internally consistent with the listed properties; the only additional assumption needed for the full claim is the regularity on L, which the paper already flags. Hence the verdict moves from UNVERDICTED to CONDITIONAL once that condition is verified.","tokens_in":1845,"tokens_out":394,"duration_ms":23227,"concrete_test":"State the precise regularity condition used in the Taylor-expansion argument (likely near the paragraph deriving α). Verify whether it holds for the two standard liquidation costs L(q) = η|q| and L(q) = (η/2)q² by direct computation of the second-order term; if the derived α deviates from ½ L''(0) for either, the forced relation between frameworks is incomplete.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central claim asserts that the axioms force the preference to be entropic (hence AS is unique) and that CJ is its second-order Taylor expansion in inventory, with both φ and α pinned by the single γ. The abstract explicitly qualifies the α relation by “a mild regularity condition on the liquidation cost function.” If this condition fails (e.g., L not twice differentiable at zero, or the second-derivative term not isolated by the expansion), the terminal penalty in the CJ approximation is no longer determined by γ; the claimed “single underlying object” relation then holds only for the running penalty φ, not for the full pair (φ, α). Because the uniqueness theorem itself does not require this regularity, the load-bearing gap is confined to the Taylor-expansion step that connects the two concrete frameworks.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.3","summary":"The paper claims that five axioms on the market maker's dynamic preference functional (cash-additivity, normalization, concavity, strong dynamic consistency, and law-invariance) force the functional to be the entropic certainty-equivalent on liquidation-adjusted terminal wealth, parametrized by a single positive scalar γ. The Avellaneda-Stoikov framework is the unique representative of this axiom class. The Cartea-Jaimungal framework is its second-order Taylor expansion in inventory magnitude, with the running coefficient forced to φ = γσ²/2 and (under a mild regularity condition on the liquidation cost function) the terminal coefficient forced to α = ½L''(0). The forced relation is invertible, γ = 2φ/σ², providing a consistency cross-check on independently calibrated parameters.","tokens_in":2014,"tokens_out":549,"duration_ms":22168,"significance":"If the result holds, the paper unifies two standard inventory market-making frameworks by deriving them from the same axiomatic preference structure, showing they are manifestations of a single object rather than competing alternatives. This supplies a theoretical basis for relating their free parameters and a practical cross-check on desk calibrations. The axiomatic derivation yielding a parameter-free uniqueness result and the explicit invertible relation between γ and φ are strengths that would be valuable to the mathematical finance literature on market making.","major_comments":[{"comment":"Abstract, paragraph on the Taylor expansion: The relation α = ½ L''(0) is qualified by a 'mild regularity condition on the liquidation cost function.' The manuscript must explicitly define this condition (e.g., twice differentiability of L at zero and isolation of the second-derivative term) and state the consequences if it fails, because the uniqueness theorem itself does not invoke the condition; without it the terminal penalty in the Cartea-Jaimungal approximation is no longer pinned by the same γ, so the 'single underlying object' claim holds only for the running penalty φ.","section":"Abstract (Taylor expansion paragraph)"},{"comment":"Derivation of the entropic form (section containing the uniqueness theorem): The abstract asserts that the five axioms force the entropic certainty-equivalent. The full proof steps from strong dynamic consistency and law-invariance to this specific functional form must be checked to confirm that the subsequent Taylor-expansion step does not introduce post-hoc restrictions that affect the central claim relating the two frameworks.","section":"Uniqueness theorem derivation"}],"minor_comments":[],"recommendation":"major_revision","confidential_remarks":"The provided materials contain only the abstract; the full proof is not visible, which prevents verification of every algebraic step and contributes to the low soundness rating. The paper would be strengthened by including the complete derivation or key intermediate lemmas."},"author_rebuttal":{"model":"grok-4.3","summary":"We thank the referee for the detailed and constructive report. The comments highlight important points on clarity and the precise scope of the uniqueness result versus its approximation. We address each major comment below and indicate the revisions we will make.","responses":[{"response":"We agree that the mild regularity condition requires explicit definition and that its failure affects only the terminal coefficient. We will revise the abstract and the relevant section on the Taylor expansion to state the condition as twice continuous differentiability of the liquidation cost L at zero (with the second-derivative term isolated after the first-order term vanishes by normalization). If the condition fails, the terminal penalty α in the Cartea-Jaimungal approximation is no longer forced to equal ½L''(0) by the same γ, while the running-penalty relation φ = γσ²/2 continues to hold unconditionally from the second-order expansion of the entropic functional. The uniqueness theorem itself remains unaffected, as it concerns only the entropic form; the 'single underlying object' claim will be qualified accordingly in the revised text.","revision_made":"yes","referee_comment":"[Abstract (Taylor expansion paragraph)] Abstract, paragraph on the Taylor expansion: The relation α = ½ L''(0) is qualified by a 'mild regularity condition on the liquidation cost function.' The manuscript must explicitly define this condition (e.g., twice differentiability of L at zero and isolation of the second-derivative term) and state the consequences if it fails, because the uniqueness theorem itself does not invoke the condition; without it the terminal penalty in the Cartea-Jaimungal approximation is no longer pinned by the same γ, so the 'single underlying object' claim holds only for the running penalty φ."},{"response":"The uniqueness theorem is proved in Section 3. The argument proceeds by first invoking cash-additivity, normalization and concavity to obtain a concave monetary utility functional, then applying strong dynamic consistency to obtain a recursive representation, and finally using law-invariance to reduce the problem to a static entropic form on the terminal liquidation-adjusted wealth; the resulting functional is necessarily the entropic certainty equivalent parametrized by a single γ > 0. The Taylor-expansion step that produces the Cartea-Jaimungal running and terminal penalties is applied only after the uniqueness result has been established and does not feed back into the axiomatic derivation. Consequently, no post-hoc restrictions are introduced. We are prepared to expand the proof steps in an appendix if the editor requests further detail, but the existing derivation already separates the axiomatic uniqueness from the subsequent approximation.","revision_made":"no","referee_comment":"[Uniqueness theorem derivation] Derivation of the entropic form (section containing the uniqueness theorem): The abstract asserts that the five axioms force the entropic certainty-equivalent. The full proof steps from strong dynamic consistency and law-invariance to this specific functional form must be checked to confirm that the subsequent Taylor-expansion step does not introduce post-hoc restrictions that affect the central claim relating the two frameworks."}],"tokens_in":1565,"tokens_out":643,"duration_ms":15478,"standing_objections":[]},"desk_editor":{"model":"grok-4.3","letter":"The main takeaway is that a short list of axioms on the market maker's dynamic preference functional—cash-additivity, normalization, concavity, strong dynamic consistency, and law-invariance—pins the objective down to the entropic certainty equivalent on liquidation-adjusted wealth, so the Avellaneda-Stoikov framework is the only one that satisfies them, while Cartea-Jaimungal appears as its second-order expansion in inventory with the running penalty forced to φ = γ σ² / 2.\n\nThis relation is new. Earlier papers treated the two frameworks as separate modeling choices driven by tractability, without showing one emerges from the other via preference axioms or giving an explicit invertible map between their free parameters. The paper does a clean job of stating the axioms and isolating where the Taylor step requires the mild regularity condition on the liquidation cost to pin α = ½ L''(0).\n\nThe soft spot is exactly that regularity condition. The abstract qualifies the terminal coefficient claim with it, but does not say what happens if it fails (for example if L is not twice differentiable at zero). In that case the claimed single-object relation holds only for the running penalty and not for the full pair of coefficients. The uniqueness theorem itself does not need the condition, so the gap is confined to the approximation step that connects the two concrete models. Without the full proof it is also impossible to check whether the steps from the axioms to the entropic form are free of hidden restrictions.\n\nThe work is aimed at quantitative finance researchers who already use or calibrate inventory market-making models. A reader familiar with both frameworks will get a concrete consistency check on independently calibrated parameters.\n\nIt deserves peer review. The claim is narrow and the result, if the proof holds, supplies a useful cross-check inside an active modeling area.","headline":"The paper derives that five axioms force the Avellaneda-Stoikov objective to be the unique entropic form and Cartea-Jaimungal to be its inventory Taylor expansion with φ tied to γ, but the terminal coefficient link needs an extra regularity condition whose scope is unclear from the abstract.","tokens_in":2489,"tokens_out":467,"would_cite":false,"duration_ms":20817,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.3","headline":"Axioms on dynamic preferences force the Avellaneda-Stoikov and Cartea-Jaimungal market-making frameworks to be two views of the same entropic object with linked parameters.","keywords":["inventory market making","entropic certainty equivalent","dynamic consistency","preference axioms","Avellaneda-Stoikov model","Cartea-Jaimungal model","liquidation penalty","market making parameters"],"falsifier":"Independent calibration of both frameworks on identical market data, followed by a direct check whether the fitted values satisfy φ = γ σ² / 2 within statistical error; systematic violation on multiple assets or periods would falsify the forced uniqueness.","tokens_in":2734,"feed_emoji":"","tokens_out":814,"duration_ms":17248,"temperature":0.7,"pith_summary":"The paper establishes that cash-additivity, normalization, concavity, strong dynamic consistency, and law-invariance on a market maker's preference functional force it to take the specific form of an entropic certainty equivalent applied to liquidation-adjusted terminal wealth, parametrized by one scalar γ. This axiom set makes the Avellaneda-Stoikov model the unique representative inside the class. The Cartea-Jaimungal model then appears as the second-order Taylor expansion of that same functional in inventory size, which pins the running penalty coefficient at φ = γ σ² / 2 and, under a regularity condition, the terminal coefficient at α = ½ L''(0). A reader would care because the two frameworks are thereby shown to be different approximations of one underlying preference rather than independent modeling choices that can be calibrated separately.","feed_headline":"Axioms force Avellaneda-Stoikov and Cartea-Jaimungal to share one parameter","feed_subtitle":"The running penalty must satisfy φ = γ σ² / 2, turning two free parameters into a single scalar with an immediate calibration cross-check.","key_machinery":"The entropic certainty-equivalent on liquidation-adjusted terminal wealth, parametrized by a single positive scalar γ, which encodes the entire preference structure.","core_discovery":"Under the five listed axioms the preference functional is forced to be the entropic certainty-equivalent on liquidation-adjusted terminal wealth parametrized by a single positive scalar γ; the Avellaneda-Stoikov framework is therefore the unique model in this class, while the Cartea-Jaimungal framework is its second-order Taylor expansion in inventory magnitude with the running coefficient forced to φ = γ σ² / 2 and the terminal coefficient forced to α = ½ L''(0) under the stated regularity condition on the liquidation cost.","pith_inferences":["If the relation is observed to hold in practice, it would suggest that real market-maker risk preferences are close to entropic on liquidation-adjusted wealth.","The unification opens the possibility of importing approximation techniques or numerical methods from one literature directly into the other without re-calibrating free parameters.","Extensions to multi-asset or stochastic-volatility settings could be tested for consistency by checking whether the same γ continues to link the running and terminal penalties across assets."],"forward_implications":["The two frameworks cannot be treated as competing alternatives whose choice is driven only by tractability; they are different manifestations of one preference object.","The relation γ = 2 φ / σ² supplies an immediate consistency cross-check on any pair of independently calibrated desk parameters.","Higher-order expansions of the same entropic functional would generate further market-making approximations whose coefficients are likewise determined by the single scalar γ.","Any model that preserves the five axioms must reproduce the same entropic form and therefore the same parameter linkage."],"fun_headline_variants":["Axioms unify AS and CJ under single entropic γ","Forced relation sets φ = γσ²/2 for market makers","Cartea-Jaimungal as second-order expansion of AS","Single γ parametrizes both inventory market making","Axioms force φ to γσ²/2 in unified frameworks"],"cache_read_input_tokens":2112,"weakest_assumption_plain":"The mild regularity condition on the liquidation cost function that is used to force the terminal coefficient α to equal ½ L''(0).","fun_headline_variants_meta":{"raw":{"variants":["Axioms unify AS and CJ under single entropic γ","Forced relation sets φ = γσ²/2 for market makers","Cartea-Jaimungal as second-order expansion of AS","Single γ parametrizes both inventory market making","Axioms force φ to γσ²/2 in unified frameworks"]},"model":"grok-4.3","cost_usd":0.00565,"raw_usage":{"total_tokens":2738,"prompt_tokens":741,"num_sources_used":0,"completion_tokens":82,"cost_in_usd_ticks":56499500,"prompt_tokens_details":{"text_tokens":741,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":1915,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":741,"tokens_out":82,"duration_ms":15025,"temperature":1.0,"reasoning_tokens":1915,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-06-28T15:36:14.904399+00:00","model_set":{"reader":"grok-4.3"},"falsifier":"Independent calibration of both frameworks on identical market data, followed by a direct check whether the fitted values satisfy φ = γ σ² / 2 within statistical error; systematic violation on multiple assets or periods would falsify the forced uniqueness.","supporting_citations":[],"review_version":1}