REVIEW 3 major objections 5 minor 4 references
A note on the Diversity Owen values
T0 review · 3 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash
Pith's one-line read The Diversity Owen value is uniquely characterized by three axiom sets once a null-player axiom repairs two flawed proofs.
desk verdict Useful correction paper with one big gap: the new Theorem 3 proof is invalid as written, but the repaired proofs of Béal et al.'s theorems look sound. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The load-bearing objects are the Diversity Owen value itself, defined as the Owen value of the diversity-restricted game $v_d$, and the Harsanyi dividend decomposition $v_d = \sum_{T} \Delta_{v_d}(T)u_T$, which lets the proofs work by induction on the number $|I(v_d)|$ of dividend-bearing diverse coalitions. The Null Player for Diverse Games axiom supplies the base case and the cancellation step: null players in diverse games receive zero, so subtracting a unanimity component removes all payoff difference. Fairness within Component and Fairness through Diversity then equate payoff differences between symmetric players or between communities, which transfers the induction hypothesis from smaller games to the full game.
What would settle it
Compute the marginal contributions of $i \in S_1$ and $k \in B_p \setminus (S_1 \cup S_2)$ in the game $\Delta_{v_d}(T_1)u_{T_1}$ for a game satisfying the Case 2.2 conditions. If there is a coalition containing $T_1$ but not $i$ whose worth changes differently with $i$ and $k$, the players are not symmetric and the derivation of $f_i - DOw_i = f_k - DOw_k$ collapses. For Theorem 1, one would instead test whether the extended-game construction in Lemma 1 preserves diversity when an outside player is added to a community at its quota.
Extended reading notes
Core claim
On the paper's own terms, the central discovery is that the Diversity Owen value is indeed uniquely characterized by the two axiom systems proposed in the earlier work, and that the missing step in those proofs is a proper treatment of null players in diverse games. The paper introduces the Null Player for Diverse Games axiom ($f_i(N,v,B,d)=0$ when $i$ is null and $(N,v,B,d)$ is diverse), shows that efficiency together with the earlier null-player-out axiom implies it, and then uses it to complete the induction in the uniqueness proofs. Theorem 3 adds an alternative characterization: the value is the unique rule satisfying Efficiency, Fairness within Component, Fairness through Diversity, Independence from Non-Diverse Coalitions, and Null Player for Diverse Games. Corollary 1 weakens the balanced-contributions axiom in the second characterization and retains uniqueness.
Load-bearing premise
The alternate characterization in Theorem 3 depends on the assertion that in Case 2.2 a player $i$ inside a diverse coalition $T_1$ and a player $k$ outside $T_1$ are symmetric in the unanimity game $\Delta_{v_d}(T_1)u_{T_1}$; if that symmetry fails, the equality used to chain the players' payoff differences does not follow.
Editorial extensions
If this is right
- The two characterizations proposed in the earlier paper become valid uniqueness results, so the Diversity Owen value is the single rule satisfying either five-axiom system.
- The Null Player for Diverse Games axiom gives a simpler route to the same conclusion, replacing the more complex null-player-out axiom in the first characterization.
- Corollary 1 shows the balanced-contributions axiom can be weakened while preserving uniqueness, so the axiomatic base of the value is thinner than previously thought.
- If the corrected proofs are accepted, the value's foundation no longer depends on the flawed induction steps identified in the original paper.
Reading between the lines
- A natural extension is to ask whether the same null-player axiom repairs the analogous uniqueness proof for the Diversity Shapley value; the paper does not discuss that value.
- The corrected induction on the number of dividend-bearing coalitions may transfer to other values defined by restricting games to feasible coalitions, such as values for games with precedence or communication constraints.
- Because ND is implied by efficiency together with NPOPD, the alternative characterization shows the value can be pinned down with a weaker null-player requirement; the independence of ND from the remaining axioms is left open.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper revisits the axiomatic characterizations of the Diversity Owen value (DOw) introduced by Béal et al. (2025a). It identifies two flaws in the original uniqueness proofs, provides corrected proofs in Sections 2.1 and 3.1, introduces the axiom "Null Player for Diverse Games" (ND), and claims a new alternative characterization (Theorem 3) in which NPOPD is replaced by ND. The corrected proofs are based on adding an outside null player to create slack, applying the relevant axioms, and then using induction on the number of nonzero Harsanyi dividends. The paper also derives a weakening of IBCOPPD in Corollary 1.
Significance. If the corrected proofs of Theorems 1 and 2 are valid, the paper repairs two published uniqueness proofs and is a useful corrigendum-style contribution. The identification of specific errors in the original proofs, together with Counterexamples 1 and 2, is valuable. However, the advertised alternative characterization is not established: the proof of Theorem 3 contains a false symmetry claim in Case 2.2, and the proof of Lemma 1 has an Efficiency step that omits the payoff of the added player in the restricted game. These are load-bearing issues for two of the paper's main claims, so the manuscript needs substantial revision before those claims can be accepted.
major comments (3)
- [§2.2, Case 2.2] The claim that "players i and k are symmetric in w1" is false. Because S1 = Bp ∩ T1, the chosen player i ∈ S1 belongs to T1 while k ∈ Bp \(S1 ∪ S2) does not belong to T1. In the unanimity game u_{T1}, a member and a non-member have different marginal contributions: for S = T1 \ {i}, u_{T1}(S ∪ i) − u_{T1}(S) = 1, whereas u_{T1}(S ∪ k) − u_{T1}(S) = 0. The same objection applies to the asserted symmetry of j and k in w2. Therefore Fairness within Component cannot be applied to obtain the two displayed equalities, and the conclusion that the deviations f_i − DOw_i are constant on Bp does not follow. Theorem 3 is thus unproven as written.
- [§2.1, proof of Lemma 1] The step marked "E=" applies Efficiency to the restricted game (N ∪ {l}) \ {i}, but Efficiency gives Σ_{j∈N\{i}} f_j(restricted) + f_l(restricted) = (v)+l((N ∪ {l}) \ {i}); the term f_l(restricted) is omitted in the displayed chain. Since f_l(restricted) = 0 has not been proved, and since NPOPD cannot remove l unless l is out in the restricted game (which is not guaranteed), the equality v(N) = f_i(N, v) + v(N \ {i}) does not follow as written. This gap is load-bearing because Lemma 1 is the basis for the ND axiom used in the corrected proof of Theorem 1. The lemma may be true, and can likely be repaired by splitting into the cases |Bk| > dk and |Bk| = dk, but the written proof needs correction.
- [§2.2, Theorem 3 proof, Case 3] The proof applies "Case 2" to the auxiliary game (v')d before Case 2 has been established as a general lemma; as written this is a forward reference inside the same induction step. This structural issue could be fixed by extracting the Case 2 argument as a separate lemma. It is secondary to the false symmetry claim in Case 2.2, but it should be addressed if Theorem 3 is retained.
minor comments (5)
- [§2.1, proof of Lemma 1] The NPOPD line in the display combines two applications of NPOPD (one removing l and one removing i) without explanation; this should be stated explicitly.
- [§2.1, Eq. (2)] The first equality in the chain for Theorem 1 uses NPOPD on the extended game with the added player l removed; this should be stated explicitly for readability.
- [§2.1, after Eq. (2)] The sentence "we immediately obtain fi(N, vd, B, d) = fi(N, vd, B, d), for all i ∈ Bk" contains an evident typo: the right-hand side should presumably be DOwi(N, vd, B, d).
- [§2.2, Case 2.2] The sentence "Then, T1 ∪ T2 ≠ N" is not used in the subsequent argument; either use it or delete it.
- [Abstract] The abstract states that the paper establishes an alternative characterization; since the proof of Theorem 3 is currently invalid, the abstract and the claims in Section 2.2 must be revised accordingly.
Circularity Check
No significant circularity: the corrected uniqueness proofs derive the Diversity Owen value from the axioms and induction; the noted flaw in Theorem 3 is a non-circular logical gap.
full rationale
The paper's central derivation is an axiomatic uniqueness proof, not an empirical prediction. The target value DOw(N,v,B,d)=Ow(N,vd,B) is defined independently of the axioms, and the proofs proceed by induction on |I(vd)|, comparing an arbitrary f satisfying the axioms with DOw through the axioms and the induction hypothesis. No parameter is fitted to data, no axiom is defined in terms of DOw, and no uniqueness theorem is imported from the same authors as an external black box. The cited facts about DOw satisfying FwC, FD, INDC, and E come from Béal et al. (2025a), by different authors, and are independently checkable from the definition; this is real evidence, not circular. Lemma 1 derives ND from E and NPOPD rather than assuming it, and the corrected proof of Theorem 1 uses that derivation legitimately. The alternative characterization in Theorem 3 is likewise an axiomatic argument. The only serious defect I see is non-circular: in Case 2.2 of Theorem 3, the assertion that i in S1 and k in Bp\(S1∪S2) are symmetric in w1=Delta_vd(T1)u_{T1} is false when i in T1 and k notin T1, so FwC cannot be applied to obtain the displayed equality. That is an invalid proof step and a correctness gap, not a reduction of the conclusion to the premises; the theorem may be true but is unproven as written. Because no claimed prediction or definitional equivalence reduces to its own inputs, the circularity score is 0.
Assumptions & free parameters
assumptions (4)
- ad hoc to paper Players i and k with i ∈ T1 and k ∉ T1 are symmetric in the unanimity game u_{T1}.
- standard math The extended game (v)+l has the same Harsanyi dividend support as v.
- standard math The induction principle on |I(vd)| is valid.
- domain assumption Definitions and axioms from Béal et al. (2025a) are correct.
Cite this review
Pith. "Pith review of A note on the Diversity Owen values." pith.science (2026). https://pith.science/paper/L3QZURZU
@misc{pith2026250524171,
author = {Pith},
title = {Pith review of: A note on the Diversity Owen values},
year = {2026},
howpublished = {\url{https://pith.science/paper/L3QZURZU}},
note = {Machine review of arXiv:2505.24171}
}
read the original abstract
B\'eal et al. (Int J Game Theory 54, 2025) introduce the Diversity Owen value for TU-games with diversity constraints, and provide axiomatic characterizations using the axioms of fairness and balanced contributions. However, there exist logical flaws in the proofs of the uniqueness of these characterizations. In this note we provide the corrected proofs of the characterizations by introducing the null player for diverse games axiom. Also, we establish an alternative characterization of the Diversity Owen value by modifying the axioms of the above characterizations.
Reference graph
Works this paper leans on
-
[1]
B\'eal S, Diss M, Tido Takeng R (2025a) New axiomatizations of the Diversity Owen and Shapley values. Int J Game Theory 54, 12
work page 2025
-
[2]
B\'eal S, Deschamps M, Diss M, Tido Takeng R (2025b) Cooperative games with diversity constraints. J Math Econ 116:103077
work page 2025
-
[3]
In: AW Tucker, RD Luce (ed) Contributions to the Theory of Games (Volume lV)
Harsanyi JC (1959) A bargaining model for the cooperative n -person game. In: AW Tucker, RD Luce (ed) Contributions to the Theory of Games (Volume lV). Princeton University Press, pp 325--356
work page 1959
-
[4]
Math Methods Oper Res 93:585–603
Hu XF (2021) New axiomatizations of the Owen value. Math Methods Oper Res 93:585–603
work page 2021
Reviewed August 7, 2026 · model on record in the stance chip above.
Discussion (0). Sign in to comment.