| paper | d4-journals-mss-2025-0032 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | high |
| authors would recognise it | unclear |
Theorem 3 is an expected-value identity, and Proposition 2 is a uniqueness statement; neither asserts an algorithm, complexity bound, hardness result, or solver-used formulation. The experiments are finite numerical comparisons without a stated guarantee. The opponent also makes clear that a continuous allocation model would require new definitions, but bit (a) already decides the grade.
fails bit a — no named computational result to mirror
No result has a qualifying computational mirror; the paper's allocation identities, uniqueness result, axiomatic characterizations, and finite numerical comparisons remain outside the complexity question.
I find no qualifying anchor. The paper has no named result stating that a computational problem is in P, NP-hard, FPT, or otherwise classifying its complexity; its experiments are finite numerical comparisons, not complexity results.
The closest result is Theorem 3, proved here: for a \(\mathbb{P}\)-extendable intersection-closed incomplete game \((N,\mathcal K,v)\), the UD-value equals the expected Shapley value over all positive extensions of that game. Proposition 2, also proved here, says the UD-value is unique when \(\mathcal K\) is intersection-closed and \(N\in\mathcal K\). Both are mathematical results about the allocation rule, not computational complexity claims.
One could propose a follow-on problem about computing limiting UD payoffs for a population grouped into player types. But the paper does not define a type-level worth function or a continuum limit for its finite-player game. That would require adding the core model, so it would not be an honest mirror anchored in a named computational result of this paper.
The proponent’s claim that there is no computational result to anchor a mirror is the strongest part of the case. The paper proves statements about an allocation rule: Proposition 2 gives a uniqueness condition, and Theorem 3 identifies the UD-value with an average over positive extensions. The experiments compare values for small finite games. None states the complexity of a computational problem. So the paper offers no result whose complexity landscape a continuous mirror could chart.
The best replacement anchor would be Theorem 3. One could group interchangeable players into \(\tau\) types, describe coalitions by type-mass vectors, and ask for the limiting UD payoff per type—or an algorithm to compute it from a compressed description of the game. That is a plausible research question, and high-multiplicity cooperative games would support its motivation. But it needs new machinery before it becomes a mirror of this theorem. The finite result assigns dividends to subsets of named players and averages over finite-dimensional simplices. With a continuum of agents, there are uncountably many measurable coalitions; neither the dividend sum nor a uniform distribution over positive extensions comes with a canonical analogue. A limit through finite replica games could define one, but its answer would depend on how the feasible-coalition systems and worth functions scale. Those choices define a new model and computational problem; the paper does not supply one.
Proposition 2 has the same gap. Its uniqueness proof uses the finite coalition lattice and sums over all subsets. A continuum version would need a new notion of dividend and a replacement for those sums. That may be worthwhile mathematics, but it is not a continuous complexity mirror of the stated result.
The negative case is strongest under ChoCo’s computational gate: there is no named computational result here, and the natural continuizations lead first to new allocation theory. It is weaker as a universal claim. A carefully specified high-multiplicity model of interchangeable players could make computing UD allocations a sensible follow-on problem. I cannot rule out that the authors would recognize that as their problem; I can only say it is not anchored by a computational result in this paper.
The adversarial triple: the proponent anchors on up to three named results; the opponent sees that case and must defeat every anchor; the judge decides which case convinced it. These are the pipeline’s own outputs, generated by tools/triple_run.py — no human edited them. The paper’s own text is not reproduced here beyond the quoted statement above.