| paper | d4-journals-jet-2004-0084 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | high |
| authors would recognise it | unclear |
Theorems 3 and 4 give a necessary-and-sufficient mathematical characterization of core nonemptiness, but neither states an algorithm, hardness result, or other computational procedure. A continuous-population NTU core is a plausible research direction, but it is not a computational analogue anchored in a result of this paper.
fails bit a — no named computational result to mirror
The proposed direction draws on Theorems 3 and 4's -balancedness characterization; it leaves Theorem 1's earlier characterization and Theorem 2's fixed-point result as background.
A credible continuous-population research question is nearby. Imagine a large cooperative with many members in a small number of complete types—members sharing the same preferences, utility scale, and coalition-relevant opportunities. A type’s mass \(\mu_t\) is its fraction of the cooperative. Let \(\Omega\) be the population, partitioned into those types. For each measurable coalition \(A\subseteq\Omega\), an NTU game specifies the measurable utility profiles \(V(A)\) it can achieve. The question is whether there is a grand-coalition profile \(u\in V(\Omega)\) that no measurable coalition can block: there is no \(A\) and \(v\in V(A)\) with \(v(\omega)>u(\omega)\) for almost every \(\omega\in A\). That keeps the paper’s central object—core immunity to coalition deviations—while making the population itself continuous. A large cooperative with, say, millions of members but a few dozen recurring roles or contract types is a plausible high-multiplicity regime.
The paper’s strongest nearby result is the combination of Theorems 3 and 4, both established here. Theorem 3 says that if an NTU game is \(\Pi\)-balanced for some convex-valued upper-hemicontinuous correspondence \(\Pi:\mathbb R^n\to\Delta\), then its core is nonempty. Theorem 4 says that if the core is nonempty, such a correspondence exists making the game \(\Pi\)-balanced. Together they give a necessary and sufficient condition for core nonemptiness. This is a natural *subject* to extend to a population of types and masses, but it is an existence characterization, not a computational result.
Under the requested anchor rule, this paper has no qualifying computational anchor: none of its numbered results says that a problem is in P, NP-hard, FPT, or otherwise classifies computational complexity. Theorem 1 is a characterization credited to Keiding and Thorlund-Petersen; Theorems 2–4 concern fixed-point and core-existence conditions. None supplies the required complexity claim. So I cannot honestly give one continuous computational problem per anchor or predict its complexity class.
The weak point in the positive case is also substantial: the paper specifies feasible sets for coalitions of a *finite* player set. A continuum mirror needs a new representation of feasible utility profiles for measurable coalitions, and assumptions connecting those sets to type composition. The proposed core-nonemptiness question is therefore a plausible research direction, but it is not a complexity mirror anchored in this paper. Making it a prover-ready computational problem would first require an effective finite encoding of the continuum game and a result connecting the paper’s \(\Pi\)-balancedness characterization to that encoding.
The strongest objection is at the programme’s entry point: this paper states no computational result to mirror. Theorem 1 is credited to Keiding and Thorlund-Petersen; Theorem 2 is a fixed-point tool. The authors’ Theorems 3 and 4 characterize core nonemptiness through \(\Pi\)-balancedness. None classifies an algorithmic problem by complexity, gives an algorithm, or specifies a finite input representation. A computational version of core nonemptiness can be posed, but its encoding and complexity question would have to be supplied by later work. They are not results of this paper.
There is a plausible high-multiplicity setting: many cooperative members with a few recurring roles or contract types. The obstacle is not that such a population is implausible. It is that the paper allows an arbitrary feasible set \(V(S)\) for every coalition \(S\). To preserve that generality over a continuum would require describing feasible utility profiles for all measurable coalitions, with no finite encoding given here. To obtain a finite-type model, one must add a rule that feasibility depends on the coalition’s type masses and specify how type-level utility profiles are represented. That is a sensible modelling choice, but it supplies the structure on which any computational result would rest.
The strongest version of the proposed mirror would therefore be: give a finite description of a type-mass-indexed NTU game and decide whether its core is nonempty. That could be a worthwhile project in its own right. But Theorems 3 and 4 do not make that question computational: they quantify over a convex-valued upper-hemicontinuous correspondence \(\Pi\), and the necessity theorem constructs such a correspondence from an already-existing core point. The paper gives no finite encoding for \(\Pi\), no separation problem, and no complexity claim. Its balancedness characterization is an existence theorem, not an algorithmic certificate as presented.
So the pair of theorems the other analyst identifies is the only serious candidate, and the best population-level reformulation still does not yield a computational mirror of their results. I would not claim that continuous core theory, or even a computational study of a carefully encoded mass-indexed NTU game, is intrinsically unworthy. The negative case is weaker if “mirror” is allowed to mean a new computational programme inspired by the paper’s subject. Under ChoCo’s stated scope—computational questions anchored in this paper—the paper supplies no such anchor.
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.