| paper | d4-journals-jet-2014-0059 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | high |
| authors would recognise it | yes |
The replication model makes a continuous population interpretation plausible, but none of the named results asserts an algorithm, hardness, or computational guarantee. The assignment LP appears in the proofs as an analytical tool, not as a formulation the paper uses to compute, so there is no computational result here to anchor a mirror.
fails bit a — no named computational result to mirror
None of the paper's results yields a computational mirror; Theorems 1–3 and the other named propositions concern limits, solution-set inclusions, nonemptiness, or equilibrium structure.
There is a credible population model here. A buyer type could be \((v,d)\), recording valuations and demand capacity; a seller type could be \((r,q)\), recording reservation prices and inventories. In a large wholesale market with many repeated buyer and supplier profiles, mass records the share of each type. A continuous assignment would send trade mass between buyer and seller types, subject to their per capita capacities. The paper’s replication setup is direct evidence that this high-multiplicity regime makes sense to its authors.
But the paper has no named computational result of the kind required for an anchor. Its closest results are:
These are limit and solution-set theorems, not results about an algorithm, hardness, or parameterized complexity. The assignment LP in the appendix is used in proofs; the paper does not give a named complexity result for computing or deciding anything from it.
So the honest positive case is limited: the market admits a plausible type-mass continuation, and replication supports its high-multiplicity interpretation. But there is no per-anchor computational question I can responsibly hand to a prover as a mirror of this paper’s named results. The gap is the requested one: the paper studies what happens to cooperative solution sets as markets grow, but does not chart the computational complexity of those problems.
The proponent’s strongest premise is sound: this market has a credible high-multiplicity regime. In a large wholesale market, many buyers can share a valuation-demand type \((v,d)\), and many sellers can share a reservation-price-inventory type \((r,q)\). The paper’s replication construction supports that interpretation. A mass model with per-capita capacities and type-level average trades would be a reasonable continuization. I cannot make a serious objection based on agents being individuated or on fractional trades.
The problem is that none of the three theorems the proponent names is a computational result. Theorem 1 is a limit statement about the Core as \(\rho\) grows; Theorem 2 compares Set-wise stability and competitive-equilibrium payoffs in a two-fold replica; Theorem 3 gives a non-equilibrium Core payoff for some market at each replica count. They establish facts about solution sets. They do not pose an algorithmic input-output problem or classify its complexity. That is the strongest negative ground available here: the continuous population model is plausible, but the paper offers no named computational result for a continuous mirror to inherit.
Theorem 1 is the best candidate. One could define a continuum market over the same buyer and seller types and ask whether its Core coincides with competitive equilibrium. But that is a solution-set or limit question, and the paper’s central contribution already concerns precisely that kind of convergence. To make it computational, one would need to add a new task—such as deciding membership in the continuum Core from a finite-support rational input. That may be a worthwhile project, but it is a new complexity question, not a mirror of Theorem 1.
The same gap applies to Theorems 2 and 3. A continuum version of Theorem 2 would compare two solution sets; an algorithmic version would require defining a computational task for Set-wise stability, whose blocking condition quantifies over distributions and assignments. Theorem 3’s quantifiers also range over a different market for each \(\rho\), rather than following one fixed society into a population limit. One can devise computational questions around either result, but the results themselves provide no complexity claim to mirror.
So the negative case is narrow: the market model and its continuous population are convincing, but that does not make this paper a strong ChoCo mirror. Its named results are about the economic behavior of solution concepts as markets are replicated, not about computational complexity. I cannot honestly argue that no useful continuous computational question could ever be asked about this model; I can argue that answering one would be a new research direction, rather than continuizing a computational result in the 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.