| paper | d4-journals-jet-2009-0006 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | medium |
| authors would recognise it | yes |
other→P discrete result → its continuous mirror
Answer: Polynomial-time solvable in \(\tau\) and \(L\). Aggregate finite-negotiation attainment holds exactly when \(b_t\leq 1/8\) for every type \(t\). · discrete side: other · adjudication: proved
full proof & adversarial review →
Theorem 2
Given rational masses \(\mu_t\) over a finite support of dispute-pair types, where type \(t\) specifies bias \(b_t\in(0,1/2)\) and the paper’s quadratic payoffs and uniform state \(\theta\in[0,1]\), compute typewise optimal arbitration and mediation rules and determine whether finite type-specific cheap-talk protocols attain the aggregate mediation value \(\sum_t\mu_t V_t^M\).
A mass of independent two-party disputes with complete type \(t\) recording the pair’s bias \(b_t\) and payoff structure. For each type, the decision variables are its arbitration rule, mediation rule, or finite cheap-talk equilibrium; the objective is population-weighted ex-ante decision-maker welfare, with no shared capacity constraint.
Theorem 2’s rule has \(N\) regions, and the paper gives no bound on the bit complexity of explicitly representing it for rational \(b\); this limits the algorithmic claim but not the well-posed population mirror.
fatal: False
The caseload mirror covers the optimal arbitration and mediation rules of Theorems 1 and 2 and the finite-negotiation attainability threshold of Theorem 3 after typewise aggregation; it leaves the paper’s other assumptions and applications unchanged.
I don’t see an honest positive case that meets the anchor requirement. The paper has no named computational result: it proves no hardness or tractability classification, and gives no analysis in terms of computational complexity.
Its numbered theorems are economic characterizations. Theorem 1, proved here, characterizes the optimal arbitration rule and shows it is deterministic. Theorem 2, also proved here, shows an optimal mediation rule can be implemented by filtering information and adding noise; the rule itself was proposed by Blume and Board. Theorem 3 shows that finite negotiation reaches the optimal mediated outcome if and only if \(b\le 1/8\); the paper proves the result, building on the equilibrium construction of Krishna and Morgan. These are substantive results, but none is an eligible computational anchor.
The nearest population scenario would be a large institution handling many disputes, with type specifying the parties’ preferences and payoff parameters, and mass representing the share of disputes of each type. One could choose protocols to maximize aggregate decision-maker welfare. But if disputes are independent, this simply aggregates the paper’s per-dispute analysis; adding shared capacity or other constraints would create a new problem. Also, the paper’s continuum is a prior over the state in a single dispute, not a continuum of agents. Treating that prior, or the randomized recommendation, as population continuization would miss ChoCo’s scope.
The proponent’s strongest point is also the main obstacle: Theorems 1–3 are not complexity results about a computational problem. Theorem 1 characterizes an optimal arbitration rule, Theorem 2 an optimal mediation rule, and Theorem 3 when finite negotiation matches mediation. The paper formulates no computational task whose complexity those results classify. That is a serious gap for a programme about computational questions; it is not a claim that the economic results are unimportant.
The continuum already in the paper does not fill that gap. The variable \(\theta\) describes uncertainty about the state in one dispute, and \(p(dy,d\theta)\) describes a mechanism’s random recommendations. Neither is a population of agents. Recasting the prior as population mass would continuize the wrong object.
The three theorems do not provide an obvious population mirror either. For Theorem 1, a large caseload could be represented by a distribution over dispute types, such as different biases \(b\). If each case can receive its own mechanism, the cases remain independent; population mass just weights their welfare. A shared protocol, mediator capacity limit, or other coupling could make the aggregate problem more involved, but it would add a constraint the paper does not study. Theorem 2’s randomization over recommendations is within a dispute; representing many disputes by masses does not turn that lottery into a population. Theorem 3’s threshold \(b\le 1/8\) could yield the share of cases for which negotiation matches mediation, but that is a population summary of a per-case characterization, not a complexity result about negotiation.
The strongest plausible mirror is therefore a high-volume institution handling many similar disputes. That is a sensible high-multiplicity regime, and it should not be dismissed merely because the paper studies two parties. But independent cases aggregate the paper’s results; coupled cases require new primitives such as shared capacity or a common protocol. Such an extension might be worthwhile, yet its computational questions would belong to the added institutional model rather than mirror a named result here.
So the negative case is strong for the proposed anchors, but weaker as a universal claim. The paper gives no computational result to anchor, and its existing continuum is over states, not agents. I cannot rule out a useful high-volume dispute-resolution model; the honest objection is that reaching one requires a new coupled population problem, not merely continuizing 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.