| paper | d4-journals-mss-2025-0047 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | high |
| authors would recognise it | unclear |
Theorem 2 is an axiomatic existence characterization, not a named computational result, and the proponent agrees it is not a qualifying anchor. The proposed anonymous rule over distributions may motivate a related project, but it cannot supply the missing computational result from this paper.
fails bit a — no named computational result to mirror
The proposed distributional question is motivated by Theorem 2, but it does not computationally mirror a named result; the other theorems and the profile proportions are left without a mirror.
The strongest positive case is a population version of the paper’s peer-review setting. Imagine a large grant programme with many reviewers assessing a fixed slate of proposals. Reviewers share evaluability sets and rankings: their type is \(t=(S,R)\), where \(S\) is the set of proposals they can evaluate and \(R\) is their weak order on \(S\). The society is a distribution \(\mu\) over these types. With many reviewers and a comparatively small number of recurring types, mass is a natural description.
A concrete computational question would be: given a finite set of proposals and a rational distribution \(\nu\) over evaluability sets, decide whether there exists a rule \(F\) on all distributions over \((S,R)\) with evaluability-set marginal \(\nu\), and produce such a rule if one exists. \(F\) must output a weak order on the proposals, always be transitive, satisfy Pareto whenever positive mass evaluates both proposals and unanimously prefers one, and satisfy IIA when two populations have the same distribution of pairwise judgments among reviewers who evaluate both proposals. This keeps the paper’s aggregation problem while making the population the continuous object; the outcome remains a ranking. I would tentatively expect existence recognition to be tractable if it reduces to a condition on the support of \(\nu\), but the paper does not establish that for this anonymous, distribution-based rule.
The nearest result is Theorem 2, proved in this paper. It says that for a fixed finite evaluability profile, an aggregating ranking function satisfying transitive valuedness, the Pareto criterion, and independence of irrelevant alternatives exists if and only if Condition 1 holds: every cycle in the graph of co-evaluated pairs is contained in the graph of at least one individual’s evaluable set. That is a useful motivation for the candidate question, but it is not a qualifying anchor under your criteria: it is an axiomatic existence theorem, not a named computational result establishing a complexity classification. The paper’s other named results are likewise axiomatic; its profile proportions do not supply a named complexity result either. So, strictly, there are no eligible anchors and no continuous problem here that can be presented as mirroring a computational result of the paper.
The weakest point in the candidate is substantial. Theorem 2’s construction can select particular individuals as local dictators. A population distribution over types has no distinguished individual to select, and anonymity changes the question: a positive mass of reviewers sharing a type is not the same as one named reviewer. Also, Theorem 2’s condition depends on which evaluability sets occur, not their rational masses. The proposed recognition problem is a plausible new computational study of this paper’s subject, but its central computational question—and whether the authors would accept that anonymous continuum rule as their problem—would need to be established.
The strongest objection is that this paper supplies no computational result to continuize. Theorem 2, like Theorems 1, 3, and 4, characterizes when an aggregating rule with specified axioms exists. The impossibility results are axiomatic too. Section 4 counts evaluability profiles; it does not classify the complexity of a computational problem. So the proposed grant-review model has no computational result in the paper as its anchor.
Theorem 2 is the nearest candidate, and its condition exposes the difficulty. It depends on which evaluability sets occur, not how much mass each receives. Replacing repeated reviewers with a distribution leaves the condition unchanged as long as the support is unchanged. One could study the complexity of recognizing Condition 1 from the supported sets, but that would be a new graph-recognition problem about an axiomatic existence criterion. The paper neither poses nor solves that computational problem, and the masses introduce no new computational structure into it.
The proponent’s anonymous version is a more substantial proposal, but it changes the question: it asks for rules on distributions and constrains them through aggregate pairwise judgments. The paper’s rules take reports from named individuals, and its constructions can select particular reviewers to resolve cycles. Characterizing anonymous rules over a continuum could be worthwhile in its own right; it would be a new axiomatic aggregation project, rather than a continuous mirror of a computational result here. Asking an algorithm to evaluate such a rule would require specifying the rule or adding a new selection or optimization task.
This negative case is strongest on fit with ChoCo’s computational programme, not on whether a large review population can sensibly have repeated types. That scenario is plausible. The weakness is that it could support a related project, while the paper itself offers no computational question whose continuous-population complexity landscape this project would chart.
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.