| paper | d4-journals-mss-2006-0036 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | high |
| authors would recognise it | yes |
Theorem 3 and Theorem 4 establish existence and multiplicity only for fixed \(n=5\); neither states an algorithm, complexity bound, or hardness result. The SAT runs verify those finite claims, but do not make finding a stable matching a claimed computation. The cohort-mass formulation is a coherent follow-on, but it cannot supply the missing named computational anchor.
fails bit a — no named computational result to mirror
The proposed extension concerns the stability question behind Theorems 3 and 4; it does not mirror a computational result or cover the paper's SAT formulations as named computational theorems.
The strongest positive case is a plausible follow-on problem, but it is not a valid mirror anchored in this paper’s named computational results.
The nearest anchors are Theorem 3, proved here, which says that every \(c3DSM\) instance with \(n=5\) has a stable matching, and Theorem 4, also proved here, which says that every such instance has at least two distinct stable matchings. The authors establish these claims by encoding counterexamples as SAT formulas and showing the formulas are unsatisfiable. But the theorems are finite existence guarantees, not results that classify a computational problem as tractable or hard; the paper explicitly leaves the complexity of general \(c3DSM\) open. The NP-completeness results mentioned in the introduction concern variants with incomplete acceptability or a prescribed partial matching. They are cited prior work, not numbered results of this paper, and they do not concern the paper’s base model. So none supplies the required anchor.
A concrete continuous problem suggested by the paper is Fractional Cyclic Stable Matching. The input is three finite supports of types, \(A,B,C\), with rational masses \(\mu_a,\nu_b,\rho_c\) summing to \(1\) on each side. Each type in \(A\) has a strict ranking of the \(B\)-types; each \(B\)-type ranks the \(C\)-types; each \(C\)-type ranks the \(A\)-types. A type represents a large cohort whose members share that ranking. A solution is a nonnegative mass allocation \(x_{abc}\) with the required marginals—for example, \(\sum_{b,c}x_{abc}=\mu_a\).
Call a triple \((a,b,c)\) blocking if positive mass of type \(a\) is assigned to \(B\)-types it ranks below \(b\), positive mass of type \(b\) is assigned to \(C\)-types it ranks below \(c\), and positive mass of type \(c\) is assigned to \(A\)-types it ranks below \(a\). The question is whether a stable allocation exists, meaning a feasible \(x\) with no blocking type triple; a solution could give \(x\) as binary rationals. The regime would be large populations in each cohort, with support sizes \(\tau_A,\tau_B,\tau_C\) much smaller than the population sizes.
This is a plausible continuous-population version of the underlying three-sided stability question: the population is represented by mass, and the matching is a mass allocation. But it changes the paper’s model in material ways. The paper ranks named individuals; this version ranks cohorts and lets their mass split across triples. The \(n=5\) theorems do not establish anything about this broader problem. I would leave its classification open: fractional allocations prevent an immediate hardness transfer from a discrete instance, while the stability condition introduces its own combinatorial constraints. Natural follow-up questions are whether every such society has a stable allocation, how existence can be decided in \(\tau_A,\tau_B,\tau_C\) and the mass bit length, and when discrete stable matchings do or do not correspond to stable mass allocations.
The weakest point is also decisive under the requested standard: the paper has no named computational-complexity result for its central problem to mirror. A referee might recognize Fractional Cyclic Stable Matching as a sensible new question, but would be right to call it a proposed extension rather than a continuous mirror of Theorem 3 or 4.
The cleanest objection is that neither named theorem is a computational result. Theorem 3 says every instance with \(n=5\) has a stable matching; Theorem 4 adds that each such instance has at least two. The SAT computations prove those fixed-size existence claims by ruling out counterexamples. Their running times are evidence for the proof method, not complexity bounds or algorithms for a variable-size problem. So neither theorem gives ChoCo a computational result to continuize.
The strongest proposed mirror is still a coherent follow-on: large cohorts on three sides, each type ranking the types on the next side, with rational masses and an allocation \(x_{abc}\) satisfying the mass constraints. One can define a block by requiring positive mass of each type to be assigned to less-preferred types. But retaining the paper’s model requires preferences over *named individuals*. Those identity-specific rankings do not give a fixed, compact type space as the population grows. Replacing them with rankings over cohort categories is a genuine change to the preference and blocking structure. The resulting fractional stability problem might be worth studying; it is not a consequence of Theorem 3 or 4.
This remains true in a plausible high-volume cohort scenario. A large market could have repeated types and assignments among them, but its stable-allocation question would need independent motivation and results. Theorem 4 is especially hard to carry over: “two distinct matchings” is a discrete claim, while a useful continuum counterpart would need to say how much mass the two allocations differ on. A fixed separation requirement would be a new claim.
That is a strong reason to reject this paper as a mirror candidate: its results establish finite existence facts, while the best population model changes the problem and has no computational theorem here to inherit. It does not prove the broader claim that no continuous three-sided stability problem is worthwhile. A fractional cohort model could be a good separate project; the honest weakness in the negative case is that lack of an anchor does not establish lack of value.
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.