| paper | d4-journals-geb-2024-0090 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | high |
| authors would recognise it | unclear |
Theorem 1 and Theorem 2 are axiomatic characterizations, and Theorem 3 is an impossibility result about mechanism properties; none is a named computational result under the stated criterion. The opponent's proposed mass version of tTTC could motivate a follow-on computational study, but it cannot supply the missing paper result.
fails bit a — no named computational result to mirror
No result is covered by a continuous computational mirror: Theorems 1 and 2 characterize tTTC axiomatically, while Theorem 3 and the other impossibility results concern mechanism properties.
The strongest positive case is a large exchange clearinghouse for several kinds of goods—course slots across sessions, for example—where many participants share the same complete preference and endowment pattern. Write the share of cohort \(a\) as \(\mu_a\), with \(n\gg\tau\). A continuous allocation could send mass \(x_{a,b}\) of each cohort to integral bundles \(b\), while preserving cohort mass and the supply of each good. That is a plausible high-multiplicity setting for a computational study related to typewise TTC.
But it is only a prospective research direction, not a mirror anchored in this paper’s results. There are no named computational results here asserting polynomial time, hardness, or another complexity classification. Theorem 1 characterizes the tTTC mechanism on lexicographic preferences: it is the unique mechanism that is unanimous (or onto), individually rational, strategy-proof, and non-bossy. Theorem 2 gives the same characterization for separable preferences. Theorem 3 rules out a mechanism on strict preferences that is unanimous, individually rational, and strategy-proof. These are axiomatic results, not computational ones. The paper defines a cycle algorithm for computing tTTC, but does not give a named complexity result for that computation.
There is also a modeling obstacle: every agent owns distinct objects, so a complete type includes that agent’s endowment. In the paper’s model, that can make the number of types grow with \(n\). Keeping \(\tau\) small would require treating many objects as interchangeable copies and defining a mass version of typewise TTC. That may be worthwhile, but the paper neither defines nor analyzes such a version. Since the requested anchors must be named computational results, I cannot give a continuous problem per anchor without misrepresenting the paper. The weak point is decisive: the best continuous question would be a follow-on to this paper, not a continuous mirror of one of its computational results.
The case for does not actually offer a continuous mirror of Theorems 1, 2, or 3; it names them to explain why it thinks no mirror is available. Its strongest point is that the paper states no complexity theorem. That makes a *literal* complexity mirror hard to claim, but it does not establish that there is no worthwhile computational question attached to the paper’s central mechanism.
The strongest candidate is the tTTC algorithm characterized in Theorems 1 and 2. Consider a large exchange market for course slots, with interchangeable copies of each slot and a finite set of cohorts. A cohort’s complete type records its preference pattern and the class of slots it initially owns. Give each cohort a rational mass, and scale the supply of each slot class with the population. Then ask for the exact mass allocation produced by a continuum version of typewise TTC, with running time in the number of slot classes, support size \(\tau\), and input bit length \(L\). The algorithm’s cycle structure gives this question a direct connection to the paper’s central result. Fractional cohort totals are compatible with every individual receiving an integral bundle.
The distinct-object objection does not rule out that regime. It shows that the paper’s literal instances may have large support when every object label matters. In a market where copies are genuinely interchangeable, the label is not part of an agent’s economically relevant type; preference and endowment *classes* are. That is a standard high-multiplicity setup, and the paper’s typewise exchange structure makes it a plausible setting for the authors’ problem. The required co-modification—supplies scaling with population—is part of the continuous model.
This would be an algorithmic companion to the characterization, rather than a continuous version of its strategy-proofness theorem. Theorem 3’s impossibility is an even weaker anchor: its argument turns on two named agents and particular trades, and a continuum version would need a careful replacement for individual manipulation. So the positive case is limited. But the existence of a natural, finite-support population regime and a defined mechanism to compute makes “no worthwhile continuous mirror in any scenario” too strong. The paper does not itself supply a computational result; whether the tTTC computation yields a sufficiently rich complexity question remains the main vulnerability of this case.
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.