| paper | d4-journals-jet-2021-0060 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | high |
| authors would recognise it | no |
The paper’s named results are analytical allocation characterizations and convergence claims, not computational results under the stated criterion; its CRRA comparisons are numerical without a stated guarantee. The paper also already models a unit-mass continuum, so a new complexity question about computing its contracts would not mirror a named computational result.
fails bit a — no named computational result to mirror
No results are mirrored; Corollary 3’s aggregate-allocation equality, Corollary 4’s convergence claim, and the numerical CRRA comparisons remain outside coverage.
I can’t make an honest positive case under the anchor rule: the paper has no numbered result classifying a computational problem as tractable, hard, FPT, or otherwise. Its numbered results are analytical. For example, Corollary 3 says that with logarithmic utility, the optimal aggregate allocation under private information equals the one under full information; Corollary 4 says those aggregate allocations converge to a stationary stochastic process. Neither is a complexity result. The numerical comparisons for more general CRRA preferences are not numbered complexity results either.
There is also a direct collision with a population mirror: the paper already models a unit-mass continuum of agents and tracks old agents through a measure \(\mu_t\) over promised values. A problem asking for optimal contracts given that distribution would be a computational question about the paper’s existing continuum model, not a continuous-population version of one of its discrete problems. A finite-agent approximation could be studied, but it would introduce a new problem rather than mirror a named anchor.
So there are no qualifying anchors, no continuous problem to state per anchor, and no expected complexity classification to give.
The absence of a computational result is decisive here. Lemma 1 and Corollary 3 characterize aggregate allocations analytically; Corollary 4 establishes their convergence to a stationary process. The CRRA comparisons are numerical. None classifies the complexity of an algorithmic problem. One could ask for the complexity of computing an optimal contract, but that would add a computational question the paper does not answer or state as a result.
The strongest rescue is to frame Corollary 3 computationally: given the model, compute optimal aggregate allocations under private and full information and compare them. But the paper already formulates both problems for a unit-mass population and proves the comparison in that setting. That is a question about an existing continuous model, not a population continuization. A finite-agent approximation would be a new problem, and the paper supplies no named discrete computational result for it to mirror.
The collision is unusually direct: the paper’s agents already form a continuum, and its state includes a measure over promised values. That is different from prior high-multiplicity work supporting a continuous mirror; here the paper itself already takes the population limit. The computational method used for numerical solutions does not change that or provide a complexity classification. I see no worthwhile continuous-population mirror of a result in 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.