| paper | d4-journals-geb-2020-0128 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | high |
| authors would recognise it | unclear |
The paper’s numbered Results 1–4 report experimental behavioral comparisons, and Appendix C’s propositions derive behavioral predictions; none asserts an algorithm, computational hardness, or a computable object with a complexity claim. A type distribution could support behavioral aggregation, but without a named computational result there is no qualifying mirror to anchor.
fails bit a — no named computational result to mirror
No paper result qualifies as a computational anchor; Results 1–4 and Propositions 1–5 remain experimental or behavioral results.
This paper contains no named computational result of the kind required for an anchor. Its numbered Results 1–4 report experimental findings about voting and group entry. Appendix C’s Propositions 1–5 characterize behavior under assumptions about responsibility aversion. None asserts that a computational problem is tractable, hard, or has a particular parameterized complexity.
A continuous population model could describe mass over agent types and their choices to join groups and vote. The paper does not formulate or solve a computational problem on that population, though, so I can’t give a qualifying continuous problem per anchor without inventing one.
The proponent’s strongest point is also the strongest negative case: this paper states no computational problem whose complexity it studies. Results 1–4 are experimental comparisons, and Appendix C’s propositions derive behavior from assumptions about responsibility aversion. None gives a tractability, hardness, or algorithmic result to mirror.
The most promising candidate is Result 2: represent agents by a distribution over complete types, then ask whether voluntary entry amplifies the difference between defaults. Results 3 and 4 offer related questions about which types enter. But those are behavioral comparative statics, not computational questions. With groups fixed at three and two possible defaults, expected outcomes can be computed by summing over type triples under the population distribution. With \(\tau\) types, direct enumeration takes \(O(\tau^3)\); the entry and composition calculations are similarly small. That could make a useful analytic or empirical model, but it does not expose the kind of complexity landscape ChoCo studies.
A more computational version would optimize group assignments, vary group size, or design new decision rules. Those may be worthwhile questions in their own right, but they replace the paper’s random matching and fixed three-person voting setup with a planner’s problem the authors did not study. Taking group size to grow with the population also changes the object: it shifts the paper’s individual pivotality mechanism toward a limit in which each person’s influence may vanish. That is a new model, not a stronger mirror of these results.
The negative case is therefore strong for this paper and this programme’s scope, though it cannot prove that no related computational question would interest the authors. A continuum over types could sharpen their account of selection and group composition. The case against is that doing so produces behavioral aggregation, while making it computational requires adding a problem—such as optimal assignment—that is absent from the 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.