| paper | d4-conf-ijcai-2009-0278 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | high |
| authors would recognise it | no |
None of the paper’s named results asserts an algorithm, a computational bound, or hardness; they characterize rules and bound structural quantities. The proposed manipulation-mass problem is computational, but it is not a continuous analogue of a result proved in this paper.
fails bit a — no named computational result to mirror
No result has a computational mirror here: Theorem 1 gives a characterization and order bounds, while Theorems 2–6 give consistency-degree bounds; the frequency-of-manipulability results cited in the paper are from prior work.
I can’t make a credible case with the required anchors: this paper contains no named computational result asserting that a problem is hard, tractable, or parameterized-tractable. Its named results are structural.
The closest foothold is Theorem 1, proved in this paper: for a fixed number of alternatives, a voting rule is a generalized scoring rule if and only if it is anonymous and finitely locally consistent. It also bounds the order of an equivalent generalized scoring rule from the degree of consistency, and gives a converse bound. In its proof, the authors extend homogeneous profiles from integer vote counts to rational vectors over rankings. That offers a natural continuous-population interpretation: rankings are types and their rational masses replace voter counts. But the theorem classifies rules; it does not analyze the complexity of a computational problem on those populations. Continuizing its axiomatic question would also be outside the programme’s scope.
Theorem 2’s Bucklin bounds, and Theorems 3–6’s bounds for STV, maximin, Copeland, and ranked pairs, likewise concern degree of consistency—not algorithmic complexity. So there is no sound way to write one continuous computational problem per qualifying anchor without inventing results or stretching the paper’s scope. The rational-profile construction is a genuine point of contact with continuous populations, but it does not support the positive computational case requested.
The proponent names Theorem 1 and Theorems 2–6, but none supplies an anchor for a continuous *computational* mirror. Theorem 1 characterizes rules through anonymity and finite local consistency; Theorems 2–6 bound consistency degrees. They establish structural and axiomatic facts, not the complexity of an algorithmic problem. The paper’s computational discussion is motivation: its frequency-of-manipulability results are attributed to earlier work, and its later applications of generalized scoring rules concern summarization and elicitation without a complexity result here. A continuous restatement of these theorems would therefore miss ChoCo’s target.
The stronger version of the proposed mirror fares little better as a mirror of this paper. One might ask for the minimum mass of manipulators needed to make a candidate win, given a distribution over sincere rankings. That is a coherent computational question, and the paper’s discussion of generalized scoring rules and manipulation gives it some motivation. But this paper neither defines that optimization problem nor proves a result about it; it would be a new question built from the paper’s context. The same applies to asking for algorithms to summarize a continuous electorate or determine whether its winner is fixed under unreported preferences. These may be worthwhile projects, but the paper provides no computational result for them to mirror.
The rational-profile construction in Theorem 1 makes the gap sharper: the proof already extends homogeneous profiles to rational masses over rankings. A continuous version of the paper’s actual classification would largely repeat that construction, while still asking an axiomatic question. The consistency-degree bounds likewise describe rule structure, not a population-level algorithm.
This is the strongest negative case, but it does not establish that no worthwhile continuous computational question could grow out of the paper. Voting admits plausible high-multiplicity regimes, and manipulation is a natural computational topic. The narrower conclusion is that the paper itself offers no named computational result to anchor such a mirror; the more promising manipulation-mass formulation would be an adjacent project, not a continuization of one of its results.
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.