| paper | d4-conf-atal-2013-0189 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | high |
| authors would recognise it | unclear |
Theorem 3 is a structural existence result, not a computational result under the stated criteria. The equilibrium experiments compute outcomes but provide no named algorithmic guarantee. The proposed atomless question is statable, but its loss of pivotality does not supply the missing computational anchor.
fails bit a — no named computational result to mirror
The proposed limit question covers only pair-support equilibrium existence associated with Theorem 3; it leaves finite pivotality, equilibrium enumeration, empirical outcome patterns, and the other voting mechanisms aside.
The strongest honest case is narrow, and it has an anchor problem: the paper states no numbered result classifying a computational problem as polynomial-time, NP-hard, or similar. Its only numbered theorem, Theorem 3, is proved here, though the paper gives only a proof sketch. It says that in any symmetric Bayesian plurality game with \(n\ge2\), for any pair \(c_1,c_2\), either some positive truthfulness incentive admits a Bayes–Nash equilibrium in which each voter votes for their preferred member of the pair, or one candidate Pareto-dominates the other ex ante. That is a structural existence result, not a computational-complexity result, so it does not strictly meet the requested anchor rule.
As the closest candidate for a lead mirror, consider Atomless Pair-Support Equilibrium. The instance is a finite candidate set \(C\), a finite support \(T\) of complete preference rankings with rational masses \(\mu_t>0\) summing to one, and a pair \(c_1,c_2\). Voters of type \(t\) have the paper’s utility: their rank for the winner, plus a fixed positive bonus for reporting \(t\) truthfully. The question is whether there is a symmetric pure equilibrium in which every type votes for its preferred member of \(c_1,c_2\). A yes solution gives the type-to-report strategy; a no solution identifies a positive-mass type whose top-ranked candidate lies outside the pair.
This is a natural high-multiplicity regime: a large electorate has relatively few shared preference rankings, and \(\mu_t\) records each type’s population share. The paper’s Bayesian model already uses a distribution over preference types, but it still has a finite number of voters whose realized type draws can affect pivotality. Here the society itself is the exact mass distribution. The question keeps the paper’s plurality rule, ranking types, truthful-report bonus, and equilibrium concept.
In the atomless model, though, an individual deviation changes no candidate’s vote mass. So a type can gain the truthfulness bonus without changing the outcome. The pair-supported equilibrium exists exactly when every supported type already ranks one of \(c_1,c_2\) first; then truthful reporting supplies a witness. Otherwise a type with a different top candidate can deviate truthfully and gain the bonus. Checking this condition and reporting the resulting plurality winner takes polynomial time in the support size and candidate count. The expected direction is therefore tractable.
The mirror would cover only Theorem 3’s pairwise equilibrium-existence question, not the paper’s other equilibrium findings. The paper’s enumeration experiments and its support-enumeration method do not supply an additional named complexity result to anchor a second mirror.
The soft spot is also the central one: this continuum limit removes pivotality, which is what makes the paper’s strategic equilibrium claim interesting. The atomless question is recognizable as a limit of their model, but its answer reduces to truthful voting and does not preserve Theorem 3’s substantive phenomenon. A useful follow-up would be to ask how the finite-\(n\) Bayesian equilibria behave as type frequencies converge to \(\mu\), especially as the truthfulness bonus scales against the probability of being pivotal. That could connect the mirror back to the paper’s computation, but it is a further asymptotic question, not an anchor established by this paper.
The paper offers no numbered result classifying the complexity of a computational problem. Theorem 3 is a structural existence claim about finite Bayesian games; the equilibrium counts and outcome patterns are empirical findings, and support enumeration is the method used to obtain them. That leaves a weaker foundation for a continuization project than a result whose algorithmic boundary the mirror could map.
Theorem 3 is the proponent’s closest anchor, but its mechanism is pivotality in a finite electorate. A voter’s report can change the winner, and the positive truthfulness bonus must be small enough for that possibility to outweigh it. In an atomless society with type masses \(\mu_t\), one voter’s deviation changes no vote mass and cannot change the outcome. With any positive truthfulness bonus, truthful reporting is therefore each type’s strict best response. The proposed pair-support question reduces to checking whether every supported type ranks one of the pair first; the masses do not matter. That is a small support check, and it bypasses the result’s central question about whether pivotality can overcome the bonus.
The natural extensions run into the same problem. Keeping uncertainty over the aggregate distribution does not restore an individual’s influence on it. Removing the truthfulness bonus leaves agents indifferent among reports, so the equilibrium set becomes uninformative. Making \(\epsilon\) shrink with electorate size can preserve finite-game pivotal incentives, but the resulting question concerns asymptotics of finite elections and pivotal probabilities. Those probabilities can vanish as \(n\) grows, often exponentially when the limiting vote shares have a strict plurality gap. The model still relies on finite voters to produce its strategic effects.
One could let all voters of a type deviate together as a positive-mass bloc. That restores influence on the winner, but changes the strategic unit: the paper studies individual voters who may each manipulate, not coordinated type cohorts. It would be a different game whose coordination assumption needs its own justification.
So the strongest case against a worthwhile mirror is that preserving the paper’s individual-agent equilibrium model makes atomless deviations powerless, while restoring pivotality requires finite-size effects or coordinated blocs. The negative case is not a proof that no related continuum game could be interesting. It is strongest against a faithful population-continuous mirror of 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.