| paper | d4-conf-wine-2005-0083 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | high |
| authors would recognise it | yes |
The proposed type-mass local-stability question is concrete, but the proponent acknowledges that Proposition 1 and Proposition 2 are analytic viability conditions, not computational results. The spectral-radius calculation is a computational follow-on, not a named computational result of this paper, so the required anchor gate fails.
fails bit a — no named computational result to mirror
The proposed threshold concerns the local instability mechanism behind Proposition 1, with the potentially contributing population fixed; it does not cover Proposition 2's incentive condition, equilibrium-size computation, or the simulated mixing results.
The strongest positive case is for a high-multiplicity version of the paper’s club-viability model. There is an important qualification: the paper contains no named computational result of the kind your anchor rule requires. Its two numbered propositions are analytic viability conditions; they say nothing about running time or computational complexity.
Proposition 1 (Sufficient viability condition) is proved in this paper. It says that
\[
N(0)\sum_s g(s)h(s)>(\bar{k}\rho(0+))^{-1}>0
\]
is sufficient for the empty club to be unstable, and hence for the club to be viable with a positive equilibrium size satisfying \(\bar P(n)N(n)=n\).
Proposition 2 (Necessary viability condition) is also proved here. It says that \(\rho'(0)\phi_0>0\) is necessary for viability: the incentive system must provide a positive initial response, and the club must have seed content. Neither proposition classifies an algorithmic problem, so neither is a qualifying computational anchor.
A plausible follow-on problem is High-Multiplicity ISC Local Viability. Consider a large file-sharing service with population scale \(M\), divided into \(\tau\) complete peer types. Type \(\theta\) has rational population fraction \(\mu_\theta\), demand distribution \(h_\theta\) over a finite goods set \(S\), supply distribution \(g_\theta\), contribution amount \(k_\theta\), and low-content search response \(a_\theta\). Here \(\sum_\theta\mu_\theta=1\). Within a type, peers match on every behavioral feature the model uses. The regime is credible where millions of users fall into a modest number of recurring demand, supply, and incentive profiles; it becomes less credible if those profiles are nearly all unique.
Let \(x_\theta\) be the participating fraction of type \(\theta\), and let the amount of good \(s\) in the club be
\[
q_s(x)=M\sum_\eta \mu_\eta k_\eta g_\eta(s)x_\eta.
\]
With no seed content, type \(\theta\)’s joining probability is
\[
p_\theta(x)=\sum_s h_\theta(s)\bigl(1-e^{-a_\theta q_s(x)}\bigr),
\]
and the aggregate joining and leaving dynamics are
\[
\dot x_\theta=p_\theta(x)-x_\theta.
\]
The computational question is: given these rational inputs, is the empty-club equilibrium linearly unstable? The solution is a yes/no answer. Its Jacobian at zero is \(B-I\), where
\[
B_{\theta\eta}=M\mu_\eta k_\eta a_\theta
\sum_s h_\theta(s)g_\eta(s).
\]
Thus the question is whether the spectral radius of \(B\) exceeds \(1\). For a single homogeneous type, this recovers the threshold underlying Proposition 1. For multiple types, it generalizes that local test. I would expect this problem to be tractable: construct \(B\) and compare its Perron root with \(1\), with input size governed by \(\tau\), \(|S|\), and the rational bit length.
This is a reasonable population mirror of the paper’s local viability question: it replaces individual peers with masses of behaviorally identical peers and asks for an algorithmic answer to the same empty-club instability issue. It covers the local mechanism behind Proposition 1, with the set of potentially contributing peers fixed. It does not cover Proposition 2’s question of how incentives create that set, or the paper’s club-mixing simulations. Natural next questions would be whether the local threshold characterizes nonlinear viability at the boundary, how to compute positive equilibria, and how sensitive the answer is to changing type masses.
The weak point is substantial: this is a proposed computational follow-on, not a continuous mirror of a named complexity result in the paper. The paper already studies statistical membership equilibria, but uses a finite peer population and focuses on analytical viability. Also, the problem is about peer-to-peer networks rather than social choice: it has no voting rule, collective ranking, or social-choice outcome. So the type-mass model is plausible to these authors as a continuation of their work, but it does not make this paper a strong ChoCo case.
The high-multiplicity regime is plausible here: a large sharing service could contain many peers with the same demand, supply, and incentive profile. The stronger negative case is that this does not make the paper a useful ChoCo mirror. Its two propositions are analytic viability conditions, not complexity results, and the paper formulates no algorithmic problem about choosing or manipulating a collective outcome.
The proposed local-viability anchor is well-defined, but it extends Proposition 1 by differentiating a type-level mean-field system. With finitely many types, that produces the matrix \(B\) and asks whether its spectral radius exceeds \(1\). This is a sensible calculation about the local stability of an empty club. It is not a computational social-choice problem: there is no social-choice rule or outcome, and no intervention problem in the paper for an algorithm to solve. The mirror also fixes the potentially contributing population and studies only the local boundary, while the paper notes that this population can change with club size. Restoring that dependence makes the model richer, but keeps the question in equilibrium analysis.
The more ambitious mirrors fare no better as ChoCo anchors. One could ask for a positive stable equilibrium in a multitype model, or for the cheapest incentives that make a club viable. The first is a computational version of the paper’s fixed-point analysis. The second is the closest analogue to bribery, and the discussion does motivate incentive design. But the paper does not specify intervention costs or how interventions change each type’s contribution response. Supplying those choices could define a worthwhile network-economics problem; its complexity would belong to that new model, rather than to a computational result in this paper. Likewise, the club-mixing section offers simulations, not a formal problem whose complexity can be mirrored.
The paper’s use of distributions over goods and statistical equilibrium is not itself an objection: its agents remain a finite peer population, so a population-level type model would be a real modeling change. But the resulting questions would be about peer participation and community dynamics. A social-choice version would require adding a collective decision mechanism, which would change the subject.
So the case against is strongest as a scope judgment, not as a proof that no useful continuous computation can be formulated. A type-based viability model could be worth studying in its own field. The weak point in the negative case is precisely that concession: the high-multiplicity setting is credible, and the paper’s incentive discussion could motivate a new computational project. What it does not supply is a computational social-choice result for ChoCo to continuize.
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.