| paper | d4-conf-wine-2016-0024 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | high |
| authors would recognise it | yes |
Both anchors are named computational results. But the paper already models a population of infinitesimal buyers through inverse-demand curves and proves these pricing guarantees in that setting; the proposed questions simply restate those results. Under the \(b\)-already rule, there is no discrete-to-continuous bridge left for these anchors.
fails bit b-already — the paper is already native to the continuum
The proposed mirror is already the paper's own continuous-population pricing problem, with the same algorithms and objectives.
fatal: True
The proposed mirror covers Theorems 3.2 and 5.1, which the paper already proves in its continuous market model; it adds no distinct result to bridge. A finite-support complete-type formulation or a finite-to-continuous transfer guarantee would be a separate extension.
The strongest case is the article’s own market model. In a large EV charging market, goods can be time slots, buyers can be drivers who accept particular bundles of slots, and production costs can reflect the cost of supplying each slot. There may be many buyers but relatively few bundle-preference classes. Replacing named buyers with population mass is natural here: the seller still chooses one price per item, buyers still choose their cheapest acceptable bundle, and the questions about profit and social welfare remain intact. The authors would plainly recognize this as their problem; their model already describes many infinitesimal buyers and aggregate demand.
The lead mirror is the Continuous Multiminded Posted-Price Bicriteria Problem. An instance has a finite item set \(T\), preference classes \(i=1,\ldots,\tau\) with desired bundle families \(\mathcal B_i\subseteq 2^T\), inverse-demand curves \(\lambda_i\), and production costs \(C_t\). Normalize total buyer mass to one, scaling costs to match. A posted price vector \(p\) induces an allocation: buyers in each class take a cheapest desired bundle when its price is below their value. The output is \(p\) and a consistent allocation, including the mass assigned to each bundle and the resulting item quantities \(y_t\). Profit is \(\sum_{t\in T}(p_ty_t-C_t(y_t))\); welfare is \(\sum_i\int_0^{x_i}\lambda_i(z)\,dz-\sum_{t\in T}C_t(y_t)\). The question is whether one price vector can approximate both the maximum achievable profit and the maximum social welfare.
Its anchor is Theorem 5.1, proved in this paper. For multi-minded buyers under the paper’s demand and production-cost assumptions, it gives a polynomial-time bicriteria algorithm: the same pricing solution achieves a \(\Theta(\log(\Delta)/(1-\alpha))\)-approximation to optimum profit and a \(\Theta(1/(1-\alpha))\)-approximation to optimum welfare, where \(\Delta=\ell_{\max}/\ell_{\min}\) is the ratio of largest to smallest desired bundle sizes and the demand curves are \(\alpha\)-strongly regular. The welfare factor is independent of \(\Delta\); the proof handles equal-size bundles with a constant number of candidate solutions. This is a Class A result, already established under those assumptions. It is the strongest anchor for the general mirror.
A useful, narrower companion is the Continuous Unit-Demand Posted-Price Bicriteria Problem: use the same instance and objectives, but require every desired bundle to be a singleton. Theorem 3.2, also proved here, gives a polynomial-time solution with revenue factor \(\zeta=2(1/(1-\alpha))^{1/\alpha}+\alpha/(1-\alpha)\) and welfare factor \((2-\alpha)/(1-\alpha)\), for \(\alpha<1\); at \(\alpha=0\), \(\zeta\) is understood by its limit \(2e\). The theorem’s algorithm first computes a welfare-maximizing solution and then floors item prices at \(\lambda_{\max}(1-\alpha)^{1/\alpha}\). This is also Class A. Theorem 5.1 is the lead because it covers general multi-minded buyers; Theorem 3.2 gives the sharper, more explicit unit-demand case.
The paper says in prose that computing profit-maximizing prices is NP-hard, even for unit-demand buyers with \(\alpha=0\). That would be a valuable hardness anchor, but the supplied text gives it no numbered theorem, lemma, corollary, or proposition, so I would not present it as one under your anchoring rule. It does suggest further questions: what guarantees are possible for arbitrary demand curves, where does the \(\alpha=1\) boundary become hard, and when do finite-market instances inherit the continuum’s approximation guarantees?
There is a real qualification about types. A complete buyer type here is \((\mathcal B_i,v)\): buyers with different values are not identical. The curves \(\lambda_i\) compactly encode a distribution over values within a preference class. Thus \(\tau\) counts preference classes, not the support size of complete types; a smooth curve represents a continuum of values. This fits a population-measure formulation, but it does not directly give ChoCo’s finite-support, complete-type encoding. One further result would be needed to connect finite valuation types to the smooth-demand model, for example by bounding how a finite valuation grid changes the pricing guarantees.
The weakest point is also the clearest: this paper already studies the continuous-population pricing problem. So it is excellent evidence that the mirror is sensible and author-recognizable, but the two algorithmic questions above are answered in the paper itself. A new contribution would need to establish the finite-to-continuous dictionary or analyze a distinct regime the theorems leave open; simply presenting these results as a newly discovered mirror would restate the article.
The strongest objection is that both anchors are already continuous-population results in the paper itself. The authors model “many, many infinitesimal buyers,” represent their aggregate demand by inverse-demand curves, and define welfare using integrals over buyer mass. Theorem 5.1 already gives the proposed multi-minded posted-price bicriteria algorithm; Theorem 3.2 already gives its unit-demand counterpart. Recasting either theorem as a continuous-society question would not uncover a missing computational problem. It would restate the problem and answer the paper already provides.
Changing the scenario does not rescue these anchors: the paper itself names EV charging markets and advertising markets, and its pricing model already covers both. Refining a preference class into complete types \((\mathcal B_i,v)\) with rational masses is a closer fit to ChoCo’s finite-support input. But it describes the same aggregate-demand market at a finer input resolution. The paper’s smoothness and \(\alpha\)-strong-regularity assumptions may not cover step demand from finite valuation supports; studying that boundary or proving a finite-support approximation guarantee could be worthwhile. That would be a new input-model or transfer question, though, not a continuous mirror of Theorem 3.2 or 5.1.
So the case against these two anchors is strong as a novelty objection: the paper already does the continuization and computational analysis they propose. It is not evidence that continuous populations are an ill-suited model; under ChoCo’s own criteria, the paper is evidence in their favor. The negative case is weaker against the broader universal claim: a finite-support bridge or an extension beyond the paper’s demand assumptions could still merit work. Those possibilities do not make either named theorem an unanswered mirror.
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.