| paper | d4-conf-ecai-2012-0166 |
| 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 computational, but the proposed mass assignment is already the paper’s rational-weight PWUW model: fixed unit-weight votes supply baseline mass, and eligible votes receive rational mass. For scoring rules, Corollary 8 already gives the continuous-side result, so this is a model collision rather than a new continuous-population question.
fails bit b-already — the paper is already native to the continuum
The paper already formulates the proposed \(Q^+\) scoring problem over its listed votes with rational weight assignments; the proponent notes this overlap but does not distinguish it from the proposed mass model.
fatal: True
The proposed mirror covers Theorem 2’s bounded integer-weight \(k\)-approval variants and Corollary 8’s rational-weight scoring variants; it does not cover the paper’s Copeland or ranked-pairs cases.
The strongest lead is Theorem 2, proved in the paper. It says that for every \(k\ge4\), both \(k\)-approval-PWUW-BW-N and \(k\)-approval-PWUW-BW-RW-N are NP-complete. In these variants, voters in \(V_1\) have fixed unit weight; voters in \(V_0\) receive nonnegative integer weights whose sum is bounded by \(B\). The range variant also restricts each voter’s weight to a specified interval. The paper proves hardness by reduction from EXACT COVER BY 3-SETS, first for 4-approval and then for every fixed \(k>4\).
A plausible regime is a very large consultation or election where participants fall into a small number \(\tau\) of ranking blocs. Some blocs’ represented mass is fixed; for others, the amount of participation or representation is uncertain, but can be chosen within a total cap. A type records the complete ranking and the bloc’s eligibility and weight constraints. The number of participants can be enormous while \(\tau\) remains modest.
Call the mirror Continuous \(k\)-Approval-PWUW-BW. Its input is a candidate set \(C\), \(k\ge4\), a target \(c^\star\), a fixed baseline distribution \(\nu\) over ranking types with total mass \(1\), a finite set \(U\) of eligible types, and a rational mass bound \(B\). Optionally, each eligible type \(t\) has a rational interval \([\ell_t,u_t]\). The question is whether there are masses \(x_t\ge0\), with \(\sum_{t\in U}x_t\le B\) and the specified interval bounds, such that \(c^\star\) is a co-winner. The resulting society is the normalized distribution formed from the baseline and the added masses; normalization does not change the winners. A solution is the vector \(x\).
Writing \(s_k(t,a)=1\) when \(a\) is among the top \(k\) positions of ranking \(t\), the winning condition is precisely, for every \(d\ne c^\star\),
\[
\sum_t \nu_t\bigl(s_k(t,c^\star)-s_k(t,d)\bigr)
+\sum_{t\in U}x_t\bigl(s_k(t,c^\star)-s_k(t,d)\bigr)\ge0.
\]
This is a linear feasibility problem with rational data, \(O(\tau)\) variables, and \(O(m+\tau)\) constraints. It is therefore tractable. I would classify this as Class A: the integer hardness can dissolve when bloc sizes become fractional masses. The reduction’s all-or-nothing choices among set-coded votes are exactly the feature the continuous feasible region removes.
A second, broader anchor is Corollary 8, also proved in the paper. It says that for every scoring rule, all four rational-weight variants—PWUW-Q+, PWUW-BW-Q+, PWUW-RW-Q+, and PWUW-BW-RW-Q+—are in P. Its continuous mirror, Continuous Scoring-PWUW, uses the same baseline, eligible ranking types, mass assignment, and optional bounds and intervals, replacing \(s_k\) with the score from an input scoring vector. The question is whether some permitted mass assignment makes \(c^\star\) a co-winner; the assignment is the solution. The scoring comparisons are linear inequalities, so the expected direction is again tractability by LP. Theorem 2 is my lead because it gives the hardness result whose fate under continuization is the central question; Corollary 8 shows that the paper already establishes the continuous-side answer across scoring rules.
The bridge is unusually direct here. The paper defines a vote of integer weight \(w\) as equivalent to \(w\) unit-weight votes. Thus integer weights can represent integral numbers of agents in a ranking bloc, while rational weights represent fractional bloc mass. Repeated votes with identical rankings and constraints can be grouped into support types. The paper’s own LP result makes this a mirror the authors would recognize, rather than a simplified problem that changes what is being decided.
The scope is the bounded, uncertain-weight \(k\)-approval results of Theorem 2 and their rational-weight scoring-rule counterparts in Corollary 8. It does not cover the paper’s Copeland or ranked-pairs results, nor does it claim a result for unique winners or for uncertainty about rankings as well as weights.
The weakest point is also the case’s clearest limitation: the paper already studies rational voter weights, which, after grouping identical ranking types, are very close to the continuous population model above. So this is strong evidence that the mirror is sensible, and that integer hardness can disappear in the continuous version; it is not a new mirror for this paper to discover. The next questions would be whether solutions round to large finite electorates with controlled error, how the answer changes under unique-winner requirements, and whether allowing both ranking and mass uncertainty preserves tractability.
The strongest case against both anchors is an exact model collision. In the paper’s \(Q^+\) variants, each uncertain vote already receives a nonnegative rational weight. Treat fixed unit-weight votes as baseline mass and uncertain votes as eligible types. Votes with the same ranking and the same constraints can be grouped; their assigned weights add. The resulting weighted ranking distribution can be normalized without changing the winner. That is the proposed mass-assignment model in different notation.
For Theorem 2, the proposed continuous bounded \(k\)-approval question is already covered by Corollary 8’s \(Q^+\) bounded and range-restricted variants. Theorem 2’s integral weights make discrete choices; replacing them with rational masses gives the very problem this paper already formulates and solves for scoring rules. The point against the mirror is not that its answer is easy or differs from the integer answer. It is that no new continuous computational question remains to pose.
Corollary 8 is an even clearer collision: its rational-weight scoring problem is already the proposed continuous scoring mirror, including the optional total bound and intervals. The proponent’s description correctly identifies the LP, but that also shows the proposed model and result are already in the paper. Rounding, unique-winner requirements, or joint uncertainty about rankings would be new questions, not mirrors of these two anchors.
That defeats the two anchors as *new* mirrors, but it does not establish the universal claim that no worthwhile continuous mirror exists. The paper’s rational-weight variants for rules such as Copeland and ranked pairs are formulated too, though their complexity is not settled here. Work on them could be valuable, but it would resolve open cases in the paper’s existing continuous-weight framework rather than introduce a new mirror. So the negative case is strongest on novelty: this paper has already made the continuization move. It is weaker if “worthwhile” includes solving those existing open continuous-weight problems.
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.