The Moderating Effect of Instant Runoff Voting

The paper has no numbered result asserting an algorithmic or computational-complexity classification; the cited Dodgson hardness is external and concerns another rule. The proposed exclusion-threshold and winner-distribution problems are recognizable extensions of the continuum-voter model, but they add computational objectives absent from the paper. Thus bit (a) fails, making the grade red.

The objection that survived

The threshold proposal quantifies over arbitrarily many candidate configurations and introduces an exact decision or optimization objective that the paper never studies.

What the mirror covers

The threshold proposal covers Theorem 1 and draws on Theorem 4; the distribution proposal covers Corollary 1 and Propositions 1–2. The remaining structural theorems, Lemma 1, and the stick-breaking analysis receive no computational counterpart.

Open questions for a prover

The case FOR (proponent)

The strongest honest positive case is a conditional near-miss, not a qualifying ChoCo anchor.

The paper contains no numbered theorem, lemma, corollary, or proposition asserting membership in P, NP-hardness, W[1]-hardness, FPT, or any comparable computational classification. Theorem 1, Corollary 1, and Theorems 4–7 are structural or probabilistic results; Lemma 1 and Propositions 1–2 give asymptotic or exact winner distributions. The statement that the number of cases grows exponentially is not a complexity lower bound. The cited NP-hardness of Dodgson’s rule is external and concerns a different voting rule. Thus the paper has zero eligible computational anchors under the programme’s strict rule.

If structural anchors were admissible, my lead would be a computational version of Theorem 1, “Combinatorial moderation for uniform IRV,” supported by Theorem 4’s general exclusion-zone condition.

The natural regime is a large electorate with recurring ideological cohorts: voters are distributed over a one-dimensional political spectrum, while the candidate slate and the number of distinct ranking patterns remain small. For a fixed slate, a voter type is the complete ranking induced by distance to the candidates; its mass is the measure under \(F\) of the positions producing that ranking. A finite election with rational cohort proportions is recovered by replication. This is especially plausible here because it is exactly the authors’ own model, not a new application story.

Define Universal IRV Exclusion Threshold\(_\infty\) as follows. An instance consists of a symmetric voter distribution \(F\) on \([0,1]\), given by a rational piecewise-polynomial density, and a rational \(c\in(0,\tfrac12)\). Let \(I_c=[c,1-c]\). The question is whether, for every finite tie-free candidate set \(X\subseteq[0,1]\), the presence of at least one candidate in \(I_c\) guarantees that no IRV execution can elect a candidate outside \(I_c\). Equivalently, decide whether \(I_c\) is a universal exclusion zone. The associated optimization problem maximizes \(c\), thereby finding the narrowest guaranteed moderating interval.

A YES solution is a proof of the universal exclusion property; a NO solution is a finite candidate configuration and IRV elimination sequence producing an outside winner. For uniform \(F\), Theorem 1 supplies the exact answer \(c^\star=1/6\): \([1/6,5/6]\) works and every narrower symmetric interval fails. Theorem 4 suggests a computational route for general \(F\), since its condition (5) is a one-dimensional inequality involving \(F\). For piecewise-polynomial \(F\), checking that sufficient condition is plausibly Class A: it reduces to exact minimization over finitely many algebraic pieces. The unrestricted exact threshold problem is open and could become continuum-specifically hard if the universal candidate configuration cannot be compressed.

A second, weaker but very natural follow-on is IRV Winner-Distribution\(_\infty\). Given \(F\), \(k\), \(c\), and \(\varepsilon,\delta>0\), sample \(k\) candidates independently from \(F\), form the ranking-type masses induced by each candidate slate, run IRV, and return an estimate of \(\Pr(R_k\in[c,1-c])\) within \(\varepsilon\) with probability \(1-\delta\). This directly computationalizes Corollary 1 and Propositions 1–2. Approximation is plainly Class A by sampling when \(F\) is efficiently sampleable; exact probability computation may be difficult because the candidate space is partitioned into exponentially many IRV regions.

The weakest point is decisive: both proposed problems add computational objectives that the paper does not study. They are recognizable extensions of the authors’ continuous IRV model, but they are not continuizations of a named computational result. They also move some complexity into candidate configurations and geometric integration rather than population multiplicity. So the positive case is that this is an unusually clean source model for a future ChoCo computationalization, with a credible Class-A certificate problem and a potentially interesting exact-distribution boundary. It is not, under the stated admission rule, a paper with a qualifying continuous mirror already anchored in its results.

The case AGAINST (opponent, writing after the proponent)

The proponent’s own “near-miss” diagnosis is decisive under ChoCo’s admission rule: this paper contains no numbered computational result to continuize. Theorems 1 and 4 are structural statements about exclusion zones; Corollary 1 is an asymptotic probability statement; Lemma 1 and Propositions 1–2 concern distributions and densities. None classifies an algorithmic problem. The cited NP-hardness of Dodgson’s rule is external and concerns another rule. Thus the proposed mirrors are new computational papers motivated by this model, not continuous versions of computational results in this paper.

The proposed exclusion-threshold problem is the strongest rescue, but it changes the object substantially. For a fixed candidate slate, the continuous electorate is already completely summarized by finitely many masses of ranking cells: in one-dimensional Euclidean voting, the cells are cut out by pairwise midpoints, and their masses are CDF differences. There is no remaining population-level optimization, pricing problem, or high-multiplicity decision. Theorem 1’s conclusion is already obtained after integrating the electorate into those masses.

To make the threshold problem nontrivial, one must quantify over arbitrarily many candidate configurations and optimize their positions. That is a geometric verification problem over candidate space, not a computational problem exposed by population continuization. Theorem 4 gives a sufficient CDF inequality, but checking that inequality is not the same as computing the exact exclusion threshold; the latter would require a new necessity theorem and a treatment of the unbounded candidate quantifier. Its possible difficulty is therefore speculative and belongs to a new paper, not to a result being mirrored.

The winner-distribution proposal is weaker still. In Corollary 1 and Propositions 1–2, the continuum is already the voter distribution \(F\); the randomness being averaged is in the candidate positions. Candidate count \(k\), not electorate multiplicity, drives the limiting distribution. Monte Carlo estimation is a generic sampling exercise, while exact integration over IRV regions is a candidate-configuration geometry problem. Neither supplies the high-multiplicity bridge central to ChoCo. Replicating the electorate changes nothing because the paper has already passed to its limiting measure.

I am not objecting to the cohort story. A large population of voters sharing one-dimensional ideological preferences is a perfectly sensible high-multiplicity regime. The problem is that the paper uses that regime analytically but never asks a computational question about it.

The best population-centred repair would be to define continuous IRV bribery or robustness: with a fixed candidate slate and distribution \(F\), find the minimum voter mass whose rankings must be changed to make an extreme candidate win. That is genuinely plausible ChoCo material. But it introduces a new intervention model, costs, and objective absent from the paper; it does not continuize Theorem 1 or Corollary 1. It is a generic continuous manipulation problem motivated by the paper’s geometry.

So the negative case is strong and decisive for the programme’s strict screening: there is no qualifying computational anchor here. The absolute claim that no worthwhile ChoCo paper could ever be built from this model is weaker. The bribery/robustness variant could be worthwhile, but it would be a new result inspired by the paper, not a continuous mirror of it.

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.