关于「博尔达计分与无关选项:二十一名选举人、诱饵方案与搅局者」,《不存在完美裁决器》给出的定量结论是什么?

结论:第三章重算博尔达的二十一名选举人,并发现三评审四选项时 0.2604 的剖面存在搅局者。

数据口径:多道检查意见不一时,经营者最常见的想法是:只要合成规则设计得足够聪明,总能得到一个公正、一致、无可挑剔的裁决。 阿罗 1950 年与 1951 年证明,选项不少于三个时,不存在同时满足无限制域、弱帕累托、无关选项独立与非独裁的排序合成规则。

English version

对照数据表

下表数字由本章 Python 程序原样打印,可重新运行复现。 程序只用 Python 标准库与结论吻合的打印值: 0.2604

3 reviewers, 4 options13,824 profiles
Borda : decided 13,824 spoilable 3,600 share 0.2604
plurality: decided 13,824 spoilable0 share 0.0000
Condorcet: decided 12,288 spoilable0 share 0.0000
5 reviewers, 3 options7,776 profiles
Borda : decided7,776 spoilable 1,020 share 0.1312
plurality: decided7,776 spoilable 2,880 share 0.3704
Condorcet: decided7,236 spoilable 0 share 0.0000
a Condorcet winner survives any deletion of a rival
查看完整验算程序
# Chapter 3. p33: how often does deleting a losing option change the
# winner? Exact census over all 24 ** 3 = 13,824 profiles of 3
# reviewers ranking 4 options, and all 6 ** 5 = 7,776 profiles of 5
# reviewers ranking 3 options (impartial culture weights). For each
# rule we take the set of tied winners; a profile is "spoilable" if
# deleting some option outside that set changes the set. The Condorcet
# winner, where one exists, is checked the same way.
import itertools
from fractions import Fraction


def restrict(prof, keep):
    return [tuple(c for c in o if c in keep) for o in prof]


def winners_borda(prof):
    m = len(prof[0])
    s = {}
    for o in prof:
        for pos, c in enumerate(o):
            s[c] = s.get(c, 0) + m - 1 - pos
    top = max(s.values())
    return frozenset(c for c in s if s[c] == top)


def winners_plurality(prof):
    s = {c: 0 for c in prof[0]}
    for o in prof:
        s[o[0]] += 1
    top = max(s.values())
    return frozenset(c for c in s if s[c] == top)


def condorcet(prof):
    opts = prof[0]
    for x in opts:
        if all(2 * len([o for o in prof if o.index(x) < o.index(y)])
               > len(prof) for y in opts if y != x):
            return frozenset([x])
    return None


def spoilable(rule, prof):
    w = rule(prof)
    if w is None:
        return None
    opts = prof[0]
    for d in opts:
        if d in w:
            continue
        keep = [c for c in opts if c != d]
        if rule(restrict(prof, keep)) != w:
            return True
    return False


if __name__ == "__main__":
    for n, opts in ((3, "ABCD"), (5, "ABC")):
        orders = list(itertools.permutations(opts))
        profs = list(itertools.product(orders, repeat=n))
        print("{} reviewers, {} options: {:,} profiles".format(
            n, len(opts), len(profs)))
        for name, rule in (("Borda", winners_borda),
                           ("plurality", winners_plurality),
                           ("Condorcet", condorcet)):
            hit = have = 0
            for pr in profs:
                r = spoilable(rule, pr)
                if r is None:
                    continue
                have += 1
                if r:
                    hit += 1
            f = Fraction(hit, have)
            print("  {:<9}: decided {:>6,}  spoilable {:>5,}  share"
                  " {:.4f}".format(name, have, hit, float(f)))
    print("a Condorcet winner survives any deletion of a rival")

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《阿罗不可能定理与多评审规则聚合、五件套拒绝码冲突时为何不存在完美裁决器母典》第 3 章「博尔达计分与无关选项:二十一名选举人、诱饵方案与搅局者」(版本 1.0.0). 5D Library. DOI 待发布

DOI 待发布 · 专著页正式发布后上线(本站暂无专著页) · 5DL-B-310 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026310,
  author    = {Fan, Yafu},
  title     = {Arrow's Impossibility Theorem and the Aggregation of Multiple Reviewers: Why No Perfect Arbiter Exists When Kit Refusal Codes Conflict},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论