关于「所需模型数与样本量:上限之下的票数账与标注账」,《相关误差下的多模型投票》给出的定量结论是什么?
结论:相关之下,票数账里出现了大片的 never:目标越高,能容忍的相关越低,越过门槛之后多调用多少次都无济于事。标注账则提醒店主,所有这些判断都建立在估计值上,估计值的精度按题量的平方根改善;三百道题是起点,一千道题是靠近门槛时的复核量。下一章暂时离开经营现场,回到两部原典:《尚书·洪范》里「三人占,则从二人之言」的多数决,以及孔多塞 1785 年的《论文》。
数据口径:书中参数为示意;全部计算由书中程序生成;本书是非个人化的经营分析,不构成经营建议。 书中以标注框区分已发表的证据、可运行的 Python 计算(19 段程序)、原典与作者的假说。
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
| A. smallest odd n reaching the target, p = 0.70 | ||||||
| rho | 0.80 | 0.85 | 0.90 | 0.95 | 0.99 | ceiling |
| 0.0 | 5 | 7 | 9 | 17 | 31 | 1.0000 |
| 0.01 | 5 | 7 | 11 | 19 | 47 | 1.0000 |
| 0.02 | 5 | 7 | 11 | 23 | 91 | 0.9979 |
| 0.05 | 5 | 9 | 19 | 85 | never | 0.9649 |
| 0.1 | 7 | 15 | never | never | never | 0.8996 |
| 0.2 | 21 | never | never | never | never | 0.8180 |
| B. the same at p = 0.80 | ||||||
| rho | 0.80 | 0.85 | 0.90 | 0.95 | 0.99 | ceiling |
| 0.0 | 1 | 3 | 5 | 7 | 13 | 1.0000 |
| 0.01 | 1 | 3 | 5 | 7 | 15 | 1.0000 |
| 0.02 | 1 | 3 | 5 | 7 | 17 | 1.0000 |
| 0.05 | 1 | 3 | 5 | 9 | 39 | 0.9973 |
| 0.1 | 1 | 3 | 5 | 17 | never | 0.9755 |
| 0.2 | 1 | 3 | 15 | never | never | 0.9202 |
查看完整验算程序
# Chapter 6. p61: how many models a target needs once errors correlate.
# Model: exchangeable Beta-binomial (p21) at mean accuracy p = 0.70.
# For each target, the smallest odd n whose majority reaches it, found
# by a step-2 search up to 3001, or "never" when the ceiling
# P(theta > 1/2) does not exceed the target. Panel B repeats it at
# p = 0.80. Correlations and targets are illustrative.
import math
def lbeta(a, b):
return math.lgamma(a) + math.lgamma(b) - math.lgamma(a + b)
def ab(p, rho):
s = 1 / rho - 1
return p * s, (1 - p) * s
def bb_major(n, p, rho):
if rho == 0:
lp, lq = math.log(p), math.log(1 - p)
terms = [math.exp(math.lgamma(n + 1) - math.lgamma(k + 1)
- math.lgamma(n - k + 1) + k * lp
+ (n - k) * lq)
for k in range(n // 2 + 1, n + 1)]
return math.fsum(terms)
a, b = ab(p, rho)
terms = [math.exp(math.lgamma(n + 1) - math.lgamma(k + 1)
- math.lgamma(n - k + 1) + lbeta(k + a, n - k + b)
- lbeta(a, b))
for k in range(n // 2 + 1, n + 1)]
return math.fsum(terms)
def betacf(a, b, x):
tiny = 1e-300
c, d = 1.0, 1.0 - (a + b) * x / (a + 1)
d = 1.0 / (d if abs(d) > tiny else tiny)
h = d
for m in range(1, 400):
m2 = 2 * m
for num in (m * (b - m) * x / ((a + m2 - 1) * (a + m2)),
-(a + m) * (a + b + m) * x / ((a + m2) * (a + m2 + 1))):
d = 1.0 + num * d
d = 1.0 / (d if abs(d) > tiny else tiny)
c = 1.0 + num / c
c = c if abs(c) > tiny else tiny
h *= d * c
if abs(d * c - 1.0) < 1e-15:
break
return h
def betai(a, b, x):
if x <= 0:
return 0.0
if x >= 1:
return 1.0
lead = math.exp(a * math.log(x) + b * math.log(1 - x) - lbeta(a, b))
if x < (a + 1) / (a + b + 2):
return lead * betacf(a, b, x) / a
return 1.0 - lead * betacf(b, a, 1 - x) / b
def limit(p, rho):
if rho == 0:
return 1.0 if p > 0.5 else 0.0
a, b = ab(p, rho)
return 1.0 - betai(a, b, 0.5)
def calls(p, rho, target, cap=3001):
if limit(p, rho) <= target:
return "never"
n = 1
while bb_major(n, p, rho) < target - 1e-12 and n < cap:
n += 2
return str(n) if n < cap else ">" + str(cap)
def table(p):
print(" rho 0.80 0.85 0.90 0.95 0.99 ceiling")
for rho in (0.0, 0.01, 0.02, 0.05, 0.1, 0.2):
cells = " ".join("{:>5}".format(calls(p, rho, t))
for t in (0.8, 0.85, 0.9, 0.95, 0.99))
print(" {:<6} {} {:.4f}".format(rho, cells, limit(p, rho)))
if __name__ == "__main__":
print("A. smallest odd n reaching the target, p = 0.70")
table(0.7)
print("B. the same at p = 0.80")
table(0.8)唯一数据源与引用
范雅福(Fan, Y.). (2026). 《孔多塞陪审团定理与相关误差、多模型投票何时越投越准何时同源共错母典》第 6 章「所需模型数与样本量:上限之下的票数账与标注账」(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-309 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026309,
author = {Fan, Yafu},
title = {Condorcet's Jury Theorem with Correlated Errors: When Multi-Model Voting Gets More Accurate and When Shared Origins Make Everyone Wrong Together},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}