关于「实证闭环:长寿文案证伪清单、健康寿命规划问卷与预注册方案」,《死亡率倍增与蓝区审计》给出的定量结论是什么?

结论:第八章给出长寿文案证伪清单、非医疗的健康寿命规划问卷(说明为何不能把多项风险比相乘:无因果作用的三种习惯相乘得 0.437)、蒸馏小工具「长寿声称核查表」与预注册检验方案。书中以标注框区分已发表证据、可运行计算(22 段程序)、原典与作者假说。

数据口径:书中参数为示意;全部计算由书中程序生成;本书不是医疗建议,不宣称任何产品或方法的效应,自我试验须先咨询执业医师。 书中以标注框区分已发表证据、可运行计算(22 段程序)、原典与作者假说。

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: 0.437

habit 1: crude rate ratio 0.775 (2.94 years)
habit 2: crude rate ratio 0.750 (3.32 years)
habit 3: crude rate ratio 0.753 (3.28 years)
naive product of three0.437 (9.54 years)
observed all-three vs none0.569 (6.51 years)
true causal effect of the habits1.000 (0.00 years)
查看完整验算程序
# Chapter 8. p83: Why a planning questionnaire must not multiply HRs.
# Synthetic cohort of 100,000 adults followed 10 years (fixed seed).
# A latent health factor z ~ N(0, 1) sets the hazard 0.02 exp(-0.3 z);
# three habits each switch on when 0.6 z + 0.8 e_i > 0, e_i ~ N(0, 1),
# and have NO causal effect. We estimate each habit's crude rate
# ratio (deaths per person-year, habit vs no habit), multiply the three
# as a naive questionnaire would, convert to years with B = ln2/8, and
# compare with the rate ratio actually seen for all three vs none.
# All parameters are illustrative (示意).
import math
import random

rng = random.Random(20261112)
N, H0, BETA, FU = 100000, 0.02, -0.3, 10.0
B8 = math.log(2) / 8.0
rows = []
for _ in range(N):
    z = rng.gauss(0.0, 1.0)
    habits = tuple(1 if 0.6 * z + 0.8 * rng.gauss(0.0, 1.0) > 0 else 0
                   for _ in range(3))
    t = -math.log(1.0 - rng.random()) / (H0 * math.exp(BETA * z))
    rows.append((habits, min(t, FU), 1 if t < FU else 0))


def rate(sel):
    d = len([r for r in sel if r[2]])
    py = math.fsum(r[1] for r in sel)
    return d / py


prod = 1.0
for i in range(3):
    on = rate([r for r in rows if r[0][i] == 1])
    off = rate([r for r in rows if r[0][i] == 0])
    prod *= on / off
    print("habit {}: crude rate ratio {:.3f} ({:.2f} years)".format(
        i + 1, on / off, -math.log(on / off) / B8))
allon = rate([r for r in rows if r[0] == (1, 1, 1)])
none = rate([r for r in rows if r[0] == (0, 0, 0)])
print("naive product of three: {:.3f} ({:.2f} years)".format(
    prod, -math.log(prod) / B8))
print("observed all-three vs none: {:.3f} ({:.2f} years)".format(
    allon / none, -math.log(allon / none) / B8))
print("true causal effect of the habits: 1.000 (0.00 years)")

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《冈珀茨死亡率倍增定律、超级百岁老人年龄误报与长寿蓝区数据审计母典》第 8 章「实证闭环:长寿文案证伪清单、健康寿命规划问卷与预注册方案」(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-193 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026193,
  author    = {Fan, Yafu},
  title     = {Gompertz Mortality Doubling, Age Misreporting among Supercentenarians and Data Audits of Longevity Blue Zones},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论