关于「东西方原典考辨:孙子论「气」与格朗特的死亡表」,《直播间留存的生存分析》给出的定量结论是什么?

结论:第七章考辨《孙子·军争》「朝气锐,昼气惰,暮气归」与格朗特 1662 年的死亡表(开头六年风险是中段的 1.46 倍)。

数据口径:书中的直播间、观众数、风险参数、钩子效果与节目单均为示意;全部计算由书中程序生成;本书是非个人化经营建议,钩子只能是真实的内容或写明真实库存与真实截止时刻的福利,不用会重置的倒计时,不诱导未成年人消费,平台规则以官方公示为准。 第一章说明截尾让平均停留偏低:示意直播间里把下播当离开得 408.1 秒,只算已离开者得 278.0 秒,乘积极限估计的限定平均停留是 337.9 秒。

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: 1.46

agesdeathsratiohazard/yr
0-6360.6400.0744
6-16240.6250.0470
16-26150.6250.0470
26-3690.6400.0446
36-4660.6250.0470
46-5640.6000.0511
56-6630.5000.0693
66-7620.3330.1099
mean hazard ages 6-66: 0.0510opening 0-6: 0.0744ratio 1.46
expected years lived before 76: 18.17
if ages 0-6 had the later hazard: 73.6 alive at 6, 20.45 years
查看完整验算程序
# Chapter 7. p72: Graunt's 1662 table read as a hazard curve.
# Survivors out of 100 at ages 0, 6, 16 and on to 76, as printed in
# Graunt's Observations (the infant figure was his estimate and the
# middle ages were filled in by a near-geometric rule). For each
# interval: deaths, the survival ratio, and the constant yearly hazard
# that reproduces it, -ln(S2 / S1) / years. Then the expected years
# lived before 76 (trapezoid rule) and the same quantity if the opening
# interval had the hazard of the later ones.
import math

AGES = [0, 6, 16, 26, 36, 46, 56, 66, 76]
ALIVE = [100, 64, 40, 25, 16, 10, 6, 3, 1]


def hazards(ages, alive):
    out = []
    for i in range(len(ages) - 1):
        yrs = ages[i + 1] - ages[i]
        r = alive[i + 1] / alive[i]
        out.append((ages[i], ages[i + 1], alive[i] - alive[i + 1], r,
                    -math.log(r) / yrs))
    return out


def years_lived(ages, alive):
    parts = [(alive[i] + alive[i + 1]) / 2.0 * (ages[i + 1] - ages[i])
             for i in range(len(ages) - 1)]
    return math.fsum(parts) / alive[0]


rows = hazards(AGES, ALIVE)
print(" ages     deaths  ratio  hazard/yr")
for a, b, d, r, h in rows:
    print("%3d-%-3d %8d  %5.3f  %8.4f" % (a, b, d, r, h))
later = [h for _, _, _, _, h in rows[1:7]]
mid = math.fsum(later) / len(later)
print("mean hazard ages 6-66: %.4f   opening 0-6: %.4f   ratio %.2f"
      % (mid, rows[0][4], rows[0][4] / mid))
print("expected years lived before 76: %.2f" % years_lived(AGES, ALIVE))
s6 = 100 * math.exp(-6 * mid)
flat = [100, s6] + [s6 * a / 64.0 for a in ALIVE[2:]]
print("if ages 0-6 had the later hazard: %.1f alive at 6, %.2f years"
      % (s6, years_lived(AGES, flat)))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《直播间留存卡普兰迈耶生存曲线、划走风险尖峰定位与福利钩子插入秒数母典》第 7 章「东西方原典考辨:孙子论「气」与格朗特的死亡表」(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-277 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026277,
  author    = {Fan, Yafu},
  title     = {Kaplan-Meier Retention Curves for Livestream Rooms: Locating Swipe-Away Hazard Spikes and Timing the Insertion of Hooks},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论