关于「实践与证伪方案:连载主事者的检查清单、攻心文案检查器与预注册检验」,《长篇叙事的因果图、悬念熵与伏笔回收》给出的定量结论是什么?

结论:第八章给出检查清单、八条规则的攻心文案检查器与预注册方案:比较钩子排期每组需要 3,044 名读者,评价单个钩子每组需要 583,531 名起始读者。书中以标注框区分已发表的证据、可运行的 Python 计算(21 段程序)、原典与作者的假说。

数据口径:书中的连载、读者与参数均为示意模型;全部计算由书中程序生成;本书不是法律、财务或投资建议,并坚持一条红线:不得以虚假悬念骗取点击。 第一、二章用卡恩拓扑排序与按键分组的版本检查找出时间悖论与同一属性的两个值,并说明跨卷吃书藏在推得出的先后里:示意连载第六卷的一条边与写在第 2 卷到第 6 卷之间的 10 条边相冲突。

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: 3044, 583531

z for alpha/21.9600, z for power = 0.8416
comparison p1 p2 readers per arm
kept@30, k3 vs k = 4 (p41) 0.6290 0.5940 3,044
kept@30, k3 vs k = 1 (p41) 0.6290 0.5040 246
ch 50 churn, x0.9 (p62)0.0150 0.0135 98,002
S(49)0.1679; starters per arm for the ch 50 test: 583,531
with2,000 per arm, smallest detectable kept@30 gap: 0.0433
with5,000 per arm, smallest detectable kept@30 gap: 0.0273
with 20,000 per arm, smallest detectable kept@30 gap0.0136
查看完整验算程序
# Chapter 8. p82: sample sizes for a pre-registered hook-spacing test.
# Two-sided two-proportion z test, alpha 0.05, power 0.80, readers
# randomized at chapter 1 to two hook spacings. The proportions are taken
# from p41 (share kept at chapter 30) and p62 (churn at chapter 50 among
# readers who got there, which needs the starters scaled up by the
# shifted-beta-geometric survival S(49) with a = 0.8, b = 6). Normal
# approximation, standard library NormalDist. Inputs are model outputs.
import math
from statistics import NormalDist

Z = NormalDist()
ZA, ZB = Z.inv_cdf(0.975), Z.inv_cdf(0.80)


def n_per_arm(p1, p2):
    pb = (p1 + p2) / 2
    num = (ZA * math.sqrt(2 * pb * (1 - pb))
           + ZB * math.sqrt(p1 * (1 - p1) + p2 * (1 - p2)))
    return math.ceil(num ** 2 / (p1 - p2) ** 2)


def s49(a=0.8, b=6.0, t=49):
    return math.exp(math.lgamma(b + t) - math.lgamma(a + b + t)
                    - math.lgamma(b) + math.lgamma(a + b))


print("z for alpha/2 = {:.4f}, z for power = {:.4f}".format(ZA, ZB))
print("comparison                          p1      p2   readers per arm")
rows = [("kept@30, k = 3 vs k = 4 (p41)", 0.629, 0.594),
        ("kept@30, k = 3 vs k = 1 (p41)", 0.629, 0.504),
        ("ch 50 churn, x0.9 (p62)", 0.0150, 0.0135)]
for label, a, b in rows:
    print("{:<33}  {:.4f}  {:.4f}  {:>15,}".format(label, a, b,
                                                    n_per_arm(a, b)))
n50 = n_per_arm(0.0150, 0.0135)
print("S(49) = {:.4f}; starters per arm for the ch 50 test: {:,}".format(
    s49(), math.ceil(n50 / s49())))
for n in (2000, 5000, 20000):
    lo = 0.629
    hi = next(d / 10000 for d in range(1, 6000)
              if n_per_arm(lo, lo - d / 10000) <= n)
    print("with {:>6,} per arm, smallest detectable kept@30 gap: {:.4f}"
          .format(n, hi))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《长篇叙事因果有向无环图校验、章节悬念信息熵锯齿调度与伏笔遗忘衰减回收》第 8 章「实践与证伪方案:连载主事者的检查清单、攻心文案检查器与预注册检验」(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-178 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026178,
  author    = {Fan, Yafu},
  title     = {Causal-DAG Validation of Long-Form Narrative, Sawtooth Scheduling of Chapter Suspense Entropy and Forgetting-Aware Foreshadowing Recovery},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论