关于「漂移的复利:串行复现、p 的 k 次方与修复回流」,《衍生品的出处与保真》给出的定量结论是什么?

结论:在独立误差、永不修复的假设下,每层保真 0.95 的接力在第五层只剩 0.7738;这个数字是一个警告,而不是定律。只要有一点修复,长链就停在正的水平上;但修复本身需要回到母书去对。走样又往往有方向,取平均救不了它。剩下的问题是:回到母书去对,具体对什么?一段文字和母书「相同」,应该在什么意义上相同?下一章用出处哈希回答。

数据口径:书中参数为示意,代价用记账点数而非价格;全部计算由书中程序生成。 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。

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: 0.7738, 0.95

share of facts intact after k layers (independent errors)
k p0.99 p=0.97 p=0.95 p=0.90
10.9900 0.9700 0.9500 0.9000
20.9801 0.9409 0.9025 0.8100
30.9703 0.9127 0.8574 0.7290
40.9606 0.8853 0.8145 0.6561
50.9510 0.8587 0.7738 0.5905
60.9415 0.8330 0.7351 0.5314
expected intact facts of 40 at layer 5:
38.0434.35 30.95 23.62
chance all 40 facts survive layer 3:
0.29940.0259 0.0021 0.0000
layers until the intact share drops below 0.80:
238 5 3
查看完整验算程序
# Chapter 1. p11: fidelity compounds as p**k along a chain of derivatives.
# Model: a master book carries N = 40 checkable facts. Each compilation
# step (book -> summary -> script -> slides -> quiz -> clip) copies each
# fact correctly with probability p, independently, and never repairs
# an earlier error. Layer k then keeps a fact intact with probability
# p**k, and keeps all N facts intact with probability p**(N*k).
# The p values are illustrative, not measured.
import math

N = 40
PS = (0.99, 0.97, 0.95, 0.90)
print("share of facts intact after k layers (independent errors)")
print("k  " + "".join("   p=%.2f" % p for p in PS))
for k in range(1, 7):
    print("%d  " % k + "".join("   %6.4f" % (p ** k) for p in PS))
print("expected intact facts of %d at layer 5:" % N)
print("   " + "".join("   %6.2f" % (N * p ** 5) for p in PS))
print("chance all %d facts survive layer 3:" % N)
print("   " + "".join("   %6.4f" % (p ** (N * 3)) for p in PS))
print("layers until the intact share drops below 0.80:")
cut = [math.floor(math.log(0.80) / math.log(p)) + 1 for p in PS]
print("   " + "".join("   %6d" % c for c in cut))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《母书多模态衍生编译与版权链、单部专著到播客投影片题库的结构同态流水线母典》第 1 章「漂移的复利:串行复现、p 的 k 次方与修复回流」(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-268 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026268,
  author    = {Fan, Yafu},
  title     = {Multimodal Derivative Compilation and Provenance Chains: A Structure-Preserving Pipeline from One Monograph to Podcasts, Slides and Quiz Banks},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论