关于「跨卷吃书:增量环检测、传递闭包与作者自己的遗忘」,《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:跨卷吃书有两个来源。一个是推得出的先后:示意连载在第六卷由一条看似无害的边闭合了一个环,与它冲突的是一条由 10 条明说的边组成、写在第 2 卷到第 6 卷之间的链;增量环检测在这条边写下的那一刻就能报警。另一个是作者自己的遗忘:在示意参数下,凭记忆写到第 4 卷平均已有 2.06 处矛盾,查阅设定集可以把它降到 0.51 处,但写到第 10 卷,仍有 0.813 的模拟至少出现一处。设定集、版本检查与环检测各管一段,三者合起来,才接近「零吃书」。
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。
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: 0.51, 0.813, 10, 2.06
| vols | chapters | memory: mean | P(>=1) | bible: mean | P(>=1) |
| 1 | 40 | 0.08 | 0.073 | 0.04 | 0.043 |
| 2 | 80 | 0.42 | 0.320 | 0.15 | 0.140 |
| 4 | 160 | 2.06 | 0.873 | 0.51 | 0.387 |
| 6 | 240 | 4.56 | 0.997 | 0.89 | 0.600 |
| 8 | 320 | 8.40 | 1.000 | 1.43 | 0.737 |
| 10 | 400 | 11.87 | 1.000 | 1.86 | 0.813 |
| mean gap between restatements: 100 chapters | |||||
| memory error at that gap: 0.0197 |
查看完整验算程序
# Chapter 2. p22: the author forgets too; contradictions grow with gaps.
# A serial keeps F facts. Each chapter restates R of them at random.
# Restating a fact last written g chapters earlier, an author working
# from memory errs with probability EMAX * (1 - exp(-g / LAM)), an
# Ebbinghaus-type rise in error with the gap; an author who looks the fact
# up in a setting bible errs with a flat EBIB. Fixed-seed Monte Carlo.
# All parameters are illustrative, not measured on any author.
import math
import random
F, R, CH_PER_VOL = 300, 3, 40
EMAX, LAM, EBIB = 0.05, 200.0, 0.002
RUNS = 300
rng = random.Random(20261112)
def run(vols, lookup):
last = {}
errs = []
for ch in range(vols * CH_PER_VOL):
for f in rng.sample(range(F), R):
if f in last:
g = ch - last[f]
p = EBIB if lookup else EMAX * (1 - math.exp(-g / LAM))
if rng.random() < p:
errs.append(ch // CH_PER_VOL + 1)
last[f] = ch
return errs
print("vols chapters memory: mean P(>=1) bible: mean P(>=1)")
for vols in (1, 2, 4, 6, 8, 10):
row = []
for lookup in (False, True):
counts = [len(run(vols, lookup)) for _ in range(RUNS)]
mean = math.fsum(counts) / RUNS
p1 = len([c for c in counts if c > 0]) / RUNS
row += [mean, p1]
print("{:>4} {:>8} {:>12.2f} {:>6.3f} {:>11.2f} {:>6.3f}".format(
vols, vols * CH_PER_VOL, *row))
print("mean gap between restatements: {:.0f} chapters".format(F / R))
print("memory error at that gap: {:.4f}".format(
EMAX * (1 - math.exp(-(F / R) / LAM))))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 2 章「跨卷吃书:增量环检测、传递闭包与作者自己的遗忘」(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}
}