关于「问答即泄露:冷读者的熵预算与「对上」为何几乎必然」,《铁板神数的比特账》给出的定量结论是什么?

结论:第四章是全书核心:在一个时辰的 120 个分中搜索,父母属相完全对上的概率为 0.6975,表面上 6.451 比特的罕事,真实证据只有 0.520 比特。

数据口径:书中参数为示意;全部计算由书中程序生成;本书分析结构与概率,不对任何个人作任何占断,也不构成法律或理财建议。 第五章说明一句条文的期望信息量至多 0.5307 比特,按示意规模每分钟有 19.01 人出生。

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: 0.520, 0.6975, 120, 6.451

q exact, uniform pairs: 0.00694; age-gap pairs: 0.01143
q father or mother matches: 0.1552
KP(exact)E[exact]P(partial)bits shownbits real
10.01140.0110.15526.4516.451
80.08730.0910.74046.4513.518
240.23610.2740.98246.4512.083
1200.69751.3721.00006.4510.520
3600.92504.1161.00006.4510.113
Monte Carlo, K = 120, 4000 clients: P(exact) = 0.7007
查看完整验算程序
# Chapter 4. p42: Why a match becomes certain. The reader holds K
# candidate cells (8 ke, 120 fen, or 360 fen when the neighbouring
# double-hours are admitted) and reads each cell's check verse, which
# names the father's and mother's animals. The verse pairs are drawn
# from the same population as the client's (age-gap model of p22).
# q = collision probability of the pair. With K tries the chance of at
# least one exact match is sum_c p_c (1 - (1 - p_c)^K), close to
# 1 - (1 - q)^K. A match found after K hidden tries is worth
# -log2 P(match) bits of evidence, not -log2 q.
import math
import random

GAP = [10, 18, 20, 17, 12, 8, 5, 3, 2, 0, 2, 3]
g = [x / 100 for x in GAP]
p_pair = [[g[(mo - fa) % 12] / 12 for mo in range(12)]
          for fa in range(12)]
q_uni = 1 / 144
flat = [p for row in p_pair for p in row]
q = math.fsum(p * p for p in flat)
# partial: father matches or mother matches
p_f = 1 / 12
q_part = p_f + p_f - q
print("  q exact, uniform pairs: {:.5f}; age-gap pairs: {:.5f}".format(
    q_uni, q))
print("  q father or mother matches: {:.4f}".format(q_part))
print("    K   P(exact)  E[exact]  P(partial)  bits shown  bits real")
for K in (1, 8, 24, 120, 360):
    pe = math.fsum(p * (1 - (1 - p) ** K) for p in flat)
    pp = math.fsum(p * (1 - (1 - (2 / 12 - p)) ** K) for p in flat)
    print("  {:>3}   {:8.4f}  {:8.3f}  {:10.4f}  {:10.3f}  {:9.3f}".format(
        K, pe, K * q, pp, -math.log2(q), -math.log2(pe)))

rng = random.Random(20261112)
cells = [(f, m) for f in range(12) for m in range(12)]
wts = [p_pair[f][m] for f, m in cells]
hits = 0
TR = 4000
for _ in range(TR):
    client = rng.choices(cells, wts)[0]
    verses = rng.choices(cells, wts, k=120)
    hits += client in verses
print("  Monte Carlo, K = 120, {} clients: P(exact) = {:.4f}".format(
    TR, hits / TR))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《铁板神数考刻分决策树、条文库仿射编码与信息熵套取下界》第 4 章「问答即泄露:冷读者的熵预算与「对上」为何几乎必然」(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-157 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026157,
  author    = {Fan, Yafu},
  title     = {Tieban Shenshu Ke-Fen Decision Trees, Affine Coding of the Verse Bank and Entropy Lower Bounds on Elicitation},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论