关于「仪轨即协议:重掷规则、停止规则与签筒均匀性」,《筊杯、三圣杯与签诗》给出的定量结论是什么?

结论:第六章把仪轨读成协议,指出允许重来时最终确认比例不说明任何问题,并说明一个磨损的签筒在 2400 次确认时也只有 0.650 的机会被审计发现。

数据口径:书中参数为示意;全部计算由书中程序生成。 第一章以准静态规则求单片筊杯平面朝上的机会,半圆截面的示意模型给出 0.6278,一对的圣杯随之降到 0.4673,并证明「三种结果各占三分之一」对相同的一对不可能成立。

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: 0.650, 2400

fair tube, 3000 confirmed lots: draws per lot 2.021
chi2 54.36(reject if > 77.93)
Nreject rate fairreject rate worn
3000.0800.085
6000.0450.165
12000.0550.330
24000.0400.650
worn tube: lots 1-10 share 0.2063 instead of 0.1667
查看完整验算程序
# Chapter 6. p62: drawing a lot, confirming it, and auditing the tube.
# Model: a tube holds 60 numbered lots (a common size for verse sets).
# The asker draws one and casts once; a holy cast (q = 0.5) confirms
# the lot, otherwise the lot goes back and another is drawn. If
# confirmation ignores the lot number, the confirmed lot stays uniform
# and the mean number of draws is 1 / q. The audit: in an illustrative
# worn tube, lots 1 to 10 are slightly longer and drawn with weight
# 1.3. A chi-square test with 59 degrees of freedom (critical value
# 77.93 at the 5% level) is run on seeded samples of N confirmed lots,
# 200 repetitions each, to estimate the power of the audit.
import math
import random

LOTS = 60
CRIT = 77.93
REPS = 200


def draw(rng, cum):
    x = rng.random() * cum[-1]
    lo, hi = 0, len(cum) - 1
    while lo < hi:
        mid = (lo + hi) // 2
        if cum[mid] < x:
            lo = mid + 1
        else:
            hi = mid
    return lo


def confirmed_sample(rng, cum, n, q):
    counts = [0] * LOTS
    draws = 0
    for _ in range(n):
        while True:
            draws += 1
            lot = draw(rng, cum)
            if rng.random() < q:
                counts[lot] += 1
                break
    return counts, draws


def chi2(counts, n):
    e = n / LOTS
    return math.fsum([(c - e) ** 2 / e for c in counts])


def cumulative(weights):
    return [math.fsum(weights[:k + 1]) for k in range(len(weights))]


if __name__ == "__main__":
    rng = random.Random(20261112)
    fair = cumulative([1.0] * LOTS)
    worn = cumulative([1.3] * 10 + [1.0] * (LOTS - 10))
    counts, draws = confirmed_sample(rng, fair, 3000, 0.5)
    print("fair tube, 3000 confirmed lots: draws per lot {:.3f}".format(
        draws / 3000))
    print("  chi2 {:.2f}  (reject if > {})".format(
        chi2(counts, 3000), CRIT))
    print("  N       reject rate fair   reject rate worn")
    for n in (300, 600, 1200, 2400):
        rf = 0
        rw = 0
        for _ in range(REPS):
            cf, _ = confirmed_sample(rng, fair, n, 0.5)
            cw, _ = confirmed_sample(rng, worn, n, 0.5)
            rf += chi2(cf, n) > CRIT
            rw += chi2(cw, n) > CRIT
        print("  {:>4d}    {:.3f}              {:.3f}".format(
            n, rf / REPS, rw / REPS))
    print("worn tube: lots 1-10 share {:.4f} instead of {:.4f}".format(
        13.0 / 63.0, 10.0 / 60.0))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《掷筊杯刚体落地概率、连续三圣杯似然与签诗文本巴纳姆效应母典》第 6 章「仪轨即协议:重掷规则、停止规则与签筒均匀性」(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-225 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026225,
  author    = {Fan, Yafu},
  title     = {Rigid-Body Landing Probabilities of Moon Blocks, the Likelihood of Three Consecutive Holy Casts and Barnum Effects in Temple Oracle Verses},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论