关于「实践:读一张样盘,以及一份盲配检验方案」,《紫微三十八原型与商业变现》给出的定量结论是什么?
结论:第八章用一张不含任何出生资料的样盘示范查法,并给出一份预先登记的盲配检验方案:三选一、真实命中率 0.40 时需要 325 人。书中以标注框区分已发表的证据、可运行的 Python 计算、原典与作者的假说。
数据口径:全部计算由书中程序生成;本书不是投资、职业、心理或婚姻建议。 书中以标注框区分已发表的证据、可运行的 Python 计算、原典与作者的假说。
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: 0.40, 325
| k | chance | true hit | N exact | N normal | N exact (2 ends) |
| 3 | 0.3333 | 0.40 | 325 | 318 | 415 |
| 3 | 0.3333 | 0.45 | 111 | 105 | 139 |
| 3 | 0.3333 | 0.50 | 58 | 52 | 68 |
| 6 | 0.1667 | 0.25 | 148 | 138 | 182 |
| 6 | 0.1667 | 0.30 | 62 | 57 | 80 |
| N = 100, k = 3: reject if hits >= 42; false-positive rate 0.0434 |
查看完整验算程序
# Chapter 8. p09: How many people a blind matching test needs. Each
# participant sees k anonymised readings - one written from their own
# chart's archetype entry, k - 1 decoys drawn from other families - and
# picks the one that fits. If the readings carry no chart-specific
# information (pure Barnum text), the hit rate is 1/k. We test one-sided
# with the exact binomial at alpha 0.05 (and 0.025 when the self card
# and the money card are two primary endpoints) and find the smallest N
# whose exact power reaches 0.80 for a true hit rate p1. Exact sums of
# binomial terms in logs; the scan starts below the normal-theory N.
from math import exp, lgamma, log, ceil, sqrt
from statistics import NormalDist
def tail(n, c, p):
"""P(X >= c) for X ~ Binomial(n, p)."""
lp, lq = log(p), log(1 - p)
return sum(exp(lgamma(n + 1) - lgamma(x + 1) - lgamma(n - x + 1)
+ x * lp + (n - x) * lq) for x in range(c, n + 1))
def power(n, p0, p1, alpha):
c = int(n * p0)
while tail(n, c, p0) > alpha:
c += 1
return tail(n, c, p1), c
def n_needed(p0, p1, alpha):
z = NormalDist().inv_cdf
na = ((z(1 - alpha) * sqrt(p0 * (1 - p0)) + z(0.80) *
sqrt(p1 * (1 - p1))) / (p1 - p0)) ** 2
n = max(10, int(0.8 * na))
while power(n, p0, p1, alpha)[0] < 0.80:
n += 1
return n, ceil(na)
if __name__ == "__main__":
print(" k chance true hit N exact N normal N exact (2 ends)")
for k, p1 in ((3, 0.40), (3, 0.45), (3, 0.50), (6, 0.25), (6, 0.30)):
p0 = 1 / k
n1, na = n_needed(p0, p1, 0.05)
n2, _ = n_needed(p0, p1, 0.025)
print(" {} {:.4f} {:.2f} {:>7} {:>8} {:>8}".format(
k, p0, p1, n1, na, n2))
pw, c = power(100, 1 / 3, 1 / 3, 0.05)
print(" N = 100, k = 3: reject if hits >= {}; false-positive rate"
" {:.4f}".format(c, pw))唯一数据源与引用
范雅福(Fan, Y.). (2026). 《紫微星宫全息人格与商业变现图谱》第 8 章「实践:读一张样盘,以及一份盲配检验方案」(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-116 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026116,
author = {Fan, Yafu},
title = {The Zi Wei Holographic Personality and Monetization Atlas},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}