关于「收费站出生时刻校准问卷:只读建议与预注册检验」,《出生时刻误差与交界翻转》给出的定量结论是什么?
结论:第八章给出收费站出生时刻校准问卷、交界翻转概率计算器与预注册检验:两个候选、真实命中率六成时需要 169 人,约筛查 676 份记录。书中以标注框区分已发表的证据、可运行的 Python 计算(22 段程序)、原典与作者的假说。
数据口径:本书只分析排盘规则的结构、误差传播与概率,不做任何个人占断,不宣称任何体系有预测效力;书中参数为示意,全部计算由书中程序生成。 第一章说明,示意模型中凭记忆报告的时刻有 0.2782 恰好落在时辰分界上,不取整时只有 0.0083。
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: 169, 676
| K p0 p1 n critical size power | |||||
| 2 0.5000 | 0.60 | 169 | 96 | 0.0451 | 0.8231 |
| 2 0.5000 | 0.65 | 76 | 46 | 0.0423 | 0.8262 |
| 2 0.5000 | 0.70 | 42 | 27 | 0.0442 | 0.8360 |
| 3 0.3333 | 0.45 | 121 | 50 | 0.0402 | 0.8170 |
| 3 0.3333 | 0.50 | 60 | 27 | 0.0397 | 0.8169 |
| documented times within 15 min of a boundary: 30 of every 120 | |||||
| minutes, so for K = 2 and p1 = 0.60 screen about 676 records | |||||
查看完整验算程序
# Chapter 8. p82: how many people a blinded validation test needs.
# Design: people whose birth time is documented by a hospital record to
# the minute and lies within 15 minutes of a double-hour boundary. A
# rater who never sees the record answers-to-chart scores the item
# battery and picks one of K candidate double hours. Under the null
# the pick is right with probability 1/K. Exact one-sided binomial
# test at alpha = 0.05: smallest n whose critical count gives power
# at least 0.80 against the stated true hit rate p1. No such hit rate
# has been demonstrated; the p1 values are the effect sizes worth
# detecting, not estimates.
import math
def tail(k, n, p):
return math.fsum(math.comb(n, j) * p ** j * (1 - p) ** (n - j)
for j in range(k, n + 1))
def critical(n, p0, alpha):
lo, hi = 0, n + 1
while lo < hi:
mid = (lo + hi) // 2
if tail(mid, n, p0) <= alpha:
hi = mid
else:
lo = mid + 1
return lo
def n_for(p0, p1, alpha=0.05, power=0.80):
n = 5
while True:
k = critical(n, p0, alpha)
if k <= n and tail(k, n, p1) >= power:
# 二项检验的锯齿:再确认之后几个 n 也满足
if all(tail(critical(m, p0, alpha), m, p1) >= power
for m in range(n + 1, n + 6)):
return n, k, tail(k, n, p0), tail(k, n, p1)
n += 1
print(" K p0 p1 n critical size power")
for K, p1 in ((2, 0.60), (2, 0.65), (2, 0.70), (3, 0.45), (3, 0.50)):
p0 = 1.0 / K
n, k, a, b = n_for(p0, p1)
print("%2d %6.4f %6.2f %5d %8d %8.4f %8.4f" % (K, p0, p1, n, k, a, b))
n2 = n_for(0.5, 0.6)[0]
print("documented times within 15 min of a boundary: 30 of every 120")
print("minutes, so for K = 2 and p1 = 0.60 screen about %d records"
% math.ceil(n2 / (30.0 / 120.0)))唯一数据源与引用
范雅福(Fan, Y.). (2026). 《出生时刻记忆误差传播与鲁棒定盘、交界时辰盘面翻转概率与最少前事反推题数母典》第 8 章「收费站出生时刻校准问卷:只读建议与预注册检验」(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-284 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026284,
author = {Fan, Yafu},
title = {Error Propagation from Remembered Birth Times and Robust Chart Fixing: Flip Probabilities at Double-Hour Boundaries and the Fewest Past-Event Questions},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}