关于「序贯概率比检验:要看多少周的记录才足以停下」,《远期承诺期权与对等停时》给出的定量结论是什么?
结论:第四章用序贯概率比检验给出停止边界,固定周数检验需 15.2 周。
数据口径:书中参数为示意;全部计算由书中程序生成;本书不是心理咨询、婚恋咨询或临床判断,也不做任何命理推导。 书中以标注框区分证据、可运行计算(21 段程序)、原典与作者假说。
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: 15.2
| upper boundary ln A | 2.890, lower boundary ln B = -2.251 |
| step for a missed week +0.773, for a met week -0.693 | |
| expected weeks under H0 | 7.9, under H1: 9.1 |
| fixed-length test with the same errors | 15.2 weeks |
| misses in a row that reach the upper boundary | 4 |
| met weeks in a row that reach the lower boundary | 4 |
查看完整验算程序
# Chapter 4. p41: Wald SPRT boundaries for a weekly reciprocity record
# Each week is scored 1 if the other party's observable investment met
# the agreed reciprocal floor, else 0. H0: p = 0.70 (reciprocal pace),
# H1: p = 0.35 (persistently one-sided). alpha = P(stop for H1 | H0),
# beta = P(continue as if H0 | H1). Wald (1945) approximations.
import math
from statistics import NormalDist
P0, P1, ALPHA, BETA = 0.70, 0.35, 0.05, 0.10
a = math.log((1.0 - BETA) / ALPHA)
b = math.log(BETA / (1.0 - ALPHA))
inc_miss = math.log((1.0 - P1) / (1.0 - P0))
inc_met = math.log(P1 / P0)
print("upper boundary ln A = {:.3f}, lower boundary ln B = {:.3f}".format(
a, b))
print("step for a missed week {:+.3f}, for a met week {:+.3f}".format(
inc_miss, inc_met))
def drift(p):
return (1.0 - p) * inc_miss + p * inc_met
asn0 = ((1.0 - ALPHA) * b + ALPHA * a) / drift(P0)
asn1 = ((1.0 - BETA) * a + BETA * b) / drift(P1)
print("expected weeks under H0: {:.1f}, under H1: {:.1f}".format(
asn0, asn1))
z = NormalDist()
za, zb = z.inv_cdf(1.0 - ALPHA), z.inv_cdf(1.0 - BETA)
s0, s1 = math.sqrt(P0 * (1 - P0)), math.sqrt(P1 * (1 - P1))
n_fix = ((za * s0 + zb * s1) / (P0 - P1)) ** 2
print("fixed-length test with the same errors: {:.1f} weeks".format(n_fix))
k = 0
while k * inc_miss < a:
k += 1
print("misses in a row that reach the upper boundary: {}".format(k))
m = 0
while m * inc_met > b:
m += 1
print("met weeks in a row that reach the lower boundary: {}".format(m))唯一数据源与引用
范雅福(Fan, Y.). (2026). 《亲职基因优选与排他恋爱解耦、远期空头期权钩子与对等底线停时母典》第 4 章「序贯概率比检验:要看多少周的记录才足以停下」(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-423 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026423,
author = {Fan, Yafu},
title = {Unbundling Paternal Selection from Romantic Commitment: Deferred-Promise Options, Avoidant Autonomy and Reciprocal Stopping Rules},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}