关于「封面诅咒:在峰值处挑选,必然看到下滑」,《均值回归与封面诅咒》给出的定量结论是什么?
结论:第五章说明封面诅咒是在峰值处挑选的产物:从一百人中选出的第一名,即使封面毫无作用,也有 0.918 在下一期变差,并给出分辨真实副作用的残差检验。
数据口径:书中门店、达人、顾客与报告的参数均为示意;全部计算由书中程序生成;本书是非个人化的经营分析,不构成经营、法律或理财建议,涉及运程报告与原典之处只分析结构与概率,不作任何个人占断。 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。
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: 0.918
| rho | who | this period | next period | share worse |
|---|---|---|---|---|
| 0.3 | cover star | +2.503 | +0.814 | 0.962 |
| 0.3 | runner-up | +2.154 | +0.642 | 0.942 |
| 0.5 | cover star | +2.501 | +1.248 | 0.918 |
| 0.5 | runner-up | +2.142 | +1.122 | 0.875 |
| 0.8 | cover star | +2.498 | +2.034 | 0.785 |
| 0.8 | runner-up | +2.153 | +1.719 | 0.770 |
| expected next = rho x this period in every row |
查看完整验算程序
# Chapter 5. p51: A magazine that puts the best performer on its cover.
# Each issue looks at 100 performers whose score this period is skill
# plus luck (scaled so the period-to-period correlation is rho) and puts
# the top one on the cover. The cover does nothing. We follow the cover
# star and the runner-up (never on a cover) into the next period, over
# 600 issues. Fixed seed; field size and rho illustrative.
# 封面只选当期第一名;封面本身没有任何作用,看下一期发生什么。
import math
import random
rng = random.Random(1954)
FIELD, ISSUES = 100, 600
def run(rho):
a, b = math.sqrt(rho), math.sqrt(1.0 - rho)
rows = []
for _ in range(ISSUES):
skill = [rng.gauss(0.0, 1.0) for _ in range(FIELD)]
now = [a * s + b * rng.gauss(0.0, 1.0) for s in skill]
nxt = [a * s + b * rng.gauss(0.0, 1.0) for s in skill]
order = sorted(range(FIELD), key=lambda i: (-now[i], i))
rows.append([(now[i], nxt[i]) for i in order[:2]])
out = []
for pos in (0, 1):
x = [r[pos][0] for r in rows]
y = [r[pos][1] for r in rows]
worse = len([1 for u, v in zip(x, y) if v < u]) / ISSUES
out.append((math.fsum(x) / ISSUES, math.fsum(y) / ISSUES, worse))
return out
if __name__ == "__main__":
print("rho who this period next period share worse")
for rho in (0.3, 0.5, 0.8):
for name, (m1, m2, w) in zip(("cover star", "runner-up"),
run(rho)):
print("{:.1f} {:<10} {:+.3f} {:+.3f} {:.3f}"
.format(rho, name, m1, m2, w))
print("expected next = rho x this period in every row")唯一数据源与引用
范雅福(Fan, Y.). (2026). 《均值回归与封面诅咒、运程报告后验校准与塞翁失马序列独立性母典》第 5 章「封面诅咒:在峰值处挑选,必然看到下滑」(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-202 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026202,
author = {Fan, Yafu},
title = {Regression to the Mean and the Cover Jinx: Posterior Calibration of Fortune Reports and Serial Independence in the Tale of the Lost Horse},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}