关于「限量款溢价半衰期:对数线性回归、假货污染与配货隐含价格的崩塌条件」,《二手精品鉴定与柠檬折价》给出的定量结论是什么?
结论:第六章用对数线性回归估计溢价半衰期为 8.35 个月,说明假货污染会把它压短到 7.03 个月,并给出配货隐含价格的崩塌条件。
数据口径:书中参数为示意;全部计算由书中程序生成;本书不是对任何具体交易的定价意见,也不构成个人化的法律或理财建议。 书中以标注框区分已发表的证据、可运行的 Python 计算(20 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: 8.35
sales used: 120; true lambda 0.0770 per month fitted lambda 0.0830, standard error 0.0032, R^2 0.850 fitted g0 0.854 (truth 0.80) half-life 8.35 months; 95 % band 7.76 to 9.04 first 6 months only: half-life 6.62, band 4.70 to 11.19 first 12 months only: half-life 8.03, band 6.74 to 9.91
查看完整验算程序
# Chapter 6. p61: half-life of the resale premium of a limited bag.
# Synthetic data, fixed seed: premium over retail g(t) = P(t)/R - 1
# follows g0 * 2^(-t/H) times lognormal noise, with g0 = 0.80 and
# H = 9 months; 5 verified sales a month for months 0 to 23. Fit
# ln g = alpha - lambda t by ordinary least squares; half-life is
# ln 2 / lambda, and its 95 % band comes from the slope's standard
# error. The truth (g0, H, noise 0.25) is illustrative.
import math
import random
rng = random.Random(20261112)
G0, H, SD = 0.80, 9.0, 0.25
data = []
for t in range(24):
for _ in range(5):
g = G0 * 2 ** (-t / H) * math.exp(rng.gauss(0.0, SD))
data.append((t, math.log(g)))
def ols(pts):
n = len(pts)
mt = math.fsum(t for t, _ in pts) / n
my = math.fsum(y for _, y in pts) / n
stt = math.fsum((t - mt) ** 2 for t, _ in pts)
sty = math.fsum((t - mt) * (y - my) for t, y in pts)
b = sty / stt
a = my - b * mt
res = [y - a - b * t for t, y in pts]
s2 = math.fsum(r * r for r in res) / (n - 2)
syy = math.fsum((y - my) ** 2 for _, y in pts)
return a, b, math.sqrt(s2 / stt), 1 - math.fsum(r * r for r in res) / syy
a, b, se, r2 = ols(data)
lam = -b
print("sales used: {}; true lambda {:.4f} per month".format(
len(data), math.log(2) / H))
print("fitted lambda {:.4f}, standard error {:.4f}, R^2 {:.3f}".format(
lam, se, r2))
print("fitted g0 {:.3f} (truth 0.80)".format(math.exp(a)))
print("half-life {:.2f} months; 95 % band {:.2f} to {:.2f}".format(
math.log(2) / lam, math.log(2) / (lam + 1.96 * se),
math.log(2) / (lam - 1.96 * se)))
for months in (6, 12):
sub = [p for p in data if p[0] < months]
a2, b2, se2, _ = ols(sub)
print("first {:2d} months only: half-life {:.2f}, band {:.2f} to {:.2f}"
.format(months, math.log(2) / -b2,
math.log(2) / (-b2 + 1.96 * se2),
math.log(2) / (-b2 - 1.96 * se2)))唯一数据源与引用
范雅福(Fan, Y.). (2026). 《二手精品多特征贝叶斯鉴定、阿克洛夫柠檬折价均衡与限量包款二级市场折旧半衰期》第 6 章「限量款溢价半衰期:对数线性回归、假货污染与配货隐含价格的崩塌条件」(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-171 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026171,
author = {Fan, Yafu},
title = {Multi-Feature Bayesian Authentication of Pre-Owned Luxury, Akerlof Lemon-Discount Equilibria and the Resale Half-Life of Limited Bags},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}