关于「四种格式的同一个期望:收入等价的精确积分与蒙特卡洛」,《二价暗标与最优保留价》给出的定量结论是什么?
结论:第一章用分数精确积分核对收入等价,并用两万次共同随机数的模拟说明平均相同而风险不同:二价暗标的成交价标准差 16185 元,一价暗标 9930 元。
数据口径:书中的估值分布、保有价值、买家人数与价格均为示意;全部计算由书中程序生成;本书是非个人化经营建议,也不是法律意见,正式拍卖以上架前人工复核的当地法规为准;紧迫感只能来自真实库存与真实截止,不虚构竞价者,不托价。 第七章考辨《宋史》的买扑与苏辙「立定中价,不许添刬」之议(示意模型中按自身估计出价的违约概率达 0.9691)以及希罗多德记载的巴比伦婚配拍卖(收入是嫁资的 3.373 倍)。
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: 16185, 9930
| exact expected revenue (all four, continuous): 128000 | |||||
| format | mean | 95% CI half | sd | P(<110k) | p05 |
| second-price | 127926 | 224 | 16185 | 0.1553 | 99343 |
| English | 128400 | 224 | 16172 | 0.1431 | 100000 |
| first-price | 127873 | 138 | 9930 | 0.0652 | 107944 |
| Dutch | 127370 | 138 | 9928 | 0.0652 | 107000 |
| exact sd: second-price 16000, first-price 9798 |
查看完整验算程序
# Chapter 1. p12: same mean, different spread - a seeded Monte Carlo.
# One-of-a-kind bag, n = 4 qualified bidders, private values iid
# uniform on [80,000, 160,000] yuan (illustrative). 20,000 auctions per
# format, the same value draws for every format (common random numbers).
# Second-price: truthful bids, price = second value. English: clock
# rises in steps of 1,000 yuan, price = first step at which only one
# bidder is left. First-price: equilibrium bid L + (n-1)/n (v - L).
# Dutch: clock falls in steps of 1,000, the first bidder whose bid is
# reached stops it, price = that clock level.
import math
import random
L, H, N, RUNS, STEP = 80000, 160000, 4, 20000, 1000
rng = random.Random(20261113)
draws = [sorted((rng.uniform(L, H) for _ in range(N)), reverse=True)
for _ in range(RUNS)]
def english(v):
up = L + STEP * math.ceil((v[1] - L) / STEP)
return up if v[0] >= up else up - STEP
def first(v):
return L + (N - 1) / N * (v[0] - L)
def dutch(v):
return H - STEP * math.ceil((H - first(v)) / STEP)
FORMATS = [("second-price", lambda v: v[1]), ("English", english),
("first-price", first), ("Dutch", dutch)]
exact = L + (H - L) * (N - 1) / (N + 1)
print("exact expected revenue (all four, continuous): %.0f" % exact)
print("format mean 95% CI half sd P(<110k) p05")
for name, rule in FORMATS:
rev = sorted(rule(v) for v in draws)
mean = math.fsum(rev) / RUNS
var = math.fsum((r - mean) ** 2 for r in rev) / (RUNS - 1)
sd = math.sqrt(var)
half = 1.96 * sd / math.sqrt(RUNS)
low = len([r for r in rev if r < 110000]) / RUNS
print("%-12s %7.0f %6.0f %6.0f %.4f %6.0f"
% (name, mean, half, sd, low, rev[RUNS // 20]))
sd2 = (H - L) * math.sqrt(2 * (N - 1) / ((N + 1) ** 2 * (N + 2)))
sd1 = (H - L) * (N - 1) / N * math.sqrt(N / ((N + 1) ** 2 * (N + 2)))
print("exact sd: second-price %.0f, first-price %.0f" % (sd2, sd1))唯一数据源与引用
范雅福(Fan, Y.). (2026). 《维克里二价暗标、迈尔森最优保留价与孤品奢包限量席位激励相容拍卖母典》第 1 章「四种格式的同一个期望:收入等价的精确积分与蒙特卡洛」(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-237 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026237,
author = {Fan, Yafu},
title = {Vickrey Second-Price Sealed Bids and Myerson Optimal Reserve Prices: Incentive-Compatible Auctions for One-of-a-Kind Luxury Pieces and Limited Seats},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}