关于「公盘暗标:N 人共同价值拍卖的赢家诅咒蒙特卡洛」,《翡翠鉴别与赌石赢家诅咒》给出的定量结论是什么?
结论:第五章模拟 N 人共同价值暗标:天真规则下十二人竞价的中标者平均每场亏 19.46,按「假如我已中标」修正后利润回到预留水平。
数据口径:书中参数为示意;全部计算由书中程序生成;本书不是个人化的法律或理财建议,也不是任何一件石头的鉴定意见。 第六章把三项修正折成出价上限:挑窗、六人竞价、证书可靠时为窗口估值的 0.309,只做门店检验时为 0.274;在天真买家占多数的公盘里应当离场;凯利比例在示意优势下为 0.090。
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
| naive bidders: bid = own estimate - 5 | |||
| N | mean profit | P(winner loses) | theory 5 - 15 e_N |
| 2 | -3.50 | 0.603 | -3.46 |
| 4 | -10.46 | 0.841 | -10.44 |
| 6 | -14.00 | 0.939 | -14.01 |
| 8 | -16.47 | 0.976 | -16.35 |
| 12 | -19.46 | 0.996 | -19.44 |
| 16 | -21.35 | 1.000 | -21.49 |
| 24 | -24.17 | 1.000 | -24.22 |
| N = 12, auctions split by winning bid: | |||
| lowest third | mean bid | 95.9 | mean profit -15.94 |
| middle third | mean bid | 119.1 | mean profit -19.15 |
| highest third | mean bid | 142.9 | mean profit -23.28 |
查看完整验算程序
# Chapter 5. p51: N-bidder common-value auction with naive bidders.
# 共同价值模型(参数示意):一块原石对所有买手价值相同 V ~ N(100, 20);
# 每位买手的估计 s_i = V + e_i,e_i ~ N(0, 15),无偏且相互独立。
# 天真买手按估计减固定利润 5 出价;暗标最高价中标并按己价付款。
# 固定种子蒙特卡洛,每个 N 模拟 20000 场;理论值 = 5 - 15 e_N(见 p11)。
# 再在 N = 12 时按中标价三等分,看「中标价越高,赢家亏得越多」是否成立(条件期望,而非因果)。
import math
import random
from statistics import NormalDist
ND = NormalDist()
MU, TAU, SIG, MARGIN = 100.0, 20.0, 15.0, 5.0
ROUNDS = 20000
def e_max(n, lo=-9.0, hi=9.0, steps=18000):
h = (hi - lo) / steps
vals = []
for i in range(steps + 1):
x = lo + i * h
w = 1 if i in (0, steps) else (4 if i % 2 else 2)
vals.append(w * x * n * ND.pdf(x) * ND.cdf(x) ** (n - 1))
return math.fsum(vals) * h / 3
def auction(rng, n):
v = rng.gauss(MU, TAU)
top = max(v + rng.gauss(0.0, SIG) for _ in range(n))
bid = top - MARGIN
return v, bid
def main():
print("naive bidders: bid = own estimate - %.0f" % MARGIN)
print(" N mean profit P(winner loses) theory 5 - 15 e_N")
for n in (2, 4, 6, 8, 12, 16, 24):
rng = random.Random(20261112 + n)
prof = []
for _ in range(ROUNDS):
v, b = auction(rng, n)
prof.append(v - b)
mp = math.fsum(prof) / ROUNDS
pl = len([p for p in prof if p < 0]) / ROUNDS
print("%4d %+7.2f %.3f %+6.2f"
% (n, mp, pl, MARGIN - SIG * e_max(n)))
rng = random.Random(20261112)
rows = sorted((auction(rng, 12) for _ in range(ROUNDS)),
key=lambda r: r[1])
print("N = 12, auctions split by winning bid:")
third = ROUNDS // 3
names = ("lowest third", "middle third", "highest third")
for j in range(3):
part = rows[j * third:(j + 1) * third]
mb = math.fsum(b for _, b in part) / len(part)
mp = math.fsum(v - b for v, b in part) / len(part)
print(" %-14s mean bid %6.1f mean profit %+6.2f"
% (names[j], mb, mp))
if __name__ == "__main__":
main()唯一数据源与引用
范雅福(Fan, Y.). (2026). 《翡翠静水比重贝叶斯鉴别、赌石开窗信号博弈与原石拍卖赢家诅咒母典》第 5 章「公盘暗标:N 人共同价值拍卖的赢家诅咒蒙特卡洛」(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-206 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026206,
author = {Fan, Yafu},
title = {Hydrostatic Specific Gravity and Bayesian Jadeite Identification, Window-Cut Signalling Games and the Winner's Curse in Rough-Jade Tenders},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}