关于「实证闭环:孤品暗标流程、席位二价公示与预注册检验」,《二价暗标与最优保留价》给出的定量结论是什么?

结论:第八章给出孤品暗标六步流程、席位二价公示、暗标保留价计算器与预注册检验方案:三位买家时流拍检验每组 33 件,所得检验每组 164 件。书中以标注框区分已发表的证据、可运行的 Python 计算(18 段程序)、原典与作者的假说。

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)。

English version

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: 164, 33

n arm mean sd lots per arm for this gap
2 floor111240 14539
2 r*119534 16334
gap 8294, pooled sd 1546355 lots per arm
3 floor121139 16304
3 r*126143 16029
gap 5004, pooled sd 16167164 lots per arm
5 floor133388 14142
5 r*135021 13893
gap 1632, pooled sd 140181158 lots per arm
no-sale rate n3: 0.0156 vs 0.2441 -> 33 lots per arm
查看完整验算程序
# Chapter 8. p82: how many lots a pre-registered reserve test needs.
# Design: each eligible lot is randomly assigned to "floor = s" or
# "reserve = r*" before invitations go out; outcome = seller payoff of
# the lot (price if sold, s if kept). Per-lot spread is estimated by a
# seeded Monte Carlo of the book's uniform model (values U[80,000,
# 160,000] yuan, s = 100,000, r* = 130,000), 40,000 lots per arm.
# Two-sided a = 0.05, power 0.80, unpaired:
#   lots per arm = 2 (z_a + z_b)^2 sd^2 / delta^2.
# All parameters are illustrative; a house must use its own estimates.
import math
import random
from statistics import NormalDist

L, H, S, R, RUNS = 80000.0, 160000.0, 100000.0, 130000.0, 40000
Z = NormalDist()
za, zb = Z.inv_cdf(0.975), Z.inv_cdf(0.80)
rng = random.Random(20261113)


def lot(vals, r):
    v = sorted(vals, reverse=True)
    if v[0] < r:
        return S
    return max(v[1], r) if len(v) > 1 else r


print("n  arm      mean     sd      lots per arm for this gap")
plan = {}
for n in (2, 3, 5):
    res = {}
    for name, r in (("floor", S), ("r*", R)):
        x = [lot([rng.uniform(L, H) for _ in range(n)], r)
             for _ in range(RUNS)]
        m = math.fsum(x) / RUNS
        sd = math.sqrt(math.fsum((y - m) ** 2 for y in x) / (RUNS - 1))
        res[name] = (m, sd)
    delta = res["r*"][0] - res["floor"][0]
    sd = math.sqrt((res["r*"][1] ** 2 + res["floor"][1] ** 2) / 2.0)
    need = math.ceil(2.0 * (za + zb) ** 2 * sd ** 2 / delta ** 2)
    plan[n] = need
    for name in ("floor", "r*"):
        print("%d  %-5s  %8.0f  %6.0f" % (n, name, res[name][0],
                                         res[name][1]))
    print("   gap %.0f, pooled sd %.0f -> %d lots per arm"
          % (delta, sd, need))
p0, p1 = 0.25 ** 3, 0.625 ** 3
pbar = (p0 + p1) / 2.0
k = math.ceil((za * math.sqrt(2 * pbar * (1 - pbar))
               + zb * math.sqrt(p0 * (1 - p0) + p1 * (1 - p1))) ** 2
              / (p1 - p0) ** 2)
print("no-sale rate n = 3: %.4f vs %.4f -> %d lots per arm" % (p0, p1, k))

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 8 章「实证闭环:孤品暗标流程、席位二价公示与预注册检验」(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}
}

本书其他章节结论