关于「交替出价:Rubinstein 贴现均衡、最后一轮与公平门槛」,《红蓝军合成客群沙盘》给出的定量结论是什么?
结论:程序 p31 给出 Rubinstein 均衡的切分:对称耐心 0.90 时先开口的一方拿 0.5263,出价间隔趋于零时份额只取决于耐心之比。程序 p32 显示,有限轮数会制造巨大的最后一轮优势,第 2 轮与第 3 轮相差 0.7200 个剩余。程序 p33 显示,加上公平门槛之后,均衡报价可能一个买家也说不动。那么,一群各有耐心、预算和公平感的挑剔买家,面对不同的让价策略,会把多少毛利带走?
数据口径:本书检验相反的立项命题:只用平均买家会漏掉尾部破绽,拉丁超立方合成客群在同样测试量下覆盖更多人格组合(McKay、Beckman 与 Conover 1979 年刊于《技术计量学》),多轮砍价有 Rubinstein 1982 年给出的贴现均衡。 书中参数为示意;全部计算由书中程序生成,不调用任何模型接口;本书是非个人化经营建议。
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: 0.5263, 0.90
| seller d1 buyer d2 seller x* buyer 1-x* | |
| 0.90 | 0.90 0.5263 0.4737 |
| 0.90 | 0.80 0.7143 0.2857 |
| 0.90 | 0.95 0.3448 0.6552 |
| 0.80 | 0.95 0.2083 0.7917 |
| 0.95 | 0.99 0.1681 0.8319 |
| 0.99 | 0.95 0.8403 0.1597 |
| equal patience d | 0.9: opener keeps 10/19 = 0.5263 |
| rates r1 | 0.05, r2 = 0.20 per round |
| step h | 1.00: seller share 0.8195 |
| step h | 0.50: seller share 0.8099 |
| step h | 0.10: seller share 0.8020 |
| step h | 0.01: seller share 0.8002 |
| limit r2 / (r1 + r2) | 0.8000 |
| list 1280, cost 700: buyer d2 | 0.95 wins 380.00 of surplus 580 |
查看完整验算程序
# Chapter 3. p31: the Rubinstein split and what patience is worth.
# Rubinstein (1982, Econometrica): two players alternate offers over a
# surplus of size 1; player 1 (here the seller, who opens) discounts a
# round of delay by d1, player 2 (the buyer) by d2. The unique subgame
# perfect split gives the opener x* = (1 - d2) / (1 - d1 d2), accepted
# at once. With d_i = exp(-r_i h) and h -> 0 the opener's share tends
# to r2 / (r1 + r2). Surplus = list price minus seller cost; values
# are illustrative. Exact fractions where the inputs are exact.
import math
from fractions import Fraction as F
def share(d1, d2):
return (1 - d2) / (1 - d1 * d2)
print("seller d1 buyer d2 seller x* buyer 1-x*")
for d1, d2 in ((F(9, 10), F(9, 10)), (F(9, 10), F(8, 10)),
(F(9, 10), F(95, 100)), (F(8, 10), F(95, 100)),
(F(95, 100), F(99, 100)), (F(99, 100), F(95, 100))):
x = share(d1, d2)
print(" %.2f %.2f %.4f %.4f"
% (d1, d2, x, 1 - x))
print("equal patience d = 0.9: opener keeps %s = %.4f"
% (share(F(9, 10), F(9, 10)), share(F(9, 10), F(9, 10))))
r1, r2 = 0.05, 0.20
print("rates r1 = %.2f, r2 = %.2f per round" % (r1, r2))
for h in (1.0, 0.5, 0.1, 0.01):
x = share(math.exp(-r1 * h), math.exp(-r2 * h))
print(" step h = %.2f: seller share %.4f" % (h, x))
print(" limit r2 / (r1 + r2) = %.4f" % (r2 / (r1 + r2)))
surplus = 1280 - 700
x = share(F(9, 10), F(95, 100))
print("list 1280, cost 700: buyer d2 = 0.95 wins %.2f of surplus %d"
% (float((1 - x) * surplus), surplus))唯一数据源与引用
范雅福(Fan, Y.). (2026). 《红蓝军合成客群对抗沙盘、参数化挑剔买家多轮砍价博弈与漏单破绽预检母典》第 3 章「交替出价:Rubinstein 贴现均衡、最后一轮与公平门槛」(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-259 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026259,
author = {Fan, Yafu},
title = {Red-Blue Synthetic-Customer Sandboxes: Multi-Round Bargaining against Parameterised Demanding Buyers and Pre-Launch Leak Detection},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}