关于「最优预付折扣:省下的代收费、更少的反悔与让出去的钱」,《货到付款押金与预付筛选》给出的定量结论是什么?
结论:第四章求出最优预付折扣 0.5556,约为售价的 0.0278。
数据口径:本书把一单的账拆开:示意参数下,签收一单的毛利为 4.5342,拒收一单要付出双程运费、货损与资金占用共 4.3522,盈亏平衡拒收率为 0.4202;在买家自愿付款的保守假设下,追讨要打平需要联系率与付款率之积达到 0.2068,现实中很难做到。 东南亚拒收率没有找到经同行评议的估计,书中参数全部为示意,全部计算由书中程序生成。
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: 0.0278, 0.5556
| COD-only profit per order 2.6234 u | |||||
| d | take H | take L | gain/order | piP-piC H | piP-piC L |
| 0.20 | 0.0461 | 0.0000 | 0.01336 | 0.3865 | 1.6930 |
| 0.40 | 0.1103 | 0.0070 | 0.01888 | 0.1955 | 1.5380 |
| 0.50 | 0.1429 | 0.0391 | 0.02500 | 0.1000 | 1.4605 |
| 0.60 | 0.1759 | 0.0715 | 0.02532 | 0.0045 | 1.3830 |
| 0.80 | 0.2428 | 0.1374 | 0.00821 | -0.1865 | 1.2280 |
| 1.00 | 0.3112 | 0.2046 | -0.03321 | -0.3775 | 1.0730 |
| 1.50 | 0.4885 | 0.3791 | -0.24829 | -0.8550 | 0.6855 |
| optimal discount d* = 0.5556 u (0.0278 of P) | |||||
| uptake H 0.1612, L 0.0571; gain 0.02590 u per order | |||||
| casual buyers start to switch at d = 0.3780 u |
查看完整验算程序
# Chapter 4. p41: the profit-maximising prepayment discount.
# A seller keeps COD at the market price P = 20 u and offers prepay at
# P - d. Prepay saves the COD fee (0.03) and float, costs an online
# payment fee of 0.015, and a prepaid buyer who changes his mind pays
# the return leg himself (h = 1.8 u) and does so less often: rate
# rho q with rho = 0.6 (commitment, illustrative). A buyer switches
# iff d >= phi (P - d) + rho q h, with his trust cost phi uniform on
# [0, 0.16] (mean 0.08, as in p21). Gain per order:
# G(d) = sum_i w_i s_i(d) (piP_i(d) - piC_i), refusal rates qH = 0.05
# (weight 0.75) and qL = 0.35 (weight 0.25). Grid of 1/1000 u then a
# golden-section polish. Illustrative parameters.
import math
P, C, SO, SR, A = 20.0, 13.0, 1.8, 1.8, 0.8
FEE, E, RHO, H, PHIMAX = 0.03, 0.015, 0.6, 1.8, 0.16
CAP = (C + SO) * 0.12 * 21 / 365
CCOD = C + SO + FEE * P + P * 0.12 * 10 / 365
K = math.fsum([SO, SR, 0.05 * C, CAP])
TYPES = ((0.05, 0.75), (0.35, 0.25))
def pi_cod(q):
return (1 - q) * (P - CCOD) - q * K - A
def pi_pre(q, d):
cp = C + SO + E * (P - d)
kp = math.fsum([SO, 0.05 * C, CAP, E * (P - d)])
return (1 - RHO * q) * (P - d - cp) - RHO * q * kp - A
def share(q, d, phimax=PHIMAX):
s = (d - RHO * q * H) / (phimax * (P - d))
return min(1.0, max(0.0, s))
def gain(d, phimax=PHIMAX):
return math.fsum(w * share(q, d, phimax) * (pi_pre(q, d) - pi_cod(q))
for q, w in TYPES)
def best(f, lo=0.0, hi=5.0):
d0 = max(range(int(lo * 1000), int(hi * 1000) + 1),
key=lambda i: f(i / 1000)) / 1000
a, b = max(lo, d0 - 0.001), min(hi, d0 + 0.001)
g = (math.sqrt(5) - 1) / 2
for _ in range(60):
x1, x2 = b - g * (b - a), a + g * (b - a)
if f(x1) >= f(x2):
b = x2
else:
a = x1
return (a + b) / 2
if __name__ == "__main__":
base = math.fsum(w * pi_cod(q) for q, w in TYPES)
print("COD-only profit per order {:.4f} u".format(base))
print(" d take H take L gain/order piP-piC H piP-piC L")
for d in (0.2, 0.4, 0.5, 0.6, 0.8, 1.0, 1.5):
print("{:5.2f} {:7.4f} {:7.4f} {:11.5f} {:10.4f} {:10.4f}".format(
d, share(0.05, d), share(0.35, d), gain(d),
pi_pre(0.05, d) - pi_cod(0.05),
pi_pre(0.35, d) - pi_cod(0.35)))
ds = best(gain)
print("optimal discount d* = {:.4f} u ({:.4f} of P)".format(
ds, ds / P))
print("uptake H {:.4f}, L {:.4f}; gain {:.5f} u per order".format(
share(0.05, ds), share(0.35, ds), gain(ds)))
print("casual buyers start to switch at d = {:.4f} u".format(
RHO * 0.35 * H))唯一数据源与引用
范雅福(Fan, Y.). (2026). 《货到付款拒收博弈与东南亚电商信任赤字、押金阈值与预付折扣激励设计母典》第 4 章「最优预付折扣:省下的代收费、更少的反悔与让出去的钱」(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-314 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026314,
author = {Fan, Yafu},
title = {Cash-on-Delivery Refusal Games and the Trust Deficit in Southeast Asian E-Commerce: Deposit Thresholds and Prepayment-Discount Incentive Design},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}