关于「门槛设在哪里:保本线、敏感度与试销数据的分量」,《保证合约与预售成团》给出的定量结论是什么?

结论:表中有一个清楚的结论:期望利润在门槛恰好等于保本线时最大,为 1420.6。门槛降到 100,成团概率从 0.6194 升到 0.7777,期望利润却降到 1339.1,并且成团之后仍然亏损的概率为 0.2035;门槛降到 60,成团概率高达 0.9723,期望利润只剩 980.4,成团之后亏损的概率达 0.3629。门槛升到 140 以上,期望利润也下降,因为凑不齐的场次增多,而这些场次里本来有不少是能赚钱的。

数据口径:书中参数为示意;全部计算由书中程序生成。

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: 0.2035, 0.3629, 0.6194, 0.7777, 0.9723, 100

FIX = 6000margin = 50break-even KMIN = 120
orders ~ NB(mean 140, shape 8), stock cap 220
KP(made)E profitP(loss | made)
600.9723980.40.3629
800.90171154.10.3130
1000.77771339.10.2035
1200.61941420.60.0000
1400.45681344.20.0000
1600.31421136.10.0000
mass beyond 1200 orders: 7.77e-15
查看完整验算程序
# Chapter 5. p51: where to put the threshold. A batch has a fixed
# cost FIX and each unit earns margin MAR, so the break-even count is
# KMIN = ceil(FIX / MAR). Orders N follow a negative binomial (mean
# MU, doubt shape A, as in p33) and are capped at the stock limit CAP.
# For each threshold K: chance the batch is made, expected profit
# E[(MAR x min(N, CAP) - FIX); N >= K], and the chance that a made
# batch still loses money. Illustrative: FIX = 6000, MAR = 50,
# MU = 140, A = 8, CAP = 220. Exact sums with fsum.
import math

FIX, MAR, MU, A, CAP = 6000.0, 50.0, 140.0, 8.0, 220
KMIN = math.ceil(FIX / MAR)
TOP = 1200


def nb_pmf(j):
    la, lm = math.log(A / (A + MU)), math.log(MU / (A + MU))
    return math.exp(math.lgamma(j + A) - math.lgamma(A)
                    - math.lgamma(j + 1) + A * la + j * lm)


if __name__ == "__main__":
    ps = [nb_pmf(j) for j in range(TOP + 1)]
    print("FIX = {:.0f}  margin = {:.0f}  break-even KMIN = {}".format(
        FIX, MAR, KMIN))
    print("orders ~ NB(mean {:.0f}, shape {:.0f}), stock cap {}".format(
        MU, A, CAP))
    print("   K   P(made)   E profit   P(loss | made)")
    for k in (60, 80, 100, 120, 140, 160):
        made = math.fsum(ps[k:])
        prof = math.fsum(ps[j] * (MAR * min(j, CAP) - FIX)
                         for j in range(k, TOP + 1))
        loss = math.fsum(ps[k:max(k, KMIN)])
        print("{:4d} {:9.4f} {:10.1f} {:16.4f}".format(
            k, made, prof, loss / made))
    print("mass beyond {} orders: {:.2e}".format(TOP, 1 - math.fsum(ps)))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《保证合约预售成团阈值、退款承诺与公共品搭便车均衡母典》第 5 章「门槛设在哪里:保本线、敏感度与试销数据的分量」(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-316 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026316,
  author    = {Fan, Yafu},
  title     = {Assurance Contracts and Pre-Sale Group Thresholds: Refund Promises and Free-Riding Equilibria in Public Goods},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论