关于「推荐树:分支过程的期望人数与有限市场的饱和」,《合规分销与佣金斜率》给出的定量结论是什么?

结论:第三章用分支过程说明推荐链靠单层奖励即可延伸,第一代只占六代以内新买家的 0.4196;有限市场里入会最多的一批人平均只带来 0.876 人。

数据口径:最优的线性佣金斜率由努力成本、风险厌恶与噪音决定,示意基准下是 0.5263;给上线按下线业绩分成的二级佣金,主要买到的是拉人头而不是真实零售;按本书采用的最保守读法,以下线业绩给上线计酬本身落在 2005 年《禁止传销条例》列举的行为之内,合规方案的二级斜率应取零,转介绍改由只为真实购买、退货期满才发放的单层顾客推荐奖励来争取。 书中参数为示意;全部计算由书中程序生成;本书是非个人化法律、税务或理财建议,法规只述大意,以上架前人工复核的现行文本为准。

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: 0.4196

expected new buyers per seed customer, by depth of the tree
R depth1 depth2 depth6 total extinct leaves
0.30.300 0.390 0.428 0.429 1.0000 0.7408
0.60.600 0.960 1.430 1.500 1.0000 0.5488
0.90.900 1.710 4.217 9.000 1.0000 0.4066
1.21.200 2.640 11.916 inf 0.6863 0.3012
Monte Carlo R0.6, six generations: mean 1.409
exact six-generation value1.430
share of six-generation buyers in depth 1 at R0.6: 0.4196
查看完整验算程序
# Chapter 3. p31: expected size of a referral tree, by depth.
# Galton-Watson branching (Watson and Galton 1875): every buying
# customer brings in a Poisson(R) number of new buying customers.
# Expected new buyers in generation k: R^k. Extinction probability q
# solves q = exp(R*(q - 1)). Share of buyers who refer nobody: e^-R.
# Monte Carlo check at R = 0.6 with a fixed seed (20,000 seeds, six
# generations). R values are illustrative.
import math
import random


def gens(R, depth):
    return [R ** k for k in range(1, depth + 1)]


def extinction(R):
    q = 0.0
    for _ in range(2000):
        q = math.exp(R * (q - 1.0))
    return q


def poisson(rng, lam):
    # Knuth 法:乘均匀数直到小于 e^-lam
    lim, k, p = math.exp(-lam), 0, 1.0
    while True:
        p *= rng.random()
        if p < lim:
            return k
        k += 1


print("expected new buyers per seed customer, by depth of the tree")
print("    R  depth1  depth2  depth6   total  extinct  leaves")
for R in (0.3, 0.6, 0.9, 1.2):
    g = gens(R, 6)
    tot = R / (1.0 - R) if R < 1.0 else float("inf")
    print("{:5.1f} {:7.3f} {:7.3f} {:7.3f} {:7.3f} {:8.4f} {:7.4f}".format(
        R, g[0], math.fsum(g[:2]), math.fsum(g), tot, extinction(R),
        math.exp(-R)))
rng = random.Random(20261113)
sizes = []
for _ in range(20000):
    alive, total = 1, 0
    for _ in range(6):
        nxt = 0
        for _ in range(alive):
            nxt += poisson(rng, 0.6)  # step
        total += nxt  # step
        alive = nxt
        if alive == 0:
            break
    sizes.append(total)
mean = math.fsum(sizes) / len(sizes)
print("Monte Carlo R = 0.6, six generations: mean {:.3f}".format(mean))
print("exact six-generation value:        {:.3f}".format(
    math.fsum(gens(0.6, 6))))
d1 = gens(0.6, 1)[0] / math.fsum(gens(0.6, 6))
print("share of six-generation buyers in depth 1 at R = 0.6: {:.4f}"
      .format(d1))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《合规分销激励相容、二级佣金斜率委托代理模型与反传销层级红线内转介绍最大化母典》第 3 章「推荐树:分支过程的期望人数与有限市场的饱和」(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-279 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026279,
  author    = {Fan, Yafu},
  title     = {Incentive-Compatible Compliant Referral Programmes: Commission Slopes in a Principal-Agent Model, a Second-Tier Slope of Zero and Maximal Referrals within Anti-Pyramid Limits},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论