关于「百名垂直达人:块状相关、单次合作的固定成本与风险集中度」,《达人矩阵的组合方差》给出的定量结论是什么?

结论:第六章落到百名垂直达人:单次合作固定成本 1000 元时,一百人矩阵的夏普比率是 -0.2320;一半预算给顶流时,顶流承担 0.80 的风险。

数据口径:书中借用组合理论的方差、相关、夏普比率与均值—方差前沿,只借其算术,讨论的是营销预算的分配,不是证券投资;本书不构成任何个人化的理财或投资建议,不推荐任何金融产品。 书中的达人、回报、相关、成本与预算均为示意;全部计算由书中程序生成;所有合作都须如实标注广告,不讨论购买流量或虚假评价。

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: -0.2320, 1000

fixed C best n mean sd Sharpe | best n mean utility
0100 0.2400 0.2413 0.6630 | 100 0.2400 0.1526
50020 0.2193 0.2714 0.5134 | 30 0.2090 0.1110
100014 0.2111 0.2891 0.4533 | 20 0.1987 0.0919
200010 0.1987 0.3035 0.3910 | 14 0.1821 0.0627
50005 0.1883 0.3818 0.2837 | 10 0.1367 0.0124
Sharpe by n at C1000:
n1 fixed share 0.002 mean 0.2379 sd 0.7987 Sharpe 0.1977
n5 fixed share 0.008 mean 0.2297 sd 0.3951 Sharpe 0.3788
n10 fixed share 0.017 mean 0.2193 sd 0.3087 Sharpe 0.4513
n20 fixed share 0.033 mean 0.1987 sd 0.2668 Sharpe 0.4448
n30 fixed share 0.050 mean 0.1780 sd 0.2490 Sharpe 0.3936
n50 fixed share 0.083 mean 0.1367 sd 0.2296 Sharpe 0.2468
n100 fixed share 0.167 mean 0.0333 sd 0.2011 Sharpe -0.2320
查看完整验算程序
# Chapter 6. p62: fixed cost per collaboration sets the number of creators.
# Model: budget B = 600,000 yuan; each collaboration costs a fixed C yuan
# (brief, contract, samples, content review, ad-label check) on top of
# its media fee. n creators are spread round-robin over 10 verticals
# (block correlation RW 0.30 / RB 0.06, sd 0.80, mean 0.24 on the media
# money). Net return per budget yuan: (1 - nC/B)(1 + r) - 1. We pick n
# (1 to 100) for the best Sharpe over the hurdle 0.08 and for the best
# mean - (LAM / 2) * variance with LAM = 3. Illustrative (示意).
import math

B, MU, SD, RW, RB, V = 600000, 0.24, 0.80, 0.30, 0.06, 10
HURDLE, LAM = 0.08, 3.0


def stats(n, c):
    ks = [n // V + (1 if v < n % V else 0) for v in range(V)]
    same = math.fsum(k * (k - 1) for k in ks)
    sq = math.fsum(k * k for k in ks)
    var = SD * SD * (n + RW * same + RB * (n * n - sq)) / (n * n)
    keep = 1.0 - n * c / B
    return keep * (1.0 + MU) - 1.0, keep * math.sqrt(var)


print("  fixed C   best n    mean      sd  Sharpe | best n    mean"
      "  utility")
for c in (0, 500, 1000, 2000, 5000):
    rows = [(n,) + stats(n, c) for n in range(1, 101) if n * c < B]
    bs = max(rows, key=lambda r: (r[1] - HURDLE) / r[2])
    bu = max(rows, key=lambda r: r[1] - LAM / 2 * r[2] ** 2)
    print("%9d %8d %7.4f %7.4f %7.4f | %6d %7.4f %8.4f"
          % (c, bs[0], bs[1], bs[2], (bs[1] - HURDLE) / bs[2],
             bu[0], bu[1], bu[1] - LAM / 2 * bu[2] ** 2))
print("Sharpe by n at C = 1000:")
for n in (1, 5, 10, 20, 30, 50, 100):
    m, s = stats(n, 1000)
    print("  n=%-3d fixed share %.3f  mean %.4f  sd %.4f  Sharpe %.4f"
          % (n, n * 1000 / B, m, s, (m - HURDLE) / s))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《达人矩阵头部集中度与中腰部组合套利、尾部方差夏普比率与百名垂直达人投资组合母典》第 6 章「百名垂直达人:块状相关、单次合作的固定成本与风险集中度」(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-276 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026276,
  author    = {Fan, Yafu},
  title     = {Head Concentration versus Mid-Tier Portfolios in Creator Matrices: Tail Variance, Sharpe Ratios and the Allocation of a Budget across a Hundred Niche Creators},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论