关于「平台突变与自有名录:泊松冲击下的受众存量与损失分位数」,《声誉反脆弱与受众主权》给出的定量结论是什么?

结论:第六章用泊松冲击与马尔可夫流转比较两种受众策略,二十四个月收入损失的第 99 百分位从只靠平台的 0.793 降到持续沉淀自有名录的 0.437。

数据口径:书中参数为示意;全部计算由书中程序生成;本书是非个人化经营建议,不是法律意见。 书中以标注框区分已发表的证据、可运行的 Python 计算(22 段程序)、原典与作者的假说。

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: 0.437, 0.793, 99

A. 24-month revenue loss, 10,000 runs
   strategy  mean    P50     P90     P99     P(loss > 0.30)
   A         0.152   0.084   0.454   0.793   0.1750
   B         0.085   0.047   0.250   0.437   0.0639
B. worst single month, share of normal revenue
   A: P10 0.119, P1 0.029, P(below 0.5) 0.5296
   B: P10 0.513, P1 0.451, P(below 0.5) 0.0709
查看完整验算程序
# Chapter 6. p62: loss quantiles under random platform shocks.
# Over 24 months, platform shocks arrive as a Poisson stream with
# rate 0.6 per year. Each cuts platform reach by D ~ Uniform(0.2,
# 0.9); with chance 0.7 it is a restriction lasting 3 months, else a
# lasting rule change. The owned list has its own, milder shocks
# (rate 0.1 per year, D ~ Uniform(0.1, 0.3), 2 months), since email
# delivery can also fail. Revenue is proportional to reach. Strategy
# A: all reach on the platform. B: list share 0.451 (p61, month 36).
# Loss = 1 - revenue / revenue without shocks. Illustrative values.
import math
import random

MONTHS, RUNS = 24, 10000
LAM_P, LAM_L = 0.6 / 12, 0.1 / 12


def channel_path(rng, lam, dlo, dhi, p_temp, temp_len):
    mult = [1.0] * MONTHS
    for m in range(MONTHS):
        if rng.random() < lam:
            d = rng.uniform(dlo, dhi)
            temp = rng.random() < p_temp
            end = min(MONTHS, m + temp_len) if temp else MONTHS
            for k in range(m, end):
                mult[k] *= (1.0 - d)
    return mult


def quantile(xs, q):
    s = sorted(xs)
    return s[min(len(s) - 1, int(q * len(s)))]


if __name__ == "__main__":
    rng = random.Random(40601)
    out = {"A": [], "B": []}
    worst = {"A": [], "B": []}
    for _ in range(RUNS):
        plat = channel_path(rng, LAM_P, 0.2, 0.9, 0.7, 3)
        lst = channel_path(rng, LAM_L, 0.1, 0.3, 1.0, 2)
        for name, s in (("A", 0.0), ("B", 0.451)):
            rev = [(1.0 - s) * p + s * q for p, q in zip(plat, lst)]
            out[name].append(1.0 - math.fsum(rev) / MONTHS)
            worst[name].append(min(rev))
    print("A. 24-month revenue loss, {:,} runs".format(RUNS))
    print("   strategy  mean    P50     P90     P99     P(loss > 0.30)")
    for name in ("A", "B"):
        xs = out[name]
        print("   {:<8}  {:.3f}   {:.3f}   {:.3f}   {:.3f}   {:.4f}".format(
            name, math.fsum(xs) / RUNS, quantile(xs, 0.5),
            quantile(xs, 0.9), quantile(xs, 0.99),
            len([x for x in xs if x > 0.30]) / RUNS))
    print("B. worst single month, share of normal revenue")
    for name in ("A", "B"):
        w = worst[name]
        print("   {}: P10 {:.3f}, P1 {:.3f}, P(below 0.5) {:.4f}".format(
            name, quantile(w, 0.10), quantile(w, 0.01),
            len([x for x in w if x < 0.5]) / RUNS))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《商业声誉反脆弱柔道、恶意差评证据链反制与去平台依赖私域主权母典》第 6 章「平台突变与自有名录:泊松冲击下的受众存量与损失分位数」(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-403 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026403,
  author    = {Fan, Yafu},
  title     = {Antifragile Reputation Judo, Evidence-Backed Smear Neutralisation and Platform-Independent Audience Sovereignty},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论