关于「混合捆绑与单行本该去哪里」,《书库捆绑、平台限流与直售》给出的定量结论是什么?

结论:第二章说明在同一家店里没有混合菜单胜过全集,而平台上的单行本在没有重叠时让总收入增加到 1.4026 倍。

数据口径:书中的读者、需求、比率与参数均为示意模型;全部计算由书中程序生成;书中引用的平台规则于 2026 年 11 月 11 日核对;本书不是法律、财务或税务方面的建议。 书中以标注框区分已发表的证据、可运行的 Python 计算(16 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: 1.4026

direct shop alone: 281029
overlap xdirectplatformtotalvs direct alone
0.002810291131543941831.4026
0.252574881242983817861.3585
0.502318191356673674861.3076
1.001891261566943458201.2305
查看完整验算程序
# Chapter 2. p04: Singles belong in another shop. The direct shop sells
# the catalogue at USD 140 to readers drawn as in p03 and keeps 0.95
# of it. A platform shop reaches 20000 other readers each wanting 1 to 3
# volumes (median USD 9, spread 0.5) and sells singles at USD 9.99, of
# which the author keeps 0.70 (9.99 - 0.30) = 6.783. A share x of the
# direct shop's readers can also see the platform singles and buy there
# instead if that gives a larger surplus. Revenue for 4000 direct and
# 20000 platform readers, seed 115.
import math
import random

SIZES = (17, 17, 17, 17, 16, 16)
CAT, SINGLE, KEEP_D = 140.0, 9.99, 0.95
ROYALTY = 0.70 * (9.99 - 0.30)

def direct_readers(n=4000):
    rng = random.Random(115)
    out = []
    for _ in range(n):
        if rng.random() < 0.35:
            vals = [math.exp(math.log(9.0) + 0.5 * rng.gauss(0, 1))
                    for _ in range(3)]
            out.append(vals)
            continue
        k = 1 if rng.random() < 0.6 else 2
        like = rng.sample(range(6), k)
        vals = []
        for t, size in enumerate(SIZES):
            hit, med = (0.7, 7.0) if t in like else (0.15, 3.0)
            vals += [math.exp(math.log(med) + 0.8 * rng.gauss(0, 1))
                     for _ in range(size) if rng.random() < hit]
        out.append(vals)
    return out

def platform_readers(n=20000):
    rng = random.Random(1150)
    return [[math.exp(math.log(9.0) + 0.5 * rng.gauss(0, 1))
             for _ in range(rng.choice((1, 2, 3)))] for _ in range(n)]

def run(x, with_platform):
    rng = random.Random(7)
    direct, plat = [], []
    for vals in direct_readers():
        see = with_platform and rng.random() < x
        s_cat = math.fsum(vals) - CAT
        s_one = math.fsum(v - SINGLE for v in vals if v >= SINGLE)
        n_one = len([v for v in vals if v >= SINGLE])
        if see and s_one > max(s_cat, 0.0):
            plat.append(ROYALTY * n_one)
        elif s_cat > 0:
            direct.append(KEEP_D * CAT)
    if with_platform:
        for vals in platform_readers():
            plat.append(ROYALTY * len([v for v in vals if v >= SINGLE]))
    return math.fsum(direct), math.fsum(plat)

if __name__ == "__main__":
    d0, _ = run(0.0, False)
    print("direct shop alone: {:.0f}".format(d0))
    print("overlap x   direct     platform    total    vs direct alone")
    for x in (0.0, 0.25, 0.5, 1.0):
        d, p = run(x, True)
        print("{:9.2f}   {:8.0f}   {:8.0f}   {:8.0f}   {:15.4f}".format(
            x, d, p, d + p, (d + p) / d0))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《百卷数字专著库混合捆绑、平台限流装箱与主权直售套利数学》第 2 章「混合捆绑与单行本该去哪里」(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-115 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026115,
  author    = {Fan, Yafu},
  title     = {Large-Scale Monograph Catalog Bundling, Platform Upload Throttling and Sovereign Direct-to-Reader Arbitrage},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论