关于「有限尺寸标度:互推网络的门槛在小站群里被抹平」,《跨尺度转化弹性》给出的定量结论是什么?

结论:第五章用有限尺寸标度处理站间互推网络的门槛,重新标度后不同规模间的最大差异从 3.949 倍缩到 1.297 倍。

数据口径:书中站数、流量、转化、成本与经营者均为示意;全部计算由书中程序生成;本书是非个人化经营参考,不构成对任何具体经营决策的建议,只讨论以真实主体运营、内容彼此不同的合规站群。 第一章定义转化弹性,示意中含固定开销的成本弹性从 0.158 升到 0.949,含交叉耦合的订单弹性从 0.980 塌到 0.329,饱和池里拆分试验的估计是真实铺开效果的 25.0 倍。

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: 1.297, 3.949

x = (c-1)N^(1/3)-1.0 +0.0 +1.0 +2.0
raw S/N, N125 0.1043 0.1756 0.3006 0.4902
raw S/N, N1000 0.0505 0.0923 0.1647 0.2832
raw S/N, N8000 0.0264 0.0450 0.0889 0.1590
S/N^(2/3), N125 0.522 0.878 1.503 2.451
S/N^(2/3), N1000 0.505 0.923 1.647 2.832
S/N^(2/3), N8000 0.528 0.901 1.779 3.180
largest max/min ratio across N: raw 3.949, rescaled 1.297
查看完整验算程序
# Chapter 5. p52: finite-size scaling collapse of the referral cluster.
# Model as in p51 (random referral links, mean degree c). Near c = 1
# finite-size scaling predicts S_max = N^(2/3) f(x) with the scaling
# variable x = (c - 1) N^(1/3) (critical window of the random graph).
# For each x we set c = 1 + x N^(-1/3) and average 300 seeded graphs,
# seed 20261119. Raw shares S/N differ by N; the rescaled S/N^(2/3)
# should agree across N if the scaling form holds.
import math
import random

rng = random.Random(20261119)


def largest(n, m):
    parent = list(range(n))
    size = [1] * n

    def find(x):
        while parent[x] != x:
            parent[x] = parent[parent[x]]
            x = parent[x]
        return x

    for _ in range(m):
        a, b = rng.randrange(n), rng.randrange(n - 1)
        if b >= a:
            b += 1
        ra, rb = find(a), find(b)
        if ra != rb:
            if size[ra] < size[rb]:
                ra, rb = rb, ra
            parent[rb] = ra
            size[ra] += size[rb]
    return max(size[find(i)] for i in range(n))


XS = (-1.0, 0.0, 1.0, 2.0)
NS = (125, 1000, 8000)
raw, res = {}, {}
for n in NS:
    for x in XS:
        c = 1 + x * n ** (-1 / 3)
        v = [largest(n, round(c * n / 2)) for _ in range(300)]
        s = math.fsum(v) / len(v)
        raw[n, x], res[n, x] = s / n, s / n ** (2 / 3)
print("x = (c-1)N^(1/3):     " + "    ".join("%+5.1f" % x for x in XS))
for n in NS:
    print("raw S/N,  N = %4d   " % n
          + "   ".join("%6.4f" % raw[n, x] for x in XS))
for n in NS:
    print("S/N^(2/3), N = %4d  " % n
          + "   ".join("%6.3f" % res[n, x] for x in XS))


def spread(tab):
    return max(max(tab[n, x] for n in NS) / min(tab[n, x] for n in NS)
               for x in XS)


print("largest max/min ratio across N: raw %.3f, rescaled %.3f"
      % (spread(raw), spread(res)))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《重整化群跨尺度标度律、一人作坊到百站矩阵的转化弹性守恒与经验法则崩塌母典》第 5 章「有限尺寸标度:互推网络的门槛在小站群里被抹平」(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-248 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026248,
  author    = {Fan, Yafu},
  title     = {Renormalisation Group and Cross-Scale Scaling Laws: Which Conversion Elasticities Survive from a One-Person Workshop to a Hundred-Site Matrix},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论