What does Sizing Systems and Return Minimisation conclude, in numbers, about “A Label Is Not a Measurement: Brand Dispersion of the Same "M" as a Random Effect”?

Answer: Chapter 1 writes brand offsets as a random effect: with an offset standard deviation of 2 cm, two brands' M differ by more than a full grade with probability 0.1573, and one customer needs 2.147 different labels across five brands.

Scope of the numbers: Its brands, shops, prices and body dimensions are illustrative; all calculations are generated by its programs; the book is not individual business advice, deals only with the measurement of garment sizes, and makes no health judgement. Marked boxes separate published evidence, runnable Python calculations (21 programs), classical sources and the author's hypotheses.

中文版

Data table

The numbers below are printed verbatim by the chapter's Python program and can be reproduced by running it. standard-library Python onlyprinted values that match the answer: 0.1573

E[range of 20 std normals] = 3.7350
tauP(|d|>2)P(|d|>4)P(|d|>6)range20sd(d)
1.00.15730.00470.00003.731.414
1.50.34580.05930.00475.602.121
2.00.47950.15730.03397.472.828
2.50.57160.25790.08979.343.536
3.00.63740.34580.157311.204.243
At tau = 2.0 cm one grade (4 cm) is 2.00 brand sd;
a label is a brand-local code, not a body measurement.
Show the full program
# Chapter 1. p11: How far apart two brands' "M" can be, by brand spread.
# Model: the finished chest of label M in brand b is 106 + u_b cm, with
# u_b ~ Normal(0, tau^2) a brand random effect. All values illustrative
# here. Kinley (2003) measured 1,011 pairs of trousers and found
# inconsistency inside every size category, which motivates tau > 0.
# Exact formulas: diff of two brands ~ Normal(0, 2 tau^2); the expected
# range of J brands is tau times E[range of J standard normals], which
# is computed by numerical integration (no simulation).
import math
from statistics import NormalDist

Z = NormalDist()
J = 20          # brands on one marketplace page (illustrative)


def expected_range(n, steps=4000, lo=-8.0, hi=8.0):
    # E[R] = integral of 1 - F(x)^n - (1 - F(x))^n over the real line
    h = (hi - lo) / steps
    vals = []
    for i in range(steps + 1):
        x = lo + i * h
        f = Z.cdf(x)
        w = 1.0 if i in (0, steps) else (4.0 if i % 2 else 2.0)
        vals.append(w * (1.0 - f ** n - (1.0 - f) ** n))
    return math.fsum(vals) * h / 3.0


ER = expected_range(J)
print("E[range of %d std normals] = %.4f" % (J, ER))
print("tau   P(|d|>2)  P(|d|>4)  P(|d|>6)  range%d  sd(d)" % J)
for tau in (1.0, 1.5, 2.0, 2.5, 3.0):
    sd = tau * math.sqrt(2.0)
    p = [2.0 * (1.0 - Z.cdf(k / sd)) for k in (2.0, 4.0, 6.0)]
    print("%.1f   %.4f    %.4f    %.4f    %6.2f   %.3f"
          % (tau, p[0], p[1], p[2], tau * ER, sd))
tau = 2.0
print("At tau = %.1f cm one grade (4 cm) is %.2f brand sd;" % (tau, 4 / tau))
print("a label is a brand-local code, not a body measurement.")

Single source and citation

Fan, Y. (2026). Anthropometric Principal-Component Sizing, Vanity-Size Drift and Cross-Border Return-Rate Minimisation (Version 1.0.0), Chapter 1: A Label Is Not a Measurement: Brand Dispersion of the Same "M" as a Random Effect. 5D Library. DOI pending

DOI pending · monograph page goes live after publication · 5DL-B-215 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026215,
  author    = {Fan, Yafu},
  title     = {Anthropometric Principal-Component Sizing, Vanity-Size Drift and Cross-Border Return-Rate Minimisation},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

Other chapter answers in this book