What does Likelihood-Ratio Screening of Luxury Goods conclude, in numbers, about “East and West in the Sources: Suspecting the Neighbour of the Axe, and Peirce's Thermometer of Belief”?

Answer: Chapter 7 examines the suspected neighbour of the Liezi and Peirce's log odds of 1878.

Scope of the numbers: Brands are named only generically, the parameters are illustrative, all calculations are generated by the book's programs, and the book is not individual business advice. Chapter 3 gives the effective number of independent features (four features are worth only 2.007 under the illustrative matrix) and a one-number slope of 0.571.

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 only

naive4 cues x LR 3 = LR 81 (19.08 deciban), P(guilt) 0.8100
rho n_eff exact LR deciban P(guilt)
0.04.00 81.00 19.08 0.8100
0.32.11 10.10 10.04 0.3472
0.61.43 4.80 6.82 0.2018
0.91.08 3.28 5.16 0.1472
1.01.00 3.00 4.77 0.1364
axe found in the valley (LR 0.001): P(guilt) from naive 0.0042
if the cues came only from suspicion they carry no weight:
LR of the cues given the facts1.00, P(guilt) stays 0.05
the second look (same son, belief reset): every cue read as LR 1
Show the full program
# Chapter 7. p71: the man who lost his axe, as a correlated-evidence
# model. Four cues (gait, face, speech, bearing) each read as a likelihood
# ratio of 3 toward "the neighbour's son took it". If the four readings
# share one source (the observer's own suspicion) with correlation rho,
# the exact combined ratio is 81 ** (1 / (1 + 3 rho)) for equal cues
# (Chapter 2, p23). Prior of guilt 0.05 is illustrative. The found axe is
# modelled as evidence with ratio 0.001 against guilt.
import math

N_CUE, LR_CUE, PRIOR = 4, 3.0, 0.05
DB = 10.0 / math.log(10.0)


def post(prior, lr):
    o = prior / (1.0 - prior) * lr
    return o / (1.0 + o)


naive = LR_CUE ** N_CUE
print("naive: %d cues x LR %.0f = LR %.0f (%.2f deciban), P(guilt) %.4f"
      % (N_CUE, LR_CUE, naive, DB * math.log(naive), post(PRIOR, naive)))
print("  rho   n_eff  exact LR  deciban  P(guilt)")
for rho in [0.0, 0.3, 0.6, 0.9, 1.0]:
    shrink = 1.0 + (N_CUE - 1) * rho
    lr = naive ** (1.0 / shrink)
    print("%5.1f %7.2f %9.2f %8.2f %9.4f"
          % (rho, N_CUE / shrink, lr, DB * math.log(lr), post(PRIOR, lr)))
found = 0.001
print("axe found in the valley (LR %.3f): P(guilt) from naive %.4f"
      % (found, post(post(PRIOR, naive), found)))
print("if the cues came only from suspicion they carry no weight:")
print("  LR of the cues given the facts = 1.00, P(guilt) stays %.2f"
      % PRIOR)
print("the second look (same son, belief reset): every cue read as LR 1")

Single source and citation

Fan, Y. (2026). Multi-Feature Likelihood-Ratio Screening of Pre-Owned Luxury Goods: Fusing Stitch Pitch, Hardware Spectra and Leather Grain, and the Bayesian Boundary between Lab Referral and Instant Purchase (Version 1.0.0), Chapter 7: East and West in the Sources: Suspecting the Neighbour of the Axe, and Peirce's Thermometer of Belief. 5D Library. DOI pending

DOI pending · monograph page goes live after publication · 5DL-B-281 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026281,
  author    = {Fan, Yafu},
  title     = {Multi-Feature Likelihood-Ratio Screening of Pre-Owned Luxury Goods: Fusing Stitch Pitch, Hardware Spectra and Leather Grain, and the Bayesian Boundary between Lab Referral and Instant Purchase},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

Other chapter answers in this book