What does Beyond Distillation conclude, in numbers, about “The Live Probe: CUSUM, False-Alarm Run Lengths and Writing Back into the Mother Book”?

Answer: Chapter 5 watches conversion and error rates with cumulative-sum probes and writes alarms back into the mother book, which avoids 0.7764 of recurring anomalies in simulation.

Scope of the numbers: Parameters are illustrative and all calculations are generated by the book's programs. This is general business analysis, not investment, legal or medical advice, and the code it generates is teaching material deployed to no site.

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: 0.7764

anomalies per year, A (patch the book)12.49
anomalies per year, B (symptom only)55.86
share avoided by A0.7764
expansion tasks per year under A5.98
patch types among the six classes:
exception-barring1
lemma-incorporation3
monster-barring1
real refutation1
chance each class shows up within a year (min, max): 0.9844 1.0000
Show the full program
# Chapter 5. p53: turning probe alarms into patches of the mother book.
# Six latent anomaly classes fire with weekly rates (illustrative). Each
# class has a Lakatos patch type: monster-barring (shrink the theorem's
# domain), exception-barring (list the exception), lemma-incorporation
# (add the hidden condition to the theorem) or real refutation (the rule
# is wrong and is rewritten). Policy A opens an expansion task at the
# first alarm of a class; the patch lands after 1 to 3 weeks and cuts the
# class rate to a tenth. Policy B only fixes the symptom in the organ:
# the rate stays. 52 weeks, 500 seeded replications per policy.
import math
import random
import statistics

CLASSES = [
    ("price outside band", 0.30, "monster-barring"),
    ("fee table changed", 0.25, "lemma-incorporation"),
    ("tap target on tablet", 0.20, "lemma-incorporation"),
    ("unit rounding", 0.15, "exception-barring"),
    ("refund path loop", 0.10, "real refutation"),
    ("late payment notice", 0.08, "lemma-incorporation"),
]


def poisson(lam, rng):
    lim, k, p = math.exp(-lam), 0, 1.0
    while True:
        p *= rng.random()
        if p <= lim:
            return k
        k += 1


def year(policy, rng):
    rates = [c[1] for c in CLASSES]
    due = [None] * len(CLASSES)
    total, tasks = 0, 0
    for week in range(52):
        for i in range(len(CLASSES)):
            if due[i] is not None and week >= due[i]:
                rates[i] = CLASSES[i][1] / 10.0
            n = poisson(rates[i], rng)
            total += n
            if n and policy == "A" and due[i] is None:
                due[i] = week + rng.randint(1, 3)
                tasks += 1
    return total, tasks


if __name__ == "__main__":
    ra = random.Random(37871)
    rb = random.Random(37872)
    a = [year("A", ra) for _ in range(500)]
    b = [year("B", rb) for _ in range(500)]
    ma = statistics.fmean(x[0] for x in a)
    mb = statistics.fmean(x[0] for x in b)
    print("anomalies per year, A (patch the book)  {:.2f}".format(ma))
    print("anomalies per year, B (symptom only)    {:.2f}".format(mb))
    print("share avoided by A                      {:.4f}".format(
        1 - ma / mb))
    print("expansion tasks per year under A        {:.2f}".format(
        statistics.fmean(x[1] for x in a)))
    kinds = {}
    for name, r, kind in CLASSES:
        kinds[kind] = kinds.get(kind, 0) + 1
    print("patch types among the six classes:")
    for kind in sorted(kinds):
        print("  {:<20} {}".format(kind, kinds[kind]))
    p_seen = [1 - math.exp(-52 * c[1]) for c in CLASSES]
    print("chance each class shows up within a year (min, max): "
          "{:.4f} {:.4f}".format(min(p_seen), max(p_seen)))

Single source and citation

Fan, Y. (2026). Beyond Distillation: Executable Business-Organ Synthesis, Generative Code Crystallisation and Living-System Bootstrapping (Version 1.0.0), Chapter 5: The Live Probe: CUSUM, False-Alarm Run Lengths and Writing Back into the Mother Book. 5D Library. DOI pending

DOI pending · monograph page goes live after publication · 5DL-B-378 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026378,
  author    = {Fan, Yafu},
  title     = {Beyond Distillation: Executable Business-Organ Synthesis, Generative Code Crystallisation and Living-System Bootstrapping},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

Other chapter answers in this book