What does Drift Control for Multi-Window Writing conclude, in numbers, about “Practice: Four Windows Write a Nine-Chapter Book, Five Classes of Injected Error and the Detection Ledger of Four Gates”?

Answer: Chapter 8 has four windows write a synthetic nine-chapter book with 119 injected errors in five classes: the four gates stop 0.723, 28 of the 33 escapes belong to techniques the machine cannot see by construction, and without them the detection rate is 0.945; on the clean book the first term run raises 1672 false alarms, which fall to zero after a person reviews 48 words; a serious injection audit needs at least 75 errors per class. Marked boxes separate published evidence, runnable calculations, sources and the author's hypotheses, each stating what would refute it.

Scope of the numbers: Every error rate without a source is an assumed parameter or a measurement on synthetic data; the book is not legal advice and recommends no writing tool or model. Marked boxes separate published evidence, runnable calculations, sources and the author's hypotheses, each stating what would refute it.

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: 0.723, 0.945, 119, 28, 48, 75

class caught n recall95% exact interval
term drift18 24 0.750 [0.533, 0.902]
number mismatch18 24 0.750 [0.533, 0.902]
duplicate sentence12 24 0.500 [0.291, 0.709]
citation hallucination16 20 0.800 [0.563, 0.943]
style degradation22 27 0.815 [0.619, 0.937]
all classes86 119 0.723 [0.633, 0.801]
excluding the 28 injections no machine gate is built to see86 / 91 = 0.945 [0.876, 0.982]
pre-registration: true recall 0.95, pass line: exact lower bound >0.85
n24: need >= 24 caught, probability of passing 0.292
n50: need >= 48 caught, probability of passing 0.541
n75: need >= 70 caught, probability of passing 0.828
smallest n per class with probability >0.80: 75
Show the full program
# Chapter 8. p81: exact (Clopper-Pearson) 95% intervals for the recalls p80 measured, and how many
# injected errors per class a pre-registered audit needs.
# Counts are copied from the output of p80 (caught by any gate / injected). The interval [lo, hi]
# solves P(X >= x | n, lo) = 0.025 and P(X <= x | n, hi) = 0.025, found by bisection.
# Sample size: the smallest n per class such that, if the true recall is 0.95, the exact 95% lower
# bound clears 0.85 with probability at least 0.80 (the audit's pass line is fixed before any run).
import math

COUNTS = [("term drift", 18, 24), ("number mismatch", 18, 24), ("duplicate sentence", 12, 24),
          ("citation hallucination", 16, 20), ("style degradation", 22, 27), ("all classes", 86, 119)]
BY_DESIGN = 6 + 6 + 6 + 6 + 4      # unrecorded synonym, wrong row, paraphrase, short stock, wrong claim


def tail_ge(x, n, p):
    return math.fsum(math.comb(n, k) * p ** k * (1 - p) ** (n - k) for k in range(x, n + 1))


def tail_le(x, n, p):
    return math.fsum(math.comb(n, k) * p ** k * (1 - p) ** (n - k) for k in range(0, x + 1))


def bisect(f, lo=0.0, hi=1.0):
    for _ in range(60):
        mid = (lo + hi) / 2
        if f(mid):
            hi = mid
        else:
            lo = mid
    return (lo + hi) / 2


def cp(x, n, a=0.05):
    lo = 0.0 if x == 0 else bisect(lambda p: tail_ge(x, n, p) >= a / 2)
    hi = 1.0 if x == n else bisect(lambda p: tail_le(x, n, p) < a / 2)
    return lo, hi


def power(n, true_r=0.95, line=0.85):
    xs = [x for x in range(n + 1) if cp(x, n)[0] >= line]
    if not xs:
        return 0.0, None
    return tail_ge(min(xs), n, true_r), min(xs)


if __name__ == "__main__":
    print("class                     caught    n  recall   95% exact interval")
    for name, x, n in COUNTS:
        lo, hi = cp(x, n)
        print("{:24s} {:6d} {:4d}   {:.3f}   [{:.3f}, {:.3f}]".format(name, x, n, x / n, lo, hi))
    x, n = 86, 119 - BY_DESIGN
    lo, hi = cp(x, n)
    print("excluding the {} injections no machine gate is built to see: {} / {} = {:.3f} [{:.3f}, {:.3f}]"
          .format(BY_DESIGN, x, n, x / n, lo, hi))
    print("pre-registration: true recall 0.95, pass line: exact lower bound >= 0.85")
    found = None
    for n in range(20, 400):
        pw, xmin = power(n)
        if n in (24, 50, 100) or (found is None and pw >= 0.80):
            print("   n = {:3d}: need >= {} caught, probability of passing {:.3f}".format(n, xmin, pw))
        if found is None and pw >= 0.80:
            found = n
            break
    print("   smallest n per class with probability >= 0.80: {}".format(found))

Single source and citation

Fan, Y. (2026). Drift Control for Parallel Multi-Window Book Writing, Cross-Volume Terminology Graph Audits and Zero-Hallucination Citation Verifiers (Version 0.1.0), Chapter 8: Practice: Four Windows Write a Nine-Chapter Book, Five Classes of Injected Error and the Detection Ledger of Four Gates. 5D Library. DOI pending

DOI pending · monograph page goes live after publication · 5DL-B-152 · v0.1.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026152,
  author    = {Fan, Yafu},
  title     = {Drift Control for Parallel Multi-Window Book Writing, Cross-Volume Terminology Graph Audits and Zero-Hallucination Citation Verifiers},
  year      = {2026},
  publisher = {5D Library},
  edition   = {0.1.0},
  note      = {DOI pending}
}

Other chapter answers in this book