What does Zero-Ticket Interfaces and Self-Service State Machines conclude, in numbers, about “The Empirical Loop: Field Tools and Checker Specifications”?
Answer: Chapter 8 gives the ledger, the distilled Ticket Root-Cause and Self-Service State Machine Checker, checker specifications grouped by four kits and a pre-registered test needing 1,089 orders per arm to detect a fall of three tenths. Marked boxes separate published evidence, runnable Python calculations (22 programs), classical sources and the author's hypotheses.
Scope of the numbers: Parameters are illustrative and all calculations come from the book's programs; the book is general business analysis, not individual or legal advice. Marked boxes separate published evidence, runnable Python calculations (22 programs), classical sources and the author's hypotheses.
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: 1089
| contacts 335 | self-served 171 | sent to human 164 |
| elimination rate 0.510 | 95% interval 0.457 to 0.564 | |
| completion among eligible 0.606 | ||
| week 1: contacts | 94 | elimination 0.500 |
| week 2: contacts | 76 | elimination 0.579 |
| week 3: contacts | 81 | elimination 0.519 |
| week 4: contacts | 84 | elimination 0.452 |
| checks: | ||
| C1 human reachable in <= 2 steps | green | worst page 2 steps |
| C2 recompute more than once | red | 1 order(s) |
| C3 rebind before verification | green | 0 row(s) |
| C4 genuine users blocked by lock | red | 0.0448 (7 locks) |
| C5 refund stopped by self-service | green | 0 row(s) |
| C6 rule change on n=600 per arm | red | needs 1,089 (p82) |
| human tickets per day 5.86; load 0.065 erlang |
Show the full program
# Chapter 8. p81: four weeks of the after-sale ledger and its checks
# 2,800 synthetic orders (100 a day). Each contact is a row: week,
# class, self-serviceable, finished by self-service, sent to a human,
# recompute count, rebind verified, lock hit (0.10 of not_received),
# genuine user blocked (0.4 of lock hits), refund stopped. Completion
# rates are the p21 mid-range. The checks and thresholds are the ones
# distilled into distill/tool.py. One order is given two recomputes on
# purpose, so that check C2 has something to catch.
import math
import random
NAMES = ["not_received", "wrong_input", "progress", "repeat_chase",
"link_expired", "refund", "content_q", "payment", "other"]
W = [0.22, 0.15, 0.20, 0.12, 0.08, 0.06, 0.07, 0.05, 0.05]
MID = {"not_received": 0.75, "wrong_input": 0.70, "progress": 0.60,
"repeat_chase": 0.675, "link_expired": 0.825, "content_q": 0.25}
STEPS_TO_HUMAN = {"order": 1, "delivery": 1, "resend": 2,
"recompute": 2, "progress": 1}
def wilson(k, n, z=1.96):
p = k / n
d = 1 + z * z / n
c = p + z * z / (2 * n)
h = z * math.sqrt(p * (1 - p) / n + z * z / (4 * n * n))
return (c - h) / d, (c + h) / d
rng = random.Random(40181)
rows = []
for day in range(28):
for _ in range(100):
if rng.random() >= 0.122:
continue
k = rng.choices(range(9), W)[0]
name = NAMES[k]
elig = name in MID
done = elig and rng.random() < MID[name]
lock = name == "not_received" and rng.random() < 0.10
rows.append({"week": day // 7 + 1, "cls": name, "elig": elig,
"self": done, "human": not done,
"recomputes": 1 if name == "wrong_input" else 0,
"verified": True, "lock": lock,
"false_block": lock and rng.random() < 0.4,
"refund_stopped": False})
rows[5]["recomputes"] = 2
n = len(rows)
done = 0
elig = 0
locks = 0
fb = 0
for r in rows:
done += r["self"]
elig += r["elig"]
locks += r["lock"]
fb += r["false_block"]
lo, hi = wilson(done, n)
print("contacts {} self-served {} sent to human {}".format(
n, done, n - done))
print("elimination rate {:.3f} 95% interval {:.3f} to {:.3f}".format(
done / n, lo, hi))
print("completion among eligible {:.3f}".format(done / elig))
for wk in range(1, 5):
a = 0
b = 0
for r in rows:
if r["week"] == wk:
a += 1
b += r["self"]
print(" week {}: contacts {:>3} elimination {:.3f}".format(
wk, a, b / a))
print("checks:")
worst = max(STEPS_TO_HUMAN.values())
twice = 0
for r in rows:
if r["recomputes"] > 1:
twice += 1
unver = 0
stopped = 0
for r in rows:
unver += not r["verified"]
stopped += r["refund_stopped"]
not_rec = 0
for r in rows:
not_rec += r["cls"] == "not_received"
fbr = fb / not_rec
checks = [
("C1 human reachable in <= 2 steps", "red" if worst > 2 else "green",
"worst page {} steps".format(worst)),
("C2 recompute more than once", "red" if twice else "green",
"{} order(s)".format(twice)),
("C3 rebind before verification", "red" if unver else "green",
"{} row(s)".format(unver)),
("C4 genuine users blocked by lock",
"red" if fbr > 0.02 else ("yellow" if fbr > 0.005 else "green"),
"{:.4f} ({} locks)".format(fbr, locks)),
("C5 refund stopped by self-service", "red" if stopped else "green",
"{} row(s)".format(stopped)),
("C6 rule change on n=600 per arm",
"red" if 600 < 1089 else "green", "needs 1,089 (p82)"),
]
for name, flag, why in checks:
print(" {:<35} {:<6} {}".format(name, flag, why))
a = (n - done) / 28 / 12 * 8 / 60
print("human tickets per day {:.2f}; load {:.3f} erlang".format(
(n - done) / 28, a))Single source and citation
Fan, Y. (2026). Zero-Ticket Self-Repairing Order Interfaces, Self-Service Correction State Machines and Pre-Emptive Support Elimination (Version 1.0.0), Chapter 8: The Empirical Loop: Field Tools and Checker Specifications. 5D Library. DOI pending
DOI pending · monograph page goes live after publication · 5DL-B-401 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026401,
author = {Fan, Yafu},
title = {Zero-Ticket Self-Repairing Order Interfaces, Self-Service Correction State Machines and Pre-Emptive Support Elimination},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}Other chapter answers in this book
- Chapter 1 · Root-Cause Clustering of Tickets: What a Month of After-Sale Messages Is Made Of
- Chapter 2 · The Ceiling on Elimination: Class Share Times Self-Service Completion
- Chapter 3 · The Self-Service Correction State Machine: Guards, Idempotency and Exhaustive Checking
- Chapter 4 · Abuse Limits and Rebind Verification: Stopping Scripts Without Stopping People
- Chapter 5 · One Timed Self-Service Recompute: Window Length, Precision Stamps and Misuse
- Chapter 6 · Progress Transparency and the Residual Human Desk: a Chase Model and Erlang C
- Chapter 7 · Classical Sources East and West: the Ant Hole and the Division of Mental Labour