What does The Nightly Funnel conclude, in numbers, about “The Nightly Funnel Ledger: Field Tools and Checker Specifications, a Pre-registered Test and the Conditions of Refutation”?
Answer: Chapter 8 gives the nightly funnel ledger, the calculator, checkers grouped by the four kits and a pre-registered test: under the base prior the median cumulative gain in net contribution over half a year is about +0.082, with a probability of 0.093 of gaining nothing. Its parameters are illustrative; all calculations are generated by its 23 programs; it is general business analysis, not business, legal or financial advice, and it promises no rise in conversion.
Scope of the numbers: Its parameters are illustrative; all calculations are generated by its 23 programs; it is general business analysis, not business, legal or financial advice, and it promises no rise in conversion.
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: 0.082, 0.093
| prior | weeks | P5 | P50 | P95 | P(uplift <= 0) |
| base | 13 | +0.000 | +0.040 | +0.125 | 0.304 |
| base | 26 | +0.000 | +0.082 | +0.217 | 0.093 |
| lean | 13 | +0.000 | +0.000 | +0.030 | 0.721 |
| lean | 26 | +0.000 | +0.000 | +0.061 | 0.530 |
| base+fade | 13 | +0.000 | +0.040 | +0.082 | 0.497 |
| base+fade | 26 | +0.000 | +0.040 | +0.125 | 0.266 |
Show the full program
# Chapter 8. p82: what "self-improving conversion" can honestly claim.
# One tested change per week (p72 policy "one", SE 0.032, bar 1.645).
# Prior over true effects on net contribution per visitor:
# base 0 (0.60), +4% (0.25), -4% (0.15)
# lean 0 (0.75), +3% (0.10), -3% (0.15)
# base+fade base, but half of the positive effects fade to 0 after
# 4 weeks (novelty), while harmful ones stay
# Output: quantiles of the cumulative uplift after 13 and 26 weeks and
# the probability that the loop ends at or below where it started.
import math
import random
SE, ZBAR, REPS = 0.032, 1.645, 4000
rng = random.Random(40582)
PRIORS = {"base": [(0.60, 0.0), (0.25, 0.04), (0.15, -0.04)],
"lean": [(0.75, 0.0), (0.10, 0.03), (0.15, -0.03)]}
def draw(prior):
u = rng.random()
for i in range(len(prior)):
if u < math.fsum(p for p, _ in prior[:i + 1]):
return prior[i][1]
return prior[-1][1]
def run(name, weeks):
prior = PRIORS["base" if name == "base+fade" else name]
live = []
for w in range(weeks):
f = draw(prior)
if (f + SE * rng.gauss(0, 1)) / SE > ZBAR:
fades = name == "base+fade" and f > 0 and rng.random() < 0.5
live.append((f, w + 4 if fades else None))
logs = []
for f, end in live:
if end is None or end > weeks:
logs.append(math.log1p(f))
return math.expm1(math.fsum(logs))
print("prior weeks P5 P50 P95 P(uplift <= 0)")
for name in ["base", "lean", "base+fade"]:
for weeks in (13, 26):
xs = sorted(run(name, weeks) for _ in range(REPS))
le0 = len([x for x in xs if x <= 0]) / REPS
print("{:<10} {:>6} {:>+7.3f} {:>+7.3f} {:>+7.3f} {:>11.3f}"
.format(name, weeks, xs[REPS // 20], xs[REPS // 2],
xs[19 * REPS // 20], le0))Single source and citation
Fan, Y. (2026). Autonomous Nightly Funnel Micro-Surgery, Bandit Traffic Allocation and Zero-Touch Conversion Self-Improvement (Version 1.0.0), Chapter 8: The Nightly Funnel Ledger: Field Tools and Checker Specifications, a Pre-registered Test and the Conditions of Refutation. 5D Library. DOI pending
DOI pending · monograph page goes live after publication · 5DL-B-405 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026405,
author = {Fan, Yafu},
title = {Autonomous Nightly Funnel Micro-Surgery, Bandit Traffic Allocation and Zero-Touch Conversion Self-Improvement},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}Other chapter answers in this book
- Chapter 1 · The Funnel Check-up: Six-Step Beta Posteriors, Nightly Change Priorities and a Compliant Candidate Pool
- Chapter 2 · Thompson Sampling and Regret: from Two Small Samples in 1933 to the Even Split
- Chapter 3 · The Reward Function: Net Contribution per Visitor, Not the Conversion Rate
- Chapter 4 · Non-stationary Traffic: Weekday Effects, Discounted Memory and Sliding Windows
- Chapter 5 · Elimination Rules: Minimum Exposure, Posterior Win Probability and Time-uniform Confidence Sequences
- Chapter 6 · Trustworthy Data: Negative Bias in Adaptive Data, Sample Ratio Mismatch and Guardrail Rollback
- Chapter 7 · Classical Sources East and West: "At the Limit, Change" in the Great Commentary, the Sunzi on Weak and Strong Points, and Random Allocation in Peirce and Jastrow