What does Conformal Intervals for Orders conclude, in numbers, about “The Coverage Gap of Normal Intervals: Heavy-Tailed Orders Push Every Miss to the Right”?
Answer: The Pareto rows tell the same story. With tail index 2.1, where the variance barely exists, two-sided coverage is 0.9429, the right-tail miss rate is 0.0571 and the one-sided bound misses 0.0690 of the time. Note the last column as well: the share of histories, out of the eight thousand, whose true coverage is below 0.90. It is 0.1378 in the log-normal row with shape one and only 0.0770 in the normal row. The sample standard deviation of heavy-tailed data is highly unstable. A history that happens to contain no large order underestimates it, and the interval becomes absurdly narrow. The owner cannot see this, because she holds only one history.
Scope of the numbers: Parameters are illustrative; all calculations are generated by the book's programs; the book is general business analysis, not individualised business advice. Marked boxes separate published evidence, runnable Python calculations (23 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: 0.0571, 0.0690, 0.0770, 0.1378, 0.90, 0.9429
| normal-theory 95% PI, n = 30, 8000 training sets | |||||
| shape | param | two-sided | upper miss | 1-sided miss | P(<0.90) |
| normal | 0.0 | 0.9494 | 0.0252 | 0.0502 | 0.0770 |
| lognorm | 0.5 | 0.9467 | 0.0529 | 0.0738 | 0.0929 |
| lognorm | 1.0 | 0.9415 | 0.0585 | 0.0731 | 0.1378 |
| lognorm | 1.5 | 0.9431 | 0.0569 | 0.0673 | 0.1449 |
| pareto | 3.0 | 0.9416 | 0.0584 | 0.0724 | 0.1374 |
| pareto | 2.5 | 0.9422 | 0.0578 | 0.0708 | 0.1375 |
| pareto | 2.1 | 0.9429 | 0.0571 | 0.0690 | 0.1368 |
| nominal: two-sided 0.95, upper miss 0.025, 1-sided miss 0.05 |
Show the full program
# Chapter 1. p11: real coverage of the normal 95% interval on heavy tails.
# Model: a shop has n = 30 past values (daily orders or order values)
# and states "mean +/- t * sd * sqrt(1 + 1/n)" as a 95% prediction
# interval (t = 2.0452, the 0.975 point of Student t with 29 df, from
# standard tables); the one-sided 95% upper bound uses t = 1.6991.
# Coverage of the next value is exact from the true CDF, per training
# set; we average over 8,000 training sets.
# Distributions are illustrative shapes, not data from any shop.
import math
import random
import statistics
ND = statistics.NormalDist()
def make(kind, a):
if kind == "normal":
return (lambda r: r.gauss(0.0, 1.0)), (lambda x: ND.cdf(x))
if kind == "lognorm":
def cdf(x):
return 0.0 if x <= 0 else ND.cdf(math.log(x) / a)
return (lambda r: math.exp(r.gauss(0.0, a))), cdf
def cdf(x):
return 0.0 if x <= 1.0 else 1.0 - x ** (-a)
return (lambda r: (1.0 - r.random()) ** (-1.0 / a)), cdf
def normal_pi(xs):
m = statistics.fmean(xs)
s = statistics.stdev(xs)
f = s * math.sqrt(1.0 + 1.0 / len(xs))
return m - 2.0452 * f, m + 2.0452 * f, m + 1.6991 * f
def run(kind, a, n=30, reps=8000):
rng = random.Random(20261113)
draw, cdf = make(kind, a)
cov, up, one = [], [], []
for _ in range(reps):
lo, hi, ub = normal_pi([draw(rng) for _ in range(n)])
fl, fh = cdf(lo), cdf(hi)
cov.append(fh - fl)
up.append(1.0 - fh)
one.append(1.0 - cdf(ub))
bad = len([c for c in cov if c < 0.90])
return (math.fsum(cov) / reps, math.fsum(up) / reps,
math.fsum(one) / reps, bad / reps)
if __name__ == "__main__":
cases = [("normal", 0.0), ("lognorm", 0.5), ("lognorm", 1.0),
("lognorm", 1.5), ("pareto", 3.0), ("pareto", 2.5),
("pareto", 2.1)]
print("normal-theory 95% PI, n = 30, 8000 training sets")
print("shape param two-sided upper miss 1-sided miss P(<0.90)")
for kind, a in cases:
c, up, one, bad = run(kind, a)
print("{:<9s} {:4.1f} {:.4f} {:.4f} {:.4f} {:.4f}"
.format(kind, a, c, up, one, bad))
print("nominal: two-sided 0.95, upper miss 0.025, 1-sided miss 0.05")Single source and citation
Fan, Y. (2026). Conformal Prediction and Distribution-Free Intervals: Finite-Sample Coverage Guarantees for Order Counts and Order Values (Version 1.0.0), Chapter 1: The Coverage Gap of Normal Intervals: Heavy-Tailed Orders Push Every Miss to the Right. 5D Library. DOI pending
DOI pending · monograph page goes live after publication · 5DL-B-306 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026306,
author = {Fan, Yafu},
title = {Conformal Prediction and Distribution-Free Intervals: Finite-Sample Coverage Guarantees for Order Counts and Order Values},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}Other chapter answers in this book
- Chapter 2 · The Correct Rank: Why "Plus One" Matters and Where the Finite-Sample Guarantee Comes From
- Chapter 3 · The Price of the Guarantee: Interval Width, the Shape of the Score and the Extra Paid When Normality Holds
- Chapter 4 · How Large a Calibration Set: Coverage Given the Calibration Set Follows a Beta Law
- Chapter 5 · Marginal Is Not Conditional: Traffic Bands, Customer Segments and Intervals Too Wide to Use
- Chapter 6 · When Exchangeability Breaks: Weekend Effects, Growth Drift, Promotion Shocks and Three Repairs
- Chapter 7 · Classical Sources East and West: Guo Shoujing's "The Calendar Rests on Measurement" and Neyman's Confidence Intervals
- Chapter 8 · Closing the Loop: the Weekly Order-Count Interval Sheet, the Conformal Interval Calculator and a Pre-Registered Test