What does The Fewest Questions conclude, in numbers, about “Classical Sources East and West: "Knocking at Both Ends" and Fisher's Amount of Information”?
Answer: Chapter 7 examines "knocking at both ends" in the Analects (6.417 questions on average over twelve levels against 3.667 for bisection) and Fisher's 1925 "Theory of Statistical Estimation" (median efficiency 0.6366).
Scope of the numbers: Marked boxes separate published evidence, runnable Python calculations (19 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: 3.667, 6.417
levels M inward mean inward max bisect mean bisect max ends+bis 12 6.417 11 3.667 4 5.667 24 12.458 23 4.667 5 6.667 64 32.484 63 6.000 6 8.000 question 0 log units from the truth: info 1.5625, share 1.0000 question 1 log units from the truth: info 0.4381, share 0.2804 question 2 log units from the truth: info 0.0416, share 0.0266 question 3 log units from the truth: info 0.0035, share 0.0022
Show the full program
# Chapter 7. p71: "knocking at both ends" read three ways. Exact counts
# on M ordered levels, noiseless answers, plus the 2PL information of
# an end question. All parameters illustrative.
# 读法一:从两端向内逐级叩问(问最低、问最高、再向内);
# 读法二:一开始就二分;读法三:先叩两端确认范围(2 题),再二分。
# 真值在 M 级上均匀;另算若真值可能落在范围外,读法二无法察觉。
import math
def inward(m, v):
lo, hi, k = 0, m - 1, 0
while lo < hi:
k += 1
if k % 2 == 1:
if v == lo:
return k
lo += 1
else:
if v == hi:
return k
hi -= 1
return k
def bisect_count(m, v):
lo, hi, k = 0, m - 1, 0
while lo < hi:
mid = (lo + hi) // 2
k += 1
if v <= mid:
hi = mid
else:
lo = mid + 1
return k
print("levels M inward mean inward max bisect mean bisect max ends+bis")
for m in (12, 24, 64):
a = [inward(m, v) for v in range(m)]
b = [bisect_count(m, v) for v in range(m)]
print("%-9d %-12.3f %-11d %-12.3f %-11d %.3f"
% (m, math.fsum(a) / m, max(a), math.fsum(b) / m, max(b),
2 + math.fsum(b) / m))
A = 2.5
for d in (0.0, 1.0, 2.0, 3.0):
p = 1.0 / (1.0 + math.exp(-A * d))
print("question %.0f log units from the truth: info %.4f, share %.4f"
% (d, A * A * p * (1 - p), p * (1 - p) / 0.25))Single source and citation
Fan, Y. (2026). Fisher Information Matrices and the Cramér-Rao Bound: the Fewest Questions Needed to Estimate a Budget or a Birth Time (Version 1.0.0), Chapter 7: Classical Sources East and West: "Knocking at Both Ends" and Fisher's Amount of Information. 5D Library. DOI pending
DOI pending · monograph page goes live after publication · 5DL-B-235 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026235,
author = {Fan, Yafu},
title = {Fisher Information Matrices and the Cramér-Rao Bound: the Fewest Questions Needed to Estimate a Budget or a Birth Time},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}Other chapter answers in this book
- Chapter 1 · Fisher Information and the Cramér-Rao Bound: How Precise Can a Questionnaire Be?
- Chapter 2 · The Information Ceiling of Correlated Questions: Why the Sixth Question Is Nearly Wasted
- Chapter 3 · An Item Response Model for Budget Questions: How Tightly Do Five Ladder Answers Pin Down a Budget?
- Chapter 4 · Adaptive Selection: the Maximum-Information Rule and an Information-Theoretic Floor on the Number of Questions
- Chapter 5 · The Fisher Information Matrix: Yes-Saying, Reverse-Keyed Questions and Optimal Design
- Chapter 6 · Estimation Theory for a Birth Time: Rounding Error, Sequential Tests and "It Matched Afterwards"
- Chapter 8 · Closing the Loop: a Five-Question Pre-Sale Budget Questionnaire, a Read-Only Question Order and a Pre-Registered Test