What does One Commander and the Serial Bottleneck conclude, in numbers, about “Decision Bandwidth: a Written Constitution as a Cache, a Faster Serial Part and the Span of Control”?
Answer: Chapter 6 treats a written constitution as a decision cache; at a nine-in-ten hit rate ten agents reach a speedup of 8.696.
Scope of the numbers: With illustrative parameters, cutting the serial share from a fifth to a twentieth raises the speedup by 3.325, while raising the agents from ten to a hundred adds only 1.236. Its parameters are illustrative; all calculations are generated by its programs.
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: 8.696
| h r_h s_h ceiling S(N | 10) |
| 0.00 | 12.00 0.2000 5.00 3.571 |
| 0.25 | 9.25 0.1616 6.49 4.270 |
| 0.50 | 6.50 0.1193 9.23 5.310 |
| 0.75 | 3.75 0.0725 16.00 7.018 |
| 0.90 | 2.10 0.0419 28.57 8.696 |
| 0.95 | 1.55 0.0313 38.71 9.449 |
| levers at N | 10, from S = 3.571: |
| N 10 | 20 S = 4.167 |
| N 10 | 100 S = 4.808 |
| decide 2 x faster S | 5.556 |
| decide 4 x faster S | 7.692 |
| cache h | 0.50 S = 5.310 |
| cache h | 0.90 S = 8.696 |
Show the full program
# Chapter 6. p61: A written constitution as a decision cache.
# Model: each item needs a = 48 min of agent work and one decision
# by the commander that takes r = 12 min when made from scratch, so
# one person alone needs 60 min (s = 0.20). A share h of decisions is
# already answered by a written rule (a constitution, a price table,
# a style sheet): looking it up costs 1 min. Effective review
# r_h = h x 1 + (1 - h) x 12. All speedups are measured against the
# original 60 minutes: S(N) = 60 / (r_h + 48 / N), ceiling 60 / r_h.
# Second lever (Hill and Marty: make the sequential part faster): the
# commander decides k times faster, r = 12 / k. Illustrative.
from fractions import Fraction as F
A, R, LOOKUP = 48, 12, 1
BASE = A + R
def speed(r, n):
return BASE / (r + F(A, n))
print("%6s %7s %8s %9s %9s" % ("h", "r_h", "s_h", "ceiling",
"S(N=10)"))
for h in [F(0), F(1, 4), F(1, 2), F(3, 4), F(9, 10), F(19, 20)]:
rh = h * LOOKUP + (1 - h) * R
s = rh / (A + rh)
print("%6.2f %7.2f %8.4f %9.2f %9.3f" % (
float(h), float(rh), float(s), float(BASE / rh),
float(speed(rh, 10))))
print("levers at N = 10, from S = %.3f:" % float(speed(F(R), 10)))
print(" N 10 -> 20 S = %.3f" % float(speed(F(R), 20)))
print(" N 10 -> 100 S = %.3f" % float(speed(F(R), 100)))
for k in [2, 4]:
print(" decide %d x faster S = %.3f" % (
k, float(speed(F(R, k), 10))))
for h in [F(1, 2), F(9, 10)]:
rh = h * LOOKUP + (1 - h) * R
print(" cache h = %.2f S = %.3f" % (
float(h), float(speed(rh, 10))))Single source and citation
Fan, Y. (2026). Constitutional Engineering for a One-Commander Silicon Empire: Serial Decision Bandwidth and Breaking Amdahl's Bottleneck (Version 1.0.0), Chapter 6: Decision Bandwidth: a Written Constitution as a Cache, a Faster Serial Part and the Span of Control. 5D Library. DOI pending
DOI pending · monograph page goes live after publication · 5DL-B-253 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026253,
author = {Fan, Yafu},
title = {Constitutional Engineering for a One-Commander Silicon Empire: Serial Decision Bandwidth and Breaking Amdahl's Bottleneck},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}Other chapter answers in this book
- Chapter 1 · The Ceiling of the Serial Share: Amdahl's Law and the Exact Account of Twenty Percent against Five
- Chapter 2 · Does the Problem Grow? Gustafson's Law, Per-Item Review and the Measured Serial Share
- Chapter 3 · The Commander Is a Single Server: Queueing Theory, the Finite-Source Model and Little's Law
- Chapter 4 · Batch Approval Windows: Harris's Square-Root Formula and the Price of Waiting
- Chapter 5 · Send Up Only the Exceptions: Automatic Gates, Random Audits and Coordination Overhead
- Chapter 7 · Eastern and Western Sources Examined: Han Fei's Tally of Name and Form and Smith's Pin Factory
- Chapter 8 · Closing the Loop: a Weekly Decision Log, Batch Approval Windows and a Pre-Registered Test