What does Collapsing Idea-to-Delivery Latency conclude, in numbers, about “Toolchain Friction and Gates: Why Speed Never Skips a Gate”?
Answer: Chapter 6 shows that gates should be sped up, never skipped: at a fault rate of 0.05 and an 8-hour rollback, expected hours are 0.220 with the gate and 0.400 without, and ordering checks by hit rate over run time brings the first red light from 12.11 to 3.73 minutes.
Scope of the numbers: Parameters are illustrative and all calculations are generated by the book's programs; it offers general, not individual, business guidance, promises no launch date or income, and every production deployment is carried out only after a person confirms it. Chapter 2 splits latency into four drags, decision, start-up, feedback and hand-off, with an illustrative median of 99.9 hours and P90 of 199.8 hours, the tail driven mainly by feedback and decision.
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: 0.05, 0.220, 0.400
| A. expected hours per change (g = 0.1 h, catch 0.7) | ||||
| p | R | gate | no gate | break-even g |
| 0.02 | 2 | 0.112 | 0.040 | 0.028 |
| 0.02 | 8 | 0.148 | 0.160 | 0.112 |
| 0.02 | 24 | 0.244 | 0.480 | 0.336 |
| 0.05 | 2 | 0.130 | 0.100 | 0.070 |
| 0.05 | 8 | 0.220 | 0.400 | 0.280 |
| 0.05 | 24 | 0.460 | 1.200 | 0.840 |
| 0.10 | 2 | 0.160 | 0.200 | 0.140 |
| 0.10 | 8 | 0.340 | 0.800 | 0.560 |
| 0.10 | 24 | 0.820 | 2.400 | 1.680 |
| B. lead time per change, p 0.05, R 8 h | ||||
| mode | P50 | P90 | P99 | mean |
| gate | 1.14 | 2.16 | 8.95 | 1.39 |
| no gate | 1.04 | 2.22 | 9.59 | 1.54 |
Show the full program
# Chapter 6. p62: is a gate worth its minutes? Expected cost and the tail.
# A change carries a fault with probability p. An automated gate takes
# g hours and stops a fraction c of faults before release; a fault that
# escapes is found later and costs R hours of rollback and repair. Per
# change: with gate g + p (1 - c) R, without p R; the gate pays when
# g < p c R. Part B simulates 20,000 changes (fixed seed) with build
# time lognormal (median 1 h, log-sd 0.5): lead time with the gate is
# build + g (+ a fix of 1 h when caught, + R when it escapes); without
# the gate, build (+ R when a fault ships). All values illustrative.
import math
import random
import statistics
G = 0.1
CATCH = 0.7
N = 20000
def per_change(p, r, gate):
return G + p * (1.0 - CATCH) * r if gate else p * r
def q(xs, f):
s = sorted(xs)
return s[min(len(s) - 1, int(f * len(s)))]
if __name__ == "__main__":
print("A. expected hours per change (g = 0.1 h, catch 0.7)")
print(" p R gate no gate break-even g")
for p in (0.02, 0.05, 0.10):
for r in (2.0, 8.0, 24.0):
print(" {:.2f} {:>4.0f} {:.3f} {:.3f} {:.3f}".format(
p, r, per_change(p, r, True), per_change(p, r, False),
p * CATCH * r))
rng = random.Random(62001)
with_g, without = [], []
p, r = 0.05, 8.0
for _ in range(N):
b = rng.lognormvariate(0.0, 0.5)
faulty = rng.random() < p
caught = rng.random() < CATCH
lw = b + G
if faulty:
lw += 1.0 if caught else r
with_g.append(lw)
without.append(b + (r if faulty else 0.0))
print("B. lead time per change, p 0.05, R 8 h")
print(" mode P50 P90 P99 mean")
for name, xs in (("gate", with_g), ("no gate", without)):
print(" {:<8} {:>5.2f} {:>5.2f} {:>5.2f} {:>5.2f}".format(
name, q(xs, 0.5), q(xs, 0.9), q(xs, 0.99),
statistics.fmean(xs)))Single source and citation
Fan, Y. (2026). Collapsing Idea-to-Reality Latency: Feedback-Bandwidth Limits, Zero-Friction Toolchains and High-Velocity Execution Physics (Version 1.0.0), Chapter 6: Toolchain Friction and Gates: Why Speed Never Skips a Gate. 5D Library. DOI pending
DOI pending · monograph page goes live after publication · 5DL-B-420 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026420,
author = {Fan, Yafu},
title = {Collapsing Idea-to-Reality Latency: Feedback-Bandwidth Limits, Zero-Friction Toolchains and High-Velocity Execution Physics},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}Other chapter answers in this book
- Chapter 1 · Loop Bandwidth: With the Delay Unchanged, More Gain Only Buys Oscillation
- Chapter 2 · Decomposing End-to-End Latency: Four Drags and the Slowest Tenth
- Chapter 3 · Queues: Why Full Utilisation Is Not Efficiency
- Chapter 4 · Batches and Rehearsal: the U-Shaped Cost and the Smith Predictor
- Chapter 5 · Hand-offs, Switching and Hesitation: Three Losses Along the Chain
- Chapter 7 · Classical Sources East and West: Clumsy Speed, the Ant Hole and Maxwell on Governors
- Chapter 8 · Closing the Loop: the Latency Ledger, Field Tools and Checker Specifications