What does Hard Gates and Soft-Prompt Decay conclude, in numbers, about “Where to Intercept: before the Write, after It or before Publication, and Whether a Broken Gate Passes or Refuses”?
Answer: Chapter 4 compares interception before the write, after it and before publication; the combined design E costs only 0.60 at a hook failure rate of 0.010.
Scope of the numbers: Soft reminders work in proportion; when layers are independent the residual violation rate is the product of their miss rates, and with illustrative parameters four layers still leave 0.0048, not zero. 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. printed values that match the answer: 0.010, 0.60
| design | e | escapes | blocks | cost |
|---|---|---|---|---|
| save hook fail-open | 0.001 | 0.0096 | 0.00 | 1.15 |
| save hook fail-closed | 0.001 | 0.0000 | 2.00 | 6.00 |
| E open save + closed | 0.001 | 0.0000 | 0.02 | 0.06 |
| save hook fail-open | 0.010 | 0.0960 | 0.00 | 11.52 |
| save hook fail-closed | 0.010 | 0.0000 | 20.00 | 60.00 |
| E open save + closed | 0.010 | 0.0000 | 0.20 | 0.60 |
| in E an escape needs the save hook down and the publish gate | ||||
| down at once; a closed publish gate turns that into a block |
Show the full program
# Chapter 4. p42: when the gate itself breaks, open or closed?
# A save hook crashes with probability e per call. Fail-open lets the
# save through unchecked; fail-closed blocks it. Per week: N = 2000
# saves, a share r = 0.0048 of them carry an in-class violation after
# the soft layers (p11), M = 20 publish-gate runs. Costs (minutes,
# illustrative): a violation that reaches readers 120, a wrong block
# of a clean save 3. Design E pairs a fail-open save hook with a
# fail-closed publish gate: availability where work is done, safety
# where work leaves.
N, R, M = 2000, 0.0048, 20
C_ESC, C_BLK = 120.0, 3.0
print("%-22s %7s %10s %10s %9s" % ("design", "e", "escapes", "blocks",
"cost"))
for e in (0.001, 0.01):
designs = [
("save hook fail-open", N * R * e, 0.0),
("save hook fail-closed", 0.0, N * e),
("E open save + closed", 0.0, M * e),
]
for name, esc, blk in designs:
cost = esc * C_ESC + blk * C_BLK
print("%-22s %7.3f %10.4f %10.2f %9.2f" % (name, e, esc, blk,
cost))
print("in E an escape needs the save hook down and the publish gate")
print("down at once; a closed publish gate turns that into a block")Single source and citation
Fan, Y. (2026). Hard Gates across the Client Lifecycle: the Decay of Soft-Prompt Compliance over Turns and State Machines for Interception before Publication (Version 1.0.0), Chapter 4: Where to Intercept: before the Write, after It or before Publication, and Whether a Broken Gate Passes or Refuses. 5D Library. DOI pending
DOI pending · monograph page goes live after publication · 5DL-B-256 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026256,
author = {Fan, Yafu},
title = {Hard Gates across the Client Lifecycle: the Decay of Soft-Prompt Compliance over Turns and State Machines for Interception before Publication},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}Other chapter answers in this book
- Chapter 1 · The Product Account: Why Stacked Soft Layers Never Leave a Zero Residual
- Chapter 2 · Holes That Line Up: the Floor That Common-Cause Failure Puts under Stacked Soft Layers
- Chapter 3 · Decay over Turns: an Exponential Model of Soft-Prompt Compliance, Re-Sending Periods and Reading the Log
- Chapter 5 · Exhausting the State Machine: Proving a Dirty Publication Unreachable and Finding the Shortest Counterexamples of Three Faulty Designs
- Chapter 6 · The Price of Hardness: Pattern Width, Wrong Blocks and Gates That Get Switched Off
- Chapter 7 · Classical Sources East and West: Law and the Inked Line in the Han Feizi, and Reason's Classification of Error
- Chapter 8 · Closing the Loop: the Weekly Interception Report, the Residual Calculator and a Pre-Registered Test