What does Red-Blue Synthetic-Customer Sandboxes conclude, in numbers, about “Four Kinds of Leak: Factual Errors, Invented Scarcity, Undue Pressure and Dodged Questions”?

Answer: Program p61 uses one fact sheet and four rules to stop four of five illustrative scripts, and shows that keyword rules can be fooled by an invented countdown. Program p62 shows that a script answering only price passes the average-buyer test while dodging seven tenths of the objections. Program p63 shows that the script with a resetting countdown converts best and is still stopped at the gate. The idea behind these checks is not new: a military classic of more than two thousand years ago and a war game of two centuries ago were both concerned with seeing oneself and the other side clearly before battle. The next chapter returns to those sources.

Scope of the numbers: Parameters are illustrative; all calculations are generated by the book's programs, which call no model interface. This is not individual business advice.

中文版

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: 62

scriptbuyleaveconvlow95leaveverdict
A honest621180.25830.20710.4917PASS
B fake timer751010.31250.25720.4208BLOCK red line
C thin521300.21670.16920.5417HOLD unproven
D cold501600.20830.16170.6667FAIL leave
buyers needed for low95 > 0.20 at true rate 0.25: 230
buyers needed for low95 > 0.20 at true rate 0.30: 70
Show the full program
# Chapter 6. p63: a release gate in which red lines come first.
# The arena sandbox blocks a script on any red-line hit regardless of
# conversion, then asks for conversion >= 0.20 and leaving <= 0.60
# (its documented defaults). Here the conversion test uses the Wilson
# 95 per cent lower bound, so a lucky small run cannot pass. Counts out
# of 240 synthetic buyers are illustrative.
import math

Z = 1.959964


def wilson(k, n):
    p = k / n
    c = (p + Z * Z / (2 * n)) / (1 + Z * Z / n)
    h = Z * math.sqrt(p * (1 - p) / n + Z * Z / (4 * n * n))
    return c - h / (1 + Z * Z / n), c + h / (1 + Z * Z / n)


RUNS = (("A honest", 62, 118, 0), ("B fake timer", 75, 101, 1),
        ("C thin", 52, 130, 0), ("D cold", 50, 160, 0))
print("script         buy  leave  conv    low95   leave  verdict")
for name, buy, leave, red in RUNS:
    n = 240
    lo, _ = wilson(buy, n)
    if red:
        v = "BLOCK red line"
    elif leave / n > 0.60:
        v = "FAIL leave"
    elif lo < 0.20:
        v = "HOLD unproven"
    else:
        v = "PASS"
    print("%-13s  %3d  %5d  %.4f  %.4f  %.4f  %s"
          % (name, buy, leave, buy / n, lo, leave / n, v))
for true in (0.25, 0.30):
    n = 10
    while wilson(round(true * n), n)[0] < 0.20:
        n += 10
    print("buyers needed for low95 > 0.20 at true rate %.2f: %d"
          % (true, n))

Single source and citation

Fan, Y. (2026). Red-Blue Synthetic-Customer Sandboxes: Multi-Round Bargaining against Parameterised Demanding Buyers and Pre-Launch Leak Detection (Version 1.0.0), Chapter 6: Four Kinds of Leak: Factual Errors, Invented Scarcity, Undue Pressure and Dodged Questions. 5D Library. DOI pending

DOI pending · monograph page goes live after publication · 5DL-B-259 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026259,
  author    = {Fan, Yafu},
  title     = {Red-Blue Synthetic-Customer Sandboxes: Multi-Round Bargaining against Parameterised Demanding Buyers and Pre-Launch Leak Detection},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

Other chapter answers in this book