What does Human-AI Mutual Feeding and the Falsification Ratchet conclude, in numbers, about “Multi-Agent Falsification: Majority Error and the Effective Number of Votes under Correlation”?
Answer: Chapter 6 computes multi-agent falsification: correlated same-source agents add votes without adding protection, and the independent prediction of 0.0579 for five votes holds only under independence.
Scope of the numbers: Parameters are illustrative and all calculations are generated by the book's programs. Chapter 2 treats the jagged frontier: with illustrative parameters, clicking confirm outside the frontier is right 0.604 of the time against 0.800 for answering alone, while falsifying first and then comparing leads on both sides.
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: 0.0579
| Panel error, independent agents | |
| e k | 1 k=3 k=5 k=7 k=9 |
| 0.05 | 0.0500 0.0073 0.0012 0.0002 0.0000 |
| 0.10 | 0.1000 0.0280 0.0086 0.0027 0.0009 |
| 0.20 | 0.2000 0.1040 0.0579 0.0333 0.0196 |
| 0.30 | 0.3000 0.2160 0.1631 0.1260 0.0988 |
| 0.40 | 0.4000 0.3520 0.3174 0.2898 0.2666 |
| Smallest odd k with panel error < 0.01 | |
| e | 0.10: k = 5 |
| e | 0.20: k = 13 |
| e | 0.30: k = 31 |
| e | 0.40: k = 133 |
Show the full program
# Chapter 6. p61: majority verdicts of k independent falsifying agents.
# Each agent errs with probability e, independently (Condorcet case).
# The panel errs when more than half err:
# P_k = sum_{j > k/2} C(k, j) e^j (1 - e)^(k - j) (k odd)
# Computed exactly with fractions, then shown as decimals. Also the
# smallest odd k bringing the panel error below 0.01. A single agent
# reviewing its own output is k = 1 at best. Illustrative error rates.
import math
from fractions import Fraction
def panel(k, e):
e = Fraction(e)
terms = [math.comb(k, j) * e ** j * (1 - e) ** (k - j)
for j in range(k // 2 + 1, k + 1)]
acc = Fraction(0)
for t in terms:
acc += t
return acc
if __name__ == "__main__":
print("Panel error, independent agents")
print(" e k=1 k=3 k=5 k=7 k=9")
for e in ("0.05", "0.10", "0.20", "0.30", "0.40"):
cells = [float(panel(k, Fraction(e))) for k in (1, 3, 5, 7, 9)]
print(" {} {:.4f} {:.4f} {:.4f} {:.4f} {:.4f}".format(
e, *cells))
print("Smallest odd k with panel error < 0.01")
for e in ("0.10", "0.20", "0.30", "0.40"):
k = 1
while float(panel(k, Fraction(e))) >= 0.01:
k += 2
print(" e = {}: k = {}".format(e, k))Single source and citation
Fan, Y. (2026). Co-Evolutionary Human-AI Writing Spirals, Falsification-Gated Ratchets and Symbiotic Capability Compounding (Version 1.0.0), Chapter 6: Multi-Agent Falsification: Majority Error and the Effective Number of Votes under Correlation. 5D Library. DOI pending
DOI pending · monograph page goes live after publication · 5DL-B-374 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026374,
author = {Fan, Yafu},
title = {Co-Evolutionary Human-AI Writing Spirals, Falsification-Gated Ratchets and Symbiotic Capability Compounding},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}Other chapter answers in this book
- Chapter 1 · Variance Collapse under Self-Feeding: Why the Tails Go First
- Chapter 2 · The Jagged Frontier: Why Clicking Confirm Loses Accuracy, and Why Better Individuals Make a Narrower Crowd
- Chapter 3 · The Gain from Injected Outliers: How Worse-on-Average Candidates Raise the Best
- Chapter 4 · The Falsification Gate: Sequential Boundaries, Sample Sizes and the Leak Rate
- Chapter 5 · A Robust Ratchet: Random Walks with Misclassification, Rollback Audits and the Condition for Compounding
- Chapter 7 · Classical Sources: Teaching and Learning in the Xueji, Blue from Indigo in the Xunzi, Bacon's Idols and Popper's Conjectures
- Chapter 8 · Field Tools and Checker Specifications: the Daily Ratchet Ledger, Red-Yellow-Green Criteria and a Pre-registered Refutation