关于「四方对账:自动匹配率、容差与跨主体重复入账」,《多主体清分与审计就绪账本》给出的定量结论是什么?
结论:A 段显示,这 2,000 条里有 79 条真实错误。容差为 0.00 时,自动匹配率 0.8985,队列 203 条,没有任何真错误被放过。容差放到 0.01,匹配率 0.9200,队列 160 条。放到 0.03,匹配率 0.9605,队列 79 条,恰好只剩真错误,良性尾差全部清掉,没有一条真错误被吞掉。再放到 0.10,结果不变。放到 0.50,匹配率 0.9630,队列 74 条,却有 5 条真错误被吞掉;放到 1.00,匹配率 0.9680,队列 64 条,被吞掉的真错误升到 15 条。
数据口径:书中参数为示意,数据为合成;全部计算由书中程序生成。 第一章按 IFRS 15 的控制判断给七类合成交易打分,「到账即收入」有 3 类出错,误差有正有负,合计高估 0.2763。
对照数据表
下表数字由本章 Python 程序原样打印,可重新运行复现。 程序只用 Python 标准库与结论吻合的打印值: 0.00, 0.01, 0.03, 0.10, 0.50, 0.8985
| A. 2,000 lines, 79 genuine faults (seeded) | |||
| tolerance | auto-match | queue | faults swallowed |
| 0.00 | 0.8985 | 203 | 0 of 79 |
| 0.01 | 0.9200 | 160 | 0 of 79 |
| 0.03 | 0.9605 | 79 | 0 of 79 |
| 0.10 | 0.9605 | 79 | 0 of 79 |
| 0.50 | 0.9630 | 74 | 5 of 79 |
| 1.00 | 0.9680 | 64 | 15 of 79 |
| B. 0.03 clears every residue and swallows no fault here; | |||
| wider tolerances buy a shorter queue with hidden faults |
查看完整验算程序
# Chapter 5. p52: matching tolerance against swallowed errors. 2,000
# synthetic settlement lines; 0.90 agree exactly, 0.06 differ by a
# benign rounding or conversion residue of 0.01 to 0.03, and 0.04 carry
# a genuine fault of 0.20 to 5.00 (fee off contract, partial refund).
# Rates are illustrative. A line auto-matches when |difference| is
# within the tolerance; a genuine fault that auto-matches is swallowed.
import random
from decimal import Decimal
rng = random.Random(40252)
lines = []
for _ in range(2000):
u = rng.random()
if u < 0.90:
lines.append(("exact", Decimal(0)))
elif u < 0.96:
lines.append(("residue", Decimal(rng.randint(1, 3)) / 100 *
(1 if rng.random() < 0.5 else -1)))
else:
lines.append(("fault", Decimal(rng.randint(20, 500)) / 100 *
(1 if rng.random() < 0.5 else -1)))
if __name__ == "__main__":
n_fault = 0
for k, _ in lines:
n_fault += 1 if k == "fault" else 0
print("A. 2,000 lines, {} genuine faults (seeded)".format(n_fault))
print(" tolerance auto-match queue faults swallowed")
for tol in ("0.00", "0.01", "0.03", "0.10", "0.50", "1.00"):
t = Decimal(tol)
auto = swallowed = 0
for k, d in lines:
if abs(d) <= t:
auto += 1
swallowed += 1 if k == "fault" else 0
print(" {:>9} {:>8.4f} {:>5} {:>4} of {}".format(
tol, auto / len(lines), len(lines) - auto, swallowed,
n_fault))
print("B. 0.03 clears every residue and swallows no fault here;")
print(" wider tolerances buy a shorter queue with hidden faults")唯一数据源与引用
范雅福(Fan, Y.). (2026). 《跨国数字业务多主体清分、跨境税基合规边界与一键审计就绪账本母典》第 5 章「四方对账:自动匹配率、容差与跨主体重复入账」(版本 1.0.0). 5D Library. DOI 待发布
DOI 待发布 · 专著页正式发布后上线(本站暂无专著页) · 5DL-B-402 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026402,
author = {Fan, Yafu},
title = {Multi-Jurisdiction Digital Revenue Clearing, Cross-Border Tax-Base Compliance Boundaries and Audit-Ready Automated Ledgers},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}