关于「原典考辨:《荀子·解蔽》、《老子》第七十一章与笛卡尔的四条规则」,《解释深度剃刀与校准》给出的定量结论是什么?
结论:整体复购率是 0.3250。只读老客的人,样本 20 时误差 0.2494;样本增到 1,280,误差只降到 0.2244,几乎不动。分层抽样的人,样本 20 时误差 0.0984,到 1,280 降到 0.0124。最后一行给出一隅读者的偏差下限:0.2250,正是老客复购率与整体复购率之差。
数据口径:书中参数为示意,数据全部合成,全部计算由书中程序生成。 第一章在合成数据里量出装置类自评下降 1.283 分,事实类下降区间从 -0.263 到 0.063,跨过零,并给出剃刀与检索的分流规则。
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: 0.0124, 0.0984, 0.2244, 0.2250, 0.2494, 0.3250
| true overall repeat rate: 0.3250 | ||
| sample n | one-corner RMSE | stratified RMSE |
| 20 | 0.2494 | 0.0984 |
| 80 | 0.2310 | 0.0499 |
| 320 | 0.2261 | 0.0253 |
| 1,280 | 0.2244 | 0.0124 |
| bias floor of the one-corner reader: 0.2250 |
查看完整验算程序
# Chapter 7. p71: "blinded by one corner" as a bias that more data keeps.
# A buying service sells to five buyer groups with different repeat
# rates. A decision maker who only reads messages from the group she
# already serves samples one corner; a stratified reader samples every
# group in proportion. We compare the root mean square error of the
# estimated overall repeat rate over 1,000 repetitions as the sample
# grows. One-corner error stops falling at its bias. Illustrative rates.
import math
import random
rng = random.Random(39671)
GROUPS = [("loyal", 0.25, 0.55), ("gift", 0.20, 0.30), ("trial", 0.25, 0.15),
("bulk", 0.15, 0.40), ("abroad", 0.15, 0.20)]
TRUE = math.fsum(w * r for _, w, r in GROUPS)
REPS = 1000
def one_corner(n):
r = GROUPS[0][2]
return len([1 for _ in range(n) if rng.random() < r]) / n
def stratified(n):
parts = []
for _, w, r in GROUPS:
k = max(1, round(w * n))
parts.append(w * len([1 for _ in range(k) if rng.random() < r]) / k)
return math.fsum(parts)
def rmse(fn, n):
return math.sqrt(math.fsum((fn(n) - TRUE) ** 2
for _ in range(REPS)) / REPS)
print("true overall repeat rate: {:.4f}".format(TRUE))
print("sample n one-corner RMSE stratified RMSE")
for n in (20, 80, 320, 1280):
print(" {:5,} {:.4f} {:.4f}".format(
n, rmse(one_corner, n), rmse(stratified, n)))
print("bias floor of the one-corner reader: {:.4f}".format(
GROUPS[0][2] - TRUE))唯一数据源与引用
范雅福(Fan, Y.). (2026). 《解释深度错觉破除、认知边界校准与真知假知剥离剃刀母典》第 7 章「原典考辨:《荀子·解蔽》、《老子》第七十一章与笛卡尔的四条规则」(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-396 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026396,
author = {Fan, Yafu},
title = {Shattering the Illusion of Explanatory Depth, Epistemic Calibration and the Razor Separating True Knowledge from Assumed Familiarity},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}