关于「决策疲劳日程:重排不等式与休息的动态规划」,《自我损耗与饥饿法官》给出的定量结论是什么?
结论:第六章用重排不等式与动态规划计算排序与休息规则:在小效应世界,「大事先做」只节省总金额的 0.0100。
数据口径:书中参数为示意;全部计算由书中程序生成;本书不是心理或医疗建议,书中流程与工具为非个人化经营建议。 本书把两项证据放到预注册复制、元分析偏差检验与案件排序模拟面前,看它们还剩多少。
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: 0.0100
| orders enumerated: 40320, total stakes 130 | |||||
| delta | descending | random mean | ascending | best=desc | saving |
| 0.000 | 13.000 | 13.000 | 13.000 | yes | 0.000 |
| 0.002 | 13.390 | 13.910 | 14.430 | yes | 0.520 |
| 0.005 | 13.975 | 15.275 | 16.575 | yes | 1.300 |
| 0.020 | 16.900 | 22.100 | 27.300 | yes | 5.200 |
| at delta 0.005 the rule saves 0.0100 of total stakes |
查看完整验算程序
# Chapter 6. p61: Highest stakes first - exact check over all orders.
# Eight decisions with stakes 50, 30, 20, 12, 8, 5, 3, 2 (units of money,
# illustrative). The chance of a costly error at position t is
# e0 + delta * (t - 1), e0 = 0.10. Expected loss = sum of stake * error
# chance. All 40320 orders are enumerated for each delta; by the
# rearrangement inequality descending stakes should be best. Delta 0.005
# sits near the pooled replication; 0.02 is a strong-depletion world.
# 高风险先决:枚举全部排列,验证重排不等式。
import math
from itertools import permutations
STAKES = (50, 30, 20, 12, 8, 5, 3, 2)
E0 = 0.10
def loss(order, delta):
return math.fsum(s * (E0 + delta * t) for t, s in enumerate(order))
total = math.fsum(STAKES)
print("orders enumerated: {}, total stakes {:.0f}".format(
math.factorial(len(STAKES)), total))
print("delta descending random mean ascending best=desc saving")
for delta in (0.0, 0.002, 0.005, 0.02):
vals = [loss(p, delta) for p in permutations(STAKES)]
desc = loss(sorted(STAKES, reverse=True), delta)
asc = loss(sorted(STAKES), delta)
mean = math.fsum(vals) / len(vals)
best = min(vals)
print("{:.3f} {:9.3f} {:10.3f} {:9.3f} {:8s} {:6.3f}".format(
delta, desc, mean, asc, "yes" if abs(best - desc) < 1e-9
else "no", mean - desc))
d = 0.005
gain = (math.fsum(loss(p, d) for p in permutations(STAKES)) /
math.factorial(len(STAKES)) - loss(sorted(STAKES)[::-1], d))
print("at delta {:.3f} the rule saves {:.4f} of total stakes".format(
d, gain / total))唯一数据源与引用
范雅福(Fan, Y.). (2026). 《自我损耗大规模复制失败、法官饥饿效应排序伪影与决策疲劳日程母典》第 6 章「决策疲劳日程:重排不等式与休息的动态规划」(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-196 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026196,
author = {Fan, Yafu},
title = {The Failed Replication of Ego Depletion, the Ordering Artefact behind the Hungry Judges and Decision-Fatigue Scheduling},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}