关于「东西方原典考辨:计然「旱则资舟,水则资车」与汤普森 1933 年的两样本问题」,《业务线周度配额》给出的定量结论是什么?

结论:第七章考辨《史记·货殖列传》所载计然「旱则资舟,水则资车」与汤普森 1933 年的两样本问题:逆着当下准备只在规律周期且时滞超过四分之一周期时有利。

数据口径:核对吉廷斯 1979 年的论文与后续文献以后,本书修正了立项蓝图:吉廷斯指数的最优性只在单线经营、闲置线不变、换线无成本、几何贴现时成立,周度拆分预算、换线成本与会变的市场都在定理之外;锁死是一个概率说法,两条示意线上按累计回报全押有 0.2133 的重复在下半年完全押错,但在每单位贴现 0.9 的短视界里贪心相对指数只损失 0.0050 到 0.0166。 书中参数为示意,毛利与成本用记账点数而非价格;全部计算由书中程序生成。

English version

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

P(same regime after lag L); Jiran rule pays 1 - P
LcyclejittermarkovJiran wins in
01.00001.00001.0000-
10.83330.83340.8333-
20.66670.66690.7222-
30.50000.50030.6481-
40.33330.33370.5988cycle,jitter
50.16670.23410.5658cycle,jitter
60.00000.20120.5439cycle,jitter
70.16670.23310.5293cycle,jitter
80.33330.33250.5195cycle,jitter
查看完整验算程序
# Chapter 7. p71: when does "in drought, lay in boats" pay?
# Model: two regimes, dry and wet, each lasting 6 periods on average.
# Stock prepared now is used L periods later and pays 1 if it suits the
# regime then, 0 if not (illustrative). The follower prepares for the
# regime of today; the Jiran rule prepares for the opposite one.
# Payoff of the follower = P(regime at t + L equals regime at t).
#   cycle  : a fixed 12-period cycle, 6 dry then 6 wet
#   jitter : spells of 4 to 8 periods, uniform (mean 6), fixed seed
#   markov : stay with probability 5/6 each period (mean spell 6),
#            P_same = (1 + (2/3)^L) / 2 exactly
# The Jiran rule wins exactly when P_same < 0.5.
import random


def same_share(path, lag):
    hit = 0
    for t in range(len(path) - lag):
        if path[t] == path[t + lag]:
            hit += 1
    return hit / (len(path) - lag)


rng = random.Random(20261113)
jit, state = [], 0
while len(jit) < 200000:
    jit.extend([state] * rng.randint(4, 8))
    state = 1 - state
cyc = [(t // 6) % 2 for t in range(200000)]
print("P(same regime after lag L); Jiran rule pays 1 - P")
print("%3s %8s %8s %8s %16s" % ("L", "cycle", "jitter", "markov",
                                "Jiran wins in"))
for lag in range(0, 9):
    c = same_share(cyc, lag)
    j = same_share(jit, lag)
    m = (1 + (2 / 3) ** lag) / 2
    wins = [name for name, v in (("cycle", c), ("jitter", j),
                                 ("markov", m)) if v < 0.5]
    print("%3d %8.4f %8.4f %8.4f %16s" % (lag, c, j, m,
                                          ",".join(wins) or "-"))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《多业务线算力与预算动态裁决、吉廷斯指数探索与收费站代购出版线周度配额母典》第 7 章「东西方原典考辨:计然「旱则资舟,水则资车」与汤普森 1933 年的两样本问题」(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-264 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026264,
  author    = {Fan, Yafu},
  title     = {Dynamic Allocation of Compute and Budget across Business Lines: Gittins-Index Exploration and Weekly Quotas for Tollbooth, Buying and Publishing Lines},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论