关于「东西方原典考辨:《棋经十三篇》与富兰克林《下棋的道德》」,《围棋 AI 后的人类决策质量》给出的定量结论是什么?
结论:第七章考辨张拟《棋经十三篇》与富兰克林《下棋的道德》,算出「宁输数子,勿失一先」的盈亏平衡点从开局的 7.00 颗子降到终局的 0.79 颗。
数据口径:书中除论文数字外的参数均为示意,全部计算由书中程序生成;本书是非个人化经营建议,不针对任何个人决定。 第四章用香农熵度量定式集中度,示意角部局面的策略熵从 1.608 升到 2.360 比特,从众学习模型显示外部评估者让群体提早约 12 年离开局部最优。
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: 0.79, 7.00
| part 1: points on the board 361 | 360 + 1 |
| part 2: stones worth one tempo (T / 2) | |
| move T (points) break-even stones | |
| 0 | 14.00 7.00 |
| 50 | 7.96 3.98 |
| 100 | 4.72 2.36 |
| 150 | 2.99 1.50 |
| 200 | 2.07 1.03 |
| 250 | 1.57 0.79 |
| komi 7.5 is 0.536 of the opening T | |
| part 3: walling off an a x a area | |
| a corner side centre points per stone c / s / m | |
| 3 | 6 9 12 1.50 / 1.00 / 0.75 |
| 4 | 8 12 16 2.00 / 1.33 / 1.00 |
| 5 | 10 15 20 2.50 / 1.67 / 1.25 |
| 6 | 12 18 24 3.00 / 2.00 / 1.50 |
查看完整验算程序
# Chapter 7. p71: three sentences of the Classic in Thirteen Chapters
# turned into numbers
# Part 1: the board of 19 x 19 points and the classic's 360 + 1.
# Part 2: "rather lose several stones than lose the initiative". The
# value of the initiative is taken as the size T of the biggest move
# on the board (illustrative schedule T = 1 + 13 exp(-m / 80) points,
# about 14 at the start, close to twice a komi of 7.5); a captured
# stone costs about 2 points under area scoring, so the break-even
# number of stones is T / 2. Part 3: stones needed to wall off an
# a x a area in a corner, on a side and in the centre of the board,
# counted exactly as empty neighbours on the 19 x 19 grid.
import math
SIZE = 19
KOMI = 7.5
def tempo_value(move):
return 1.0 + 13.0 * math.exp(-move / 80.0)
def wall(a, where):
if where == "corner":
r0, c0 = 0, 0
elif where == "side":
r0, c0 = 0, (SIZE - a) // 2
else:
r0, c0 = (SIZE - a) // 2, (SIZE - a) // 2
inside = {(r, c) for r in range(r0, r0 + a) for c in range(c0, c0 + a)}
border = set()
for r, c in inside:
for dr, dc in ((1, 0), (-1, 0), (0, 1), (0, -1)):
q = (r + dr, c + dc)
if 0 <= q[0] < SIZE and 0 <= q[1] < SIZE and q not in inside:
border.add(q)
return len(border)
if __name__ == "__main__":
print("part 1: points on the board {} = 360 + {}".format(
SIZE * SIZE, SIZE * SIZE - 360))
print("part 2: stones worth one tempo (T / 2)")
print(" move T (points) break-even stones")
for m in (0, 50, 100, 150, 200, 250):
t = tempo_value(m)
print(" {:3d} {:5.2f} {:.2f}".format(m, t, t / 2.0))
print(" komi {} is {:.3f} of the opening T".format(
KOMI, KOMI / tempo_value(0)))
print("part 3: walling off an a x a area")
print(" a corner side centre points per stone c / s / m")
for a in (3, 4, 5, 6):
k = [wall(a, w) for w in ("corner", "side", "centre")]
pts = a * a
print(" {} {:2d} {:2d} {:2d} "
"{:.2f} / {:.2f} / {:.2f}"
.format(a, k[0], k[1], k[2], pts / k[0], pts / k[1],
pts / k[2]))唯一数据源与引用
范雅福(Fan, Y.). (2026). 《围棋人工智能后人类决策质量跃迁、新颖着法率与棋经十三篇策略熵母典》第 7 章「东西方原典考辨:《棋经十三篇》与富兰克林《下棋的道德》」(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-230 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026230,
author = {Fan, Yafu},
title = {Post-AI Leaps in Human Go Decision Quality, Novel-Move Rates and Strategic Entropy in the Thirteen-Chapter Classic of Go},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}