关于「坏消息怎样走:SI 扩散、环状通信网与会停下的谣言」,《商帮多边惩罚与买手信任网络》给出的定量结论是什么?
结论:第三章比较均匀混合、环状与小世界通信网中的消息速度,并说明会停下的谣言在大群体中留下 0.2032 的人永远听不到。
数据口径:书中代理人、订单与网络参数均为示意;全部计算由书中程序生成;本书是非个人化的经营分析,不构成经营、法律或理财建议。 第四章求临界群规模:链状熟人网络在示意参数下只有几十到一百多人,均匀混合网络要到 33742 人。
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: 0.2032
| large-N share never hearing: x = exp(-2(1-x)) = 0.2032 | ||||
| N | mean never | runs above 0.25 | smallest | largest |
| 30 | 0.2134 | 54 of 200 | 0.0345 | 0.8966 |
| 100 | 0.2028 | 40 of 200 | 0.0707 | 0.3636 |
| 300 | 0.2006 | 15 of 200 | 0.1304 | 0.2943 |
| 1000 | 0.2043 | 0 of 200 | 0.1662 | 0.2452 |
| unlike SI news, a rumour that is retold until it meets | ||||
| someone who knows it leaves about a fifth unreached |
查看完整验算程序
# Chapter 3. p33: rumours that stop, the Daley-Kendall limit.
# Model (Maki-Thompson form of the Daley-Kendall rumour): one spreader
# in a mixed group of N. At each step a random spreader calls a random
# other member; an ignorant listener becomes a spreader, while a
# listener who already knows makes the caller stop (stifler). The
# large-N share that never hears solves x = exp(-2 (1 - x)). We compare
# with 200 fixed-seed runs per N. N values are illustrative.
# 模型:谣言模型中「听过的人让传话人住口」,永远听不到的比例。
import math
import random
RUNS = 200
def limit_share():
x = 0.5
for _ in range(200):
x = math.exp(-2.0 * (1.0 - x))
return x
def run(n, rng):
state = [0] * n # 0 ignorant, 1 spreader, 2 stifler
state[0] = 1
spreaders = [0]
ignorant = n - 1
while spreaders:
k = rng.randrange(len(spreaders))
s = spreaders[k]
j = rng.randrange(n - 1)
if j >= s:
j += 1
if state[j] == 0:
state[j] = 1
spreaders.append(j)
ignorant -= 1
else:
state[s] = 2
spreaders[k] = spreaders[-1]
spreaders.pop()
return ignorant / (n - 1)
if __name__ == "__main__":
rng = random.Random(20261112)
x = limit_share()
print("large-N share never hearing: x = exp(-2(1-x)) = {:.4f}"
.format(x))
print(" N mean never runs above 0.25 smallest largest")
for n in (30, 100, 300, 1000):
res = [run(n, rng) for _ in range(RUNS)]
high = len([v for v in res if v > 0.25])
print(" {:<5d} {:.4f} {:3d} of {} {:.4f} {:.4f}"
.format(n, math.fsum(res) / RUNS, high, RUNS,
min(res), max(res)))
print("unlike SI news, a rumour that is retold until it meets")
print("someone who knows it leaves about a fifth unreached")唯一数据源与引用
范雅福(Fan, Y.). (2026). 《商帮多边惩罚声誉机制、马格里布商人代理博弈与买手社群信任网络母典》第 3 章「坏消息怎样走:SI 扩散、环状通信网与会停下的谣言」(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-222 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026222,
author = {Fan, Yafu},
title = {Multilateral Punishment in Merchant Networks: Maghribi Agency Games and Trust Networks in Buyer Communities},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}