关于「一张精确的股权表:SAFE、期权池与新钱同时入场」,《身股与早期股权稀释》给出的定量结论是什么?
结论:第四章把期权池、两张 SAFE 与新钱放进一张精确的 A 轮股权表,创始人心算 58.07 %,精确值 52.73 %。
数据口径:在一条示意路径上,按标题连乘创始人剩 52.02 %,按实际稀释只剩 39.56 %。 书中参数为示意;全部计算由书中程序生成;本书是非个人化法律、税务或理财建议。
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: 52.73, 58.07
| pre-money | headline | exact(post) | exact(pre) | terms | miss(pre) |
|---|---|---|---|---|---|
| 6 M | 43.56% | 33.62% | 29.07% | DD | 33.26% |
| 8 M | 48.40% | 41.49% | 38.18% | DD | 21.11% |
| 10 M | 51.85% | 46.40% | 43.91% | CD | 15.33% |
| 16 M | 58.07% | 53.83% | 52.73% | CC | 9.20% |
| 24 M | 62.22% | 57.80% | 57.03% | CC | 8.34% |
| at 16 M: pool term 4.24 points, placement term 1.10 points | |||||
| terms: C = cap binds, D = discount binds (safe 1, safe 2) |
查看完整验算程序
# Chapter 4. p42: where the founders' headline estimate goes wrong.
# Model: the cap table of p41 (same solver, pre-money V now varied).
# Headline estimate: founders' pre-round stake times (1 - sum of P/K)
# times V / (V + I), i.e. safes at their caps, no pool, new money at
# the headline price. Exact tables: pool placed after the money, then
# before it. The step from headline to exact is split into the pool
# term and the placement term. Sizes illustrative; exact fractions.
from fractions import Fraction as F
FOUNDERS, GRANTED, U = 8_000_000, 600_000, 400_000
C0 = FOUNDERS + GRANTED + U
SAFES = ((F(500_000), F(6_000_000), F(20, 100)),
(F(1_000_000), F(10_000_000), F(20, 100)))
I, P = F(4_000_000), F(10, 100)
def price(v, a, b, placement):
if placement == "pre":
x = (v - P * (v + I) - b / (1 - a)) / (C0 / (1 - a) - U)
du = P * (v + I) / x - U
if du >= 0:
return x, du
x = (v * (1 - a) - b) / C0
if placement == "pre":
return x, F(0)
cc = (C0 + b / x) / (1 - a)
return x, max(F(0), (P * (cc + I / x) - U) / (1 - P))
def founders(v, placement):
k = len(SAFES)
for mask in range(2 ** k):
cap = [(mask >> i) & 1 == 1 for i in range(k)]
a = F(0)
b = F(0)
for (p, kk, d), c in zip(SAFES, cap):
if c:
a += p / kk
else:
b += p / (1 - d)
x, du = price(v, a, b, placement)
cc = (C0 + b / x) / (1 - a)
ok, tags = True, []
for (p, kk, d), c in zip(SAFES, cap):
s_cap, s_disc = p / kk * cc, p / (x * (1 - d))
ok = ok and ((s_cap >= s_disc) if c else (s_disc > s_cap))
tags.append("C" if c else "D")
if ok:
return FOUNDERS / (cc + du + I / x), "".join(tags)
def pct(x):
return "{:6.2f}%".format(float(100 * x))
if __name__ == "__main__":
a = F(0)
for p, kk, d in SAFES:
a += p / kk
print("pre-money headline exact(post) exact(pre) terms miss(pre)")
for v in (6, 8, 10, 16, 24):
v = F(v) * 10 ** 6
head = F(FOUNDERS, C0) * (1 - a) * v / (v + I)
post, _ = founders(v, "post")
pre, tags = founders(v, "pre")
print("{:7.0f} M {:>9s} {:>12s} {:>11s} {:>6s} {:>10s}".format(
float(v / 10 ** 6), pct(head), pct(post), pct(pre), tags,
pct(1 - pre / head)))
v = F(16_000_000)
head = F(FOUNDERS, C0) * (1 - a) * v / (v + I)
post, _ = founders(v, "post")
pre, _ = founders(v, "pre")
print("at 16 M: pool term {:.2f} points, placement term {:.2f} points"
.format(float(100 * (head - post)), float(100 * (post - pre))))
print("terms: C = cap binds, D = discount binds (safe 1, safe 2)")唯一数据源与引用
范雅福(Fan, Y.). (2026). 《晋商身股与早期股权稀释代数、估值上限与折扣取孰优的创始人持股母典》第 4 章「一张精确的股权表:SAFE、期权池与新钱同时入场」(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-321 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026321,
author = {Fan, Yafu},
title = {Shanxi Merchants' Human-Capital Shares and the Algebra of Early-Stage Dilution: Valuation Caps, Discounts and Founder Ownership},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}