关于「实战工具与检查器规格:低波动窗口看跌头寸台账与预注册检验」,《低波动窗口的看跌期权》给出的定量结论是什么?
结论:第八章给出实战工具与检查器规格:试算器「低波动窗口看跌头寸双希腊值情景与凯利仓位上限试算器」、按四件套分组的检查项、台账字段与预注册序贯检验,盈亏平衡胜率 0.2941 时,固定样本检验需要 62 笔交易。书中以标注框区分已发表的证据、可运行的 Python 计算(22 段程序)、原典与作者的假说。
数据口径:书中的「做空」只指损失有上限的看跌敞口,即买入看跌期权与看跌借方价差;全部价格以合成指数的指数点表示,不使用任何历史行情。 书中参数为示意;全部计算由书中程序生成。
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: 0.2941, 62
break-even hit rate L / (W + L) = 0.2941 upper bound ln A +2.7726, lower bound ln B -1.5581 step per win +0.4055, per loss -0.2412 Wald average sample number under H0: 28.4 trades Wald average sample number under H1: 38.3 trades simulated under H0: accept H1 0.0432, mean N 31.3, 90th pct N 66 simulated under H1: accept H1 0.8194, mean N 42.3, 90th pct N 82 fixed-sample one-sided z-test needs 62 trades
查看完整验算程序
# Chapter 8. p82: a pre-registered sequential test of one's own hit rate.
# Break-even hit rate p_be = L / (W + L), with W = mean gain of a winning
# put and L = mean loss of a losing one, both in premium multiples
# (illustrative W 1.8, L 0.75). Wald's (1945) SPRT tests H0: p = 0.30
# against H1: p = 0.45 with alpha 0.05 and beta 0.20. Printed: the
# boundaries, the log-likelihood steps, Wald's average sample numbers,
# a 20,000-run simulation (truncated at 400 trades), and the fixed
# sample size a one-sided binomial z-test would need.
import math
import random
from statistics import NormalDist
W, L = 1.8, 0.75
P0, P1, ALPHA, BETA = 0.30, 0.45, 0.05, 0.20
print("break-even hit rate L / (W + L) = {:.4f}".format(L / (W + L)))
A = math.log((1.0 - BETA) / ALPHA)
B = math.log(BETA / (1.0 - ALPHA))
win = math.log(P1 / P0)
loss = math.log((1.0 - P1) / (1.0 - P0))
print("upper bound ln A {:+.4f}, lower bound ln B {:+.4f}".format(A, B))
print("step per win {:+.4f}, per loss {:+.4f}".format(win, loss))
for name, p, oc in (("H0", P0, 1.0 - ALPHA), ("H1", P1, BETA)):
ez = p * win + (1.0 - p) * loss
asn = (oc * B + (1.0 - oc) * A) / ez
print("Wald average sample number under {}: {:.1f} trades".format(
name, asn))
rng = random.Random(42582)
for name, p in (("H0", P0), ("H1", P1)):
acc, ns = 0, []
for _ in range(20000):
z, n = 0.0, 0
while B < z < A and n < 400:
n += 1
z += win if rng.random() < p else loss
acc += z >= A
ns.append(n)
ns.sort()
print("simulated under {}: accept H1 {:.4f}, mean N {:.1f},"
" 90th pct N {}".format(name, acc / 20000,
math.fsum(ns) / 20000, ns[18000]))
za = NormalDist().inv_cdf(1.0 - ALPHA)
zb = NormalDist().inv_cdf(1.0 - BETA)
n = ((za * math.sqrt(P0 * (1.0 - P0)) + zb * math.sqrt(P1 * (1.0 - P1)))
/ (P1 - P0)) ** 2
print("fixed-sample one-sided z-test needs {} trades".format(math.ceil(n)))唯一数据源与引用
范雅福(Fan, Y.). (2026). 《宏观下行期低隐含波动率做空窗口、双希腊值凸性共振与三十日胜率边界母典》第 8 章「实战工具与检查器规格:低波动窗口看跌头寸台账与预注册检验」(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-425 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026425,
author = {Fan, Yafu},
title = {Shorting Low-Implied-Volatility Windows in Macro Downturns: Delta-Vega Dual Convexity, Bear-Rally Compression and Thirty-Day Hit-Rate Bounds},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}