Pair trading 策略 - 期权
0. 引库
import pandas as pd
import numpy as np
import tushare as ts
import seaborn
from matplotlib import pyplot as plt
plt.style.use('seaborn')
%matplotlib inline
stocks_pair = ['cu', 'zn']
1. 数据准备
# 加载数据
data1 = pd.read_csv('cu.csv')[['date','close']]
data2 = pd.read_csv('zn.csv')['close']
# 按行拼接收盘价
data = pd.concat([data1, data2], axis=1)
data.set_index('date',inplace = True)
# 重命名列('cu'、'zn')
data.columns = stocks_pair
data.head()
cu | zn | |
---|---|---|
date | ||
2016/10/18 | 37300 | 18080 |
2016/10/19 | 37260 | 18200 |
2016/10/20 | 37270 | 18255 |
2016/10/21 | 37150 | 18245 |
2016/10/24 | 37320 | 18175 |
data.plot(figsize= (8,6));
2. 策略开发思路
data.corr() # 协方差矩阵
cu | zn | |
---|---|---|
cu | 1.000000 | 0.941114 |
zn | 0.941114 | 1.000000 |
# 数据可视化,看相关关系
plt.figure(figsize =(8,6))
plt.title('Stock Correlation')
plt.plot(data['cu'], data['zn'], '.');
plt.xlabel('cu')
plt.ylabel('zn')
data.dropna(inplace = True)
# 对两股票价格做线性回归(白噪声项符合正态分布)
[slope, intercept] = np.polyfit(data.iloc[:,0], data.iloc[:,1], 1).round(2)
slope,intercept
(0.48, 266.73)
(y-266.73-0.48x) 符合Stationary
# 算出 (y-266.73-0.48x) 一列
data['spread'] = data.iloc[:,1] - (data.iloc[:,0]*slope + intercept)
data.head()
cu | zn | spread | |
---|---|---|---|
date | |||
2016/10/18 | 37300 | 18080 | -90.73 |
2016/10/19 | 37260 | 18200 | 48.47 |
2016/10/20 | 37270 | 18255 | 98.67 |
2016/10/21 | 37150 | 18245 | 146.27 |
2016/10/24 | 37320 | 18175 | -5.33 |
data['spread'].plot(figsize = (8,6),title = 'Price Spread');
# 对 spread 进行标准化
data['zscore'] = (data['spread'] - data['spread'].mean())/data['spread'].std()
data.head()
cu | zn | spread | zscore | |
---|---|---|---|---|
date | ||||
2016/10/18 | 37300 | 18080 | -90.73 | -0.018872 |
2016/10/19 | 37260 | 18200 | 48.47 | 0.192004 |
2016/10/20 | 37270 | 18255 | 98.67 | 0.268053 |
2016/10/21 | 37150 | 18245 | 146.27 | 0.340163 |
2016/10/24 | 37320 | 18175 | -5.33 | 0.110502 |
# 可视化标准化后的值
data['zscore'].plot(figsize = (10,8),title = 'Z-score')
plt.axhline(1.5)
plt.axhline(0)
plt.axhline(-1.5)
产生交易信号
data['position_1'] = np.where(data['zscore'] > 1.5, 1, np.nan)
data['position_1'] = np.where(data['zscore'] < -1.5, -1, data['position_1'])
data['position_1'] = np.where(abs(data['zscore']) < 0.5, 0, data['position_1'])
data['position_1'] = data['position_1'].fillna(method = 'ffill')
data['position_1'].plot(ylim=[-1.1, 1.1], figsize=(10, 6),title = 'Trading Signal_Uptrade');
data['position_2'] = -np.sign(data['position_1'])
data['position_2'].plot(ylim=[-1.1, 1.1], figsize=(10, 6),title = 'Trading Signal_Downtrade');
3. 计算策略年化收益并可视化
# 算离散收益率
data['returns_1'] = np.log(data['cu'] / data['cu'].shift(1))
data['returns_2'] = np.log(data['zn'] / data['zn'].shift(1))
# 算策略列
data['strategy'] = 0.5*(data['position_1'].shift(1) * data['returns_1']) + 0.5*(data['position_2'].shift(1) * data['returns_2'])
# 计算累积收益率
data[['returns_1','returns_2','strategy']].dropna().cumsum().apply(np.exp).tail(1)
returns_1 | returns_2 | strategy | |
---|---|---|---|
date | |||
2018/6/6 | 1.410724 | 1.378042 | 1.35536 |
# 画出累积收益率
data[['returns_1','returns_2','strategy']].dropna().cumsum().apply(np.exp).plot(figsize=(10, 8),title = 'Strategy_Backtesting');
# 计算年化收益率
data[['returns_1','returns_2','strategy']].dropna().mean() * 252
returns_1 0.216785
returns_2 0.202018
strategy 0.191562
dtype: float64
# 计算年化风险
data[['returns_1','returns_2','strategy']].dropna().std() * 252 ** 0.5
returns_1 0.158409
returns_2 0.210702
strategy 0.054305
dtype: float64
# 策略累积收益率
data['cumret'] = data['strategy'].dropna().cumsum().apply(np.exp)
# 策略累积最大值
data['cummax'] = data['cumret'].cummax()
# 算回撤序列
drawdown = (data['cummax'] - data['cumret'])
# 算最大回撤
drawdown.max()
0.026175344283502433
小结
策略的思考
- 对多只ETF进行配对交易,是很多实盘量化基金的交易策略;
策略的风险和问题:
-
Spread不回归的风险,当市场结构发生重大改变时,用过去历史回归出来的Spread会发生不回归的重大风险;
-
中国市场做空受到限制,策略中有部分做空的收益是无法获得的;
-
回归系数需要Rebalancing;
-
策略没有考虑交易成本和其他成本;