Paired trading

来源：https://uqer.io/community/share/54895a8df9f06c31c3950ca0

配对交易

策略思路

寻找走势相关且股价相近的一对股票，根据其价格变动买卖

策略实现

历史前五日的Pearson相关系数若大于给定的阈值则触发买卖操作

from scipy.stats.stats import pearsonr

start = datetime(2013, 1, 1)
end   = datetime(2014, 12, 1)
benchmark = 'HS300'
universe = ['000559.XSHE', '600126.XSHG']
capital_base = 1e6 

corlen = 5

def initialize(account):
    add_history('hist', corlen)
    account.cutoff = 0.9
    account.prev_prc1 = 0
    account.prev_prc2 = 0
    account.prev_prcb = 0

def handle_data(account, data):
    stk1 = universe[0]
    stk2 = universe[1]

    prc1 = data[stk1]['closePrice']
    prc2 = data[stk2]['closePrice']
    prcb = data['HS300']['return']

    px1 = account.hist[stk1]['closePrice'].values
    px2 = account.hist[stk2]['closePrice'].values
    pxb = account.hist['HS300']['return'].values

    corval, pval = pearsonr(px1, px2)

    mov1, mov2 = adj(prc1, prc2, prcb, account.prev_prc1, account.prev_prc2, account.prev_prcb)


    amount =1e4 / prc2
    if (mov1 > 0) and (abs(corval) > account.cutoff):
        order(stk2, amount)
    elif (mov1 < 0) and (abs(corval) > account.cutoff):
        if (account.position.stkpos.get(stk2, 0) > amount):
            order(stk2, -amount)
        else:
            order_to(stk2, 0)

    amount =1e4 / prc1
    if (mov2 > 0) and (abs(corval) > account.cutoff):
        order(stk1, amount)
    elif (mov2 < 0) and (abs(corval) > account.cutoff):
        if (account.position.stkpos.get(stk1, 0) > amount):
            order(stk1, -amount)
        else:
            order_to(stk1, 0)

    account.prev_prc1 = prc1
    account.prev_prc2 = prc2
    account.prev_prcb = prcb


def dmv(curr, prev):
    delta = curr / prev - 1
    return delta

def adj(x, y, base, prev_x, prev_y, prev_base):
    dhs = dmv(base, prev_base)
    dx = dmv(x, prev_x) - dhs
    dy = dmv(y, prev_y) - dhs
    return (dx, dy)

min(bt.cash)

232096.85369499651

import pandas as pd
import numpy as np
from datetime import datetime

import quartz
import quartz.backtest as qb
import quartz.performance as qp
from quartz.api import *

from scipy.stats.stats import pearsonr

start = datetime(2013, 1, 1)                # 回测起始时间
end   = datetime(2014, 12, 1)                # 回测结束时间
benchmark = 'HS300'                            # 使用沪深 300 作为参考标准
capital_base = 1e6                          # 起始资金

corlen = 5

def initialize(account):                    # 初始化虚拟账户状态
    add_history('hist', corlen)
    account.cutoff = 0.9
    account.prev_prc1 = 0
    account.prev_prc2 = 0
    account.prev_prcb = 0

def handle_data(account, data):                # 每个交易日的买入卖出指令
    stk1 = universe[0]
    stk2 = universe[1]

    prc1 = data[stk1]['closePrice']
    prc2 = data[stk2]['closePrice']
    prcb = data['HS300']['return']

    px1 = account.hist[stk1]['closePrice'].values
    px2 = account.hist[stk2]['closePrice'].values
    pxb = account.hist['HS300']['return'].values

    corval, pval = pearsonr(px1, px2)

    mov1, mov2 = adj(prc1, prc2, prcb, account.prev_prc1, account.prev_prc2, account.prev_prcb)

    #amount = int( 0.08 * capital_base / prc2)
    amount =1e4 / prc2
    if (mov1 > 0) and (abs(corval) > account.cutoff):
        order(stk2, amount)
    elif (mov1 < 0) and (abs(corval) > account.cutoff):
        if (account.position.stkpos.get(stk2, 0) > amount):
            order(stk2, -amount)
        else:
            order_to(stk2, 0)

    #amount = int(0.08 * capital_base / prc1)
    amount =1e4 / prc1
    if (mov2 > 0) and (abs(corval) > account.cutoff):
        order(stk1, amount)
    elif (mov2 < 0) and (abs(corval) > account.cutoff):
        if (account.position.stkpos.get(stk1, 0) > amount):
            order(stk1, -amount)
        else:
            order_to(stk1, 0)

    account.prev_prc1 = prc1
    account.prev_prc2 = prc2
    account.prev_prcb = prcb


def dmv(curr, prev):
    delta = curr / prev - 1
    return delta

def adj(x, y, base, prev_x, prev_y, prev_base):
    dhs = dmv(base, prev_base)
    dx = dmv(x, prev_x) - dhs
    dy = dmv(y, prev_y) - dhs
    return (dx, dy)


pool_raw = pd.read_csv("po.pair.2012.csv")
pool = []
for i in range(len(pool_raw)):
    s1, s2 = pool_raw.loc[i].tolist()
    if [s2, s1] not in pool:
        pool.append([s1, s2])

outfile = []
for i, universe in enumerate(pool):
    print i
    try:
        bt = qb.backtest(start, end, benchmark, universe, capital_base, initialize = initialize, handle_data = handle_data)
        perf = qp.perf_parse(bt)
        outfile.append(universe + [perf["annualized_return"], perf["sharpe"]])
    except:
        pass    

keys = ['stock1', 'stock2', 'annualized_return', 'sharpe']
outdict = {}
outfile =  zip(*sorted(outfile, key=lambda x:x[2], reverse=True))
for i,k in enumerate(keys):
    outdict[k] = outfile[i]
outdict = pd.DataFrame(outdict).loc[:, keys]
outdict

['000066.XSHE', '000707.XSHE']
['000066.XSHE', '600117.XSHG']
['000066.XSHE', '600126.XSHG']
['000066.XSHE', '600819.XSHG']
['000089.XSHE', '600035.XSHG']
['000089.XSHE', '600037.XSHG']
['000089.XSHE', '600595.XSHG']
['000159.XSHE', '000967.XSHE']
['000159.XSHE', '600595.XSHG']
['000417.XSHE', '000541.XSHE']
['000417.XSHE', '000685.XSHE']
['000417.XSHE', '600875.XSHG']
['000425.XSHE', '000528.XSHE']
['000507.XSHE', '600391.XSHG']
['000541.XSHE', '000987.XSHE']
['000541.XSHE', '600330.XSHG']
['000541.XSHE', '600883.XSHG']
['000554.XSHE', '000707.XSHE']
['000559.XSHE', '600026.XSHG']
['000559.XSHE', '600126.XSHG']
['000559.XSHE', '600477.XSHG']
['000559.XSHE', '600581.XSHG']
['000559.XSHE', '601666.XSHG']
['000635.XSHE', '000707.XSHE']
['000635.XSHE', '600068.XSHG']
['000635.XSHE', '600117.XSHG']
['000635.XSHE', '600188.XSHG']
['000635.XSHE', '600295.XSHG']
['000635.XSHE', '600550.XSHG']
['000635.XSHE', '600819.XSHG']
['000635.XSHE', '601168.XSHG']
['000635.XSHE', '601233.XSHG']
['000650.XSHE', '600261.XSHG']
['000683.XSHE', '000936.XSHE']
['000683.XSHE', '600595.XSHG']
['000685.XSHE', '000988.XSHE']
['000685.XSHE', '601101.XSHG']
['000698.XSHE', '000949.XSHE']
['000707.XSHE', '000911.XSHE']
['000707.XSHE', '000969.XSHE']
['000707.XSHE', '000987.XSHE']
['000707.XSHE', '600117.XSHG']
['000707.XSHE', '600295.XSHG']
['000707.XSHE', '600550.XSHG']
['000707.XSHE', '600831.XSHG']
['000707.XSHE', '601168.XSHG']
['000707.XSHE', '601233.XSHG']
['000708.XSHE', '600327.XSHG']
['000709.XSHE', '601107.XSHG']
['000709.XSHE', '601618.XSHG']
['000717.XSHE', '600282.XSHG']
['000717.XSHE', '600307.XSHG']
['000717.XSHE', '600808.XSHG']
['000761.XSHE', '600320.XSHG']
['000761.XSHE', '600548.XSHG']
['000822.XSHE', '600117.XSHG']
['000830.XSHE', '600068.XSHG']
['000830.XSHE', '600320.XSHG']
['000830.XSHE', '600550.XSHG']
['000877.XSHE', '601519.XSHG']
['000898.XSHE', '600022.XSHG']
['000898.XSHE', '600808.XSHG']
['000911.XSHE', '600550.XSHG']
['000916.XSHE', '600033.XSHG']
['000916.XSHE', '600035.XSHG']
['000916.XSHE', '600126.XSHG']
['000930.XSHE', '600026.XSHG']
['000932.XSHE', '600569.XSHG']
['000933.XSHE', '600348.XSHG']
['000933.XSHE', '600595.XSHG']
['000936.XSHE', '600477.XSHG']
['000937.XSHE', '600348.XSHG']
['000937.XSHE', '600508.XSHG']
['000937.XSHE', '600997.XSHG']
['000937.XSHE', '601001.XSHG']
['000939.XSHE', '600819.XSHG']
['000967.XSHE', '600879.XSHG']
['000969.XSHE', '600831.XSHG']
['000973.XSHE', '600460.XSHG']
['000987.XSHE', '600636.XSHG']
['000987.XSHE', '600827.XSHG']
['000987.XSHE', '601001.XSHG']
['600008.XSHG', '600035.XSHG']
['600012.XSHG', '600428.XSHG']
['600020.XSHG', '600033.XSHG']
['600020.XSHG', '600035.XSHG']
['600026.XSHG', '600068.XSHG']
['600026.XSHG', '600089.XSHG']
['600026.XSHG', '600126.XSHG']
['600026.XSHG', '600307.XSHG']
['600026.XSHG', '600331.XSHG']
['600026.XSHG', '600375.XSHG']
['600026.XSHG', '600581.XSHG']
['600026.XSHG', '600963.XSHG']
['600026.XSHG', '601666.XSHG']
['600026.XSHG', '601898.XSHG']
['600033.XSHG', '600035.XSHG']
['600035.XSHG', '600126.XSHG']
['600035.XSHG', '600269.XSHG']
['600035.XSHG', '600307.XSHG']
['600035.XSHG', '600586.XSHG']
['600037.XSHG', '600327.XSHG']
['600068.XSHG', '600126.XSHG']
['600068.XSHG', '600269.XSHG']
['600068.XSHG', '600320.XSHG']
['600068.XSHG', '600550.XSHG']
['600068.XSHG', '601001.XSHG']
['600068.XSHG', '601666.XSHG']
['600089.XSHG', '600581.XSHG']
['600100.XSHG', '600117.XSHG']
['600117.XSHG', '600295.XSHG']
['600117.XSHG', '600339.XSHG']
['600117.XSHG', '601168.XSHG']
['600117.XSHG', '601233.XSHG']
['600126.XSHG', '600282.XSHG']
['600126.XSHG', '600327.XSHG']
['600126.XSHG', '600569.XSHG']
['600126.XSHG', '600581.XSHG']
['600126.XSHG', '600808.XSHG']
['600126.XSHG', '600963.XSHG']
['600160.XSHG', '600449.XSHG']
['600160.XSHG', '601216.XSHG']
['600160.XSHG', '601311.XSHG']
['600188.XSHG', '600295.XSHG']
['600188.XSHG', '601001.XSHG']
['600231.XSHG', '600282.XSHG']
['600269.XSHG', '601618.XSHG']
['600282.XSHG', '600307.XSHG']
['600282.XSHG', '600569.XSHG']
['600282.XSHG', '600808.XSHG']
['600282.XSHG', '600963.XSHG']
['600307.XSHG', '600581.XSHG']
['600307.XSHG', '600808.XSHG']
['600307.XSHG', '600963.XSHG']
['600320.XSHG', '600548.XSHG']
['600320.XSHG', '601600.XSHG']
['600330.XSHG', '600883.XSHG']
['600330.XSHG', '601268.XSHG']
['600331.XSHG', '600581.XSHG']
['600348.XSHG', '600508.XSHG']
['600348.XSHG', '600997.XSHG']
['600348.XSHG', '601001.XSHG']
['600368.XSHG', '600527.XSHG']
['600375.XSHG', '600581.XSHG']
['600391.XSHG', '601100.XSHG']
['600449.XSHG', '601311.XSHG']
['600449.XSHG', '601519.XSHG']
['600460.XSHG', '601908.XSHG']
['600477.XSHG', '600581.XSHG']
['600508.XSHG', '600546.XSHG']
['600508.XSHG', '600997.XSHG']
['600522.XSHG', '600973.XSHG']
['600550.XSHG', '600831.XSHG']
['600569.XSHG', '600808.XSHG']
['600569.XSHG', '600963.XSHG']
['600581.XSHG', '600963.XSHG']
['600581.XSHG', '601001.XSHG']
['600581.XSHG', '601168.XSHG']
['600581.XSHG', '601666.XSHG']
['600586.XSHG', '601268.XSHG']
['600595.XSHG', '601001.XSHG']
['600595.XSHG', '601168.XSHG']
['600595.XSHG', '601666.XSHG']
['600688.XSHG', '600871.XSHG']
['600785.XSHG', '600827.XSHG']
['600808.XSHG', '600963.XSHG']
['600827.XSHG', '601001.XSHG']
['600875.XSHG', '601001.XSHG']
['600883.XSHG', '601268.XSHG']
['601001.XSHG', '601101.XSHG']
['601001.XSHG', '601168.XSHG']
['601001.XSHG', '601666.XSHG']
['601101.XSHG', '601666.XSHG']
['601168.XSHG', '601666.XSHG']

174 rows × 4 columns

a = list(outfile[2])
'percentage of outperform HS300: %f' % (1.*len([x for x in a if x>0.117]) / len(a))

'percentage of outperform HS300: 0.741379'