from __future__ import print_function, absolute_import, unicode_literals
import numpy as np
from gm.api import *
try:
import statsmodels.tsa.stattools as ts
except:
print('请安装statsmodels库')
'''
本策略根据EG两步法(1.序列同阶单整2.OLS残差平稳)判断序列具有协整关系之后(若无协整关系则全平仓位不进行操作)
通过计算两个真实价格序列回归残差的0.9个标准差上下轨,并在价差突破上轨的时候做空价差,价差突破下轨的时候做多价差
并在回归至标准差水平内的时候平仓
回测数据为:SHFE.rb1801和SHFE.rb1805的1min数据
回测时间为:2017-09-25 08:00:00到2017-10-01 15:00:00
'''
def cointegration_test(series01, series02):
urt_rb1801 = ts.adfuller(np.array(series01), 1)[1]
urt_rb1805 = ts.adfuller(np.array(series01), 1)[1]
if (urt_rb1801 > 0.1 and urt_rb1805 > 0.1) or (urt_rb1801 < 0.1 and urt_rb1805 < 0.1):
urt_diff_rb1801 = ts.adfuller(np.diff(np.array(series01)), 1)[1]
urt_diff_rb1805 = ts.adfuller(np.diff(np.array(series01), 1))[1]
if urt_diff_rb1801 < 0.1 and urt_diff_rb1805 < 0.1:
matrix = np.vstack([series02, np.ones(len(series02))]).T
beta, c = np.linalg.lstsq(matrix, series01)[0]
resid = series01 - beta * series02 - c
if ts.adfuller(np.array(resid), 1)[1] > 0.1:
result = 0.0
else:
result = 1.0
return beta, c, resid, result
else:
result = 0.0
return 0.0, 0.0, 0.0, result
else:
result = 0.0
return 0.0, 0.0, 0.0, result
def init(context):
context.goods = ['SHFE.rb1801', 'SHFE.rb1805']
subscribe(symbols=context.goods, frequency='60s', count=801, wait_group=True)
def on_bar(context, bars):
close_01 = context.data(symbol=context.goods[0], frequency='60s', count=801, fields='close')['close'].values
close_02 = context.data(symbol=context.goods[1], frequency='60s', count=801, fields='close')['close'].values
beta, c, resid, result = cointegration_test(close_01, close_02)
if not result:
print('协整检验不通过,全平所有仓位')
order_close_all()
return
mean = np.mean(resid)
up = mean + 0.9 * np.std(resid)
down = mean - 0.9 * np.std(resid)
resid_new = close_01[-1] - beta * close_02[-1] - c
position_01_long = context.account().position(symbol=context.goods[0], side=PositionSide_Long)
position_01_short = context.account().position(symbol=context.goods[0], side=PositionSide_Short)
if not position_01_long and not position_01_short:
if resid_new > up:
order_target_volume(symbol=context.goods[0], volume=1, order_type=OrderType_Market,
position_side=PositionSide_Short)
print(context.goods[0] + '以市价单开空仓1手')
order_target_volume(symbol=context.goods[1], volume=1, order_type=OrderType_Market,
position_side=PositionSide_Long)
print(context.goods[1] + '以市价单开多仓1手')
if resid_new < down:
order_target_volume(symbol=context.goods[0], volume=1, order_type=OrderType_Market,
position_side=PositionSide_Long)
print(context.goods[0], '以市价单开多仓1手')
order_target_volume(symbol=context.goods[1], volume=1, order_type=OrderType_Market,
position_side=PositionSide_Short)
print(context.goods[1], '以市价单开空仓1手')
elif position_01_short:
if resid_new <= up:
order_close_all()
print('价格回归,平掉所有仓位')
if resid_new < down:
order_target_volume(symbol=context.goods[0], volume=1, order_type=OrderType_Market,
position_side=PositionSide_Long)
print(context.goods[0], '以市价单开多仓1手')
order_target_volume(symbol=context.goods[1], volume=1, order_type=OrderType_Market,
position_side=PositionSide_Short)
print(context.goods[1], '以市价单开空仓1手')
elif position_01_long:
if resid_new >= down:
order_close_all()
print('价格回归,平所有仓位')
if resid_new > up:
order_target_volume(symbol=context.goods[0], volume=1, order_type=OrderType_Market,
position_side=PositionSide_Short)
print(context.goods[0], '以市价单开空仓1手')
order_target_volume(symbol=context.goods[1], volume=1, order_type=OrderType_Market,
position_side=PositionSide_Long)
print(context.goods[1], '以市价单开多仓1手')
if __name__ == '__main__':
'''
strategy_id策略ID,由系统生成
filename文件名,请与本文件名保持一致
mode实时模式:MODE_LIVE回测模式:MODE_BACKTEST
token绑定计算机的ID,可在系统设置-密钥管理中生成
backtest_start_time回测开始时间
backtest_end_time回测结束时间
backtest_adjust股票复权方式不复权:ADJUST_NONE前复权:ADJUST_PREV后复权:ADJUST_POST
backtest_initial_cash回测初始资金
backtest_commission_ratio回测佣金比例
backtest_slippage_ratio回测滑点比例
'''
run(strategy_id='strategy_id',
filename='main.py',
mode=MODE_BACKTEST,
token='token_id',
backtest_start_time='2017-09-25 08:00:00',
backtest_end_time='2017-10-01 16:00:00',
backtest_adjust=ADJUST_PREV,
backtest_initial_cash=500000,
backtest_commission_ratio=0.0001,
backtest_slippage_ratio=0.0001)