Simple Bollinger Bands

来源；https://uqer.io/community/share/54b5c1b4f9f06c276f651a15

策略思路： 在一段时间内，股票价格可以认为是在一定水平上下波动。例如，可以简单认为股价在过去100天的平均值基础上，上下波动几个标准差。

策略实现： (1) 股价高于这个波动区间，说明股价虚高，故卖出 (2) 股价低于这个波动区间，说明股价虚低，故买入

PS：本策略适用于股市平稳波动时期，在大牛市或者大熊市不太适用

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

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

import talib

start = datetime(2011, 1, 1)
end = datetime(2014, 10, 1)
benchmark = 'HS300'
universe = ['601398.XSHG', '600028.XSHG', '601988.XSHG', '600036.XSHG', '600030.XSHG',
            '601318.XSHG', '600000.XSHG', '600019.XSHG', '600519.XSHG', '601166.XSHG']
capital_base = 1000000
refresh_rate = 5
window = 200

def initialize(account):
    account.amount = 10000
    account.universe = universe
    add_history('hist', window)

def handle_data(account):

    for stk in account.universe:
        prices = account.hist[stk]['closePrice']
        if prices is None:
            return

        mu = prices.mean()
        sd = prices.std()

        upper = mu + 2*sd
        middle = mu
        lower = mu - 2*sd

        cur_pos = account.position.stkpos.get(stk, 0)
        cur_prc = prices[-1]
        if cur_prc > upper and cur_pos >= 0:
            order_to(stk, 0)
        if cur_prc < lower and cur_pos <= 0:
            order(stk, account.amount)

bt

709 rows × 6 columns

perf = qp.perf_parse(bt)
out_keys = ['annualized_return', 'volatility', 'information',
            'sharpe', 'max_drawdown', 'alpha', 'beta']

for k in out_keys:
    print '%s: %s' % (k, perf[k])

annualized_return: 0.0594823977045
volatility: 0.214571802928
information: 0.407163653335
sharpe: 0.102806600884
max_drawdown: 0.36585509048
alpha: 0.052294592526
beta: 0.434097117249

perf['cumulative_return'].plot()
perf['benchmark_cumulative_return'].plot()
pylab.legend(['current_strategy','HS300'])

<matplotlib.legend.Legend at 0x568fe90>