# 羊驼反转策略（修改版） > 来源：https://uqer.io/community/share/566c0e3cf9f06c6c8a91ceec ```py # 第一步：设置基本参数 start = '2015-01-01' # 回测起始时间 end = '2015-12-01' # 回测结束时间 capital_base = 1000000 # 起始资金 refresh_rate = 5 # 调仓频率 benchmark = 'HS300' # 策略参考标准 freq = 'd' # 策略类型，'d'表示日间策略使用日线回测 # 第二步：选择主题，设置股票池 universe = set_universe('HS300') # 股票池 import numpy as np import pandas as pd def initialize(account): # 初始化虚拟账户状态 account.stocks_num=10 def handle_data(account): # 每个交易日的买入卖出指令 if account.stocks_num==10: #第一天交易使用buylist account.stocks_num=1 keylist=[] data=DataAPI.MktStockFactorsOneDayGet(tradeDate=account.current_date,secID=account.universe,ticker=u"",field=['secID','REVS10'],pandas="1") #获取start前一日股票池中十日收益 keylist=data.dropna().sort(columns='REVS10',ascending=False).tail(10)['secID'].values.tolist() #将十日收益最差的十只股票组成list #hist_prices = account.get_attribute_history('closePrice', 1) for i in keylist: order(i,100000/account.referencePrice[i]) else: sellist=[] replacelist=[] keylist=[] for key in account.valid_secpos.keys(): keylist.append(key) sell=DataAPI.MktStockFactorsOneDayGet(tradeDate=account.current_date,secID=keylist,ticker=u"",field=['secID','REVS10'],pandas="1") #获得十日账户中所有股票的收益 sellist.append(sell.min()['secID']) #找出收益最差的股票加入sellist replace=DataAPI.MktStockFactorsOneDayGet(tradeDate=account.current_date,secID=universe,ticker=u"",field=['secID','REVS10'],pandas="1") #获得股票池中十日以来 replace=replace.set_index('secID').drop(keylist).dropna() replace=replace.sort(columns='REVS10',ascending=False).tail(1).reset_index()['secID'].values.tolist() #获得收益最差的股票作为账户中新的代替股票 keylist.remove(sellist[0]) replacelist=replacelist+replace keylist.append(replacelist[0]) #print keylist for stk in sellist: order_to(stk, 0) for stk in replacelist: order(stk,account.cash/account.referencePrice[stk]) #print account.valid_secpos ```