13.1 易經、傳統文化 · 老黃歷診股 · Python 量化交易教程

# 13.1 易經、傳統文化 · 老黃歷診股 A：聽說你今天滿倉了？ B：是啊，聽老專家說炒股要看黃歷，我早上查了，今天宜開市。你呢，倉位如何？ A：策略有賣出信號，我昨天就空倉了。 S：怪不得今天跌停，原來是空頭行動了。最近小散發現了一個管用的擇時方法:“看黃歷”。4號忌交易，熔斷；6號宜開市，紅盤。7號宜開市，滿倉殺入。然后就有了上面的對話。等等，好像哪里不對。為了讓大家科學的賺錢，小散決定驗證一下這個策略。第一步：讀入老黃歷data ```py import pandas as pd data = pd.read_csv('老黃歷診股.csv',encoding='GB18030') data = data[['calendarDate','isOpen']] buydata = data[data['isOpen'] == 1] selldata = data[data['isOpen'] == -1] print buydata.head(5) print selldata.head(5) calendarDate isOpen 3 2015-01-08 1 7 2015-01-14 1 8 2015-01-15 1 11 2015-01-20 1 27 2015-02-11 1 calendarDate isOpen 1 2015-01-06 -1 10 2015-01-19 -1 13 2015-01-22 -1 15 2015-01-26 -1 18 2015-01-29 -1 ``` 第二步A，買賣HS300ETF。根據老黃歷信號，宜開市開倉，否則平倉。 ```py buy_date = map(lambda x: x[0:4]+x[5:7]+x[8:10], buydata['calendarDate'].values.tolist()) sell_date = map(lambda x: x[0:4]+x[5:7]+x[8:10], selldata['calendarDate'].values.tolist()) start = '2015-01-01' # 回測起始時間 end = '2015-12-31' # 回測結束時間 benchmark = 'HS300' # 策略參考標準 universe = ['510310.XSHG'] # 股票池 # 策略參考標準 refresh_rate = 1 # 調倉頻率，即每 refresh_rate 個交易日執行一次 handle_data() 函數 def initialize(account): # 初始化虛擬賬戶狀態 pass def handle_data(account): # 每個交易日的買入賣出指令 today=account.current_date.strftime('%Y%m%d') for stk in account.universe: if today in buy_date: order_pct_to(stk, 1) else: if stk in account.avail_secpos: order_to(stk, 0) ``` ![](https://box.kancloud.cn/2016-07-30_579cbdb5c8a24.jpg) 第二步B，買賣HS300ETF。根據老黃歷信號，忌開市開倉，否則平倉。 ```py start = '2015-01-01' # 回測起始時間 end = '2015-12-31' # 回測結束時間 benchmark = 'HS300' # 策略參考標準 universe = ['510310.XSHG'] # 股票池 # 策略參考標準 refresh_rate = 1 # 調倉頻率，即每 refresh_rate 個交易日執行一次 handle_data() 函數 def initialize(account): # 初始化虛擬賬戶狀態 pass def handle_data(account): # 每個交易日的買入賣出指令 today=account.current_date.strftime('%Y%m%d') for stk in account.universe: if today in sell_date: order_pct_to(stk, 1) else: if stk in account.avail_secpos: order_to(stk, 0) ``` ![](https://box.kancloud.cn/2016-07-30_579cbdb5e1e63.jpg) 觀察策略結果，雖然A，B都顯著跑輸大盤，但是A相對B仍然有5%的超額收益。真是應了那句老話： ![](https://box.kancloud.cn/2016-07-31_579d7a030af90.jpg) 本策略告訴小散：炒股有風險，新手慎入。 `老黃歷診股.csv` 數據來源： 1、`DataAPI.TradeCalGet` 得到交易日歷，下載到本地。 2、根據漢典老黃歷開市日歷為`isOpen`字段賦值，宜開市為1，忌開市為-1，無提示為0。 3、上傳到Data使用。 ```py print "Buydate:",buy_date print "Selldate:",sell_date Buydate: [u'20150108', u'20150114', u'20150115', u'20150120', u'20150211', u'20150217', u'20150226', u'20150304', u'20150306', u'20150311', u'20150312', u'20150316', u'20150318', u'20150330', u'20150401', u'20150408', u'20150420', u'20150424', u'20150506', u'20150513', u'20150515', u'20150520', u'20150521', u'20150525', u'20150527', u'20150612', u'20150619', u'20150625', u'20150701', u'20150707', u'20150708', u'20150713', u'20150714', u'20150720', u'20150722', u'20150806', u'20150819', u'20150820', u'20150831', u'20150901', u'20150902', u'20150908', u'20150910', u'20150914', u'20151009', u'20151021', u'20151027', u'20151102', u'20151106', u'20151110', u'20151111', u'20151117', u'20151120', u'20151204', u'20151216'] Selldate: [u'20150106', u'20150119', u'20150122', u'20150126', u'20150129', u'20150130', u'20150203', u'20150212', u'20150213', u'20150216', u'20150305', u'20150319', u'20150326', u'20150331', u'20150402', u'20150403', u'20150409', u'20150416', u'20150423', u'20150427', u'20150428', u'20150430', u'20150505', u'20150512', u'20150522', u'20150526', u'20150605', u'20150616', u'20150626', u'20150702', u'20150715', u'20150723', u'20150727', u'20150729', u'20150807', u'20150817', u'20150825', u'20150826', u'20150828', u'20150907', u'20150915', u'20150916', u'20150917', u'20150923', u'20150925', u'20150928', u'20150929', u'20151015', u'20151016', u'20151022', u'20151028', u'20151029', u'20151030', u'20151103', u'20151104', u'20151112', u'20151116', u'20151119', u'20151124', u'20151127', u'20151201', u'20151202', u'20151203', u'20151211', u'20151218', u'20151221', u'20151223', u'20151224', u'20151225', u'20151228', u'20151230'] ```