技術分析入門【2】 —— 大家搶籌碼（06年至12年版）— 更新版 · Python 量化交易教程

# 技術分析入門【2】 —— 大家搶籌碼（06年至12年版）— 更新版 > 來源：https://uqer.io/community/share/568e6f54228e5b18e5ba296e 從社區李大大以前的帖子，稍作修改，適合現在的uqer版本，感謝李大大的無私分享！原帖地址： https://uqer.io/community/share/5541d8a4f9f06c1c3d687fef 在本篇中，我們將使用流通股份的集中程度作為指標，為大家開發如何機智的搶籌碼策略！股市里面總是有這樣的一種說法：大股東總是會快小散一步，悄悄地進村，放槍的不要。大股東會在建倉期吸收世面上的廉價籌碼，然后放出利好，逢高出貨。所以大股東的建倉期，正是小散們入場分湯的好時機！ ## 1. 數據準備好了，說了這些原理，到底靈不靈呢？來，一試便知！這里我們首先要定義什么叫大股東呢？這里我們借助中誠信的數據，獲取前十大流通股東的持股比例：數據API： `CCXE.EquMainshFCCXEGet` 獲取財報中十大流通股股東的持股比例（本API需要在數據商城購買）下面的語句查詢`600000.XSHG`浦發銀行在2014年9月30日到2014年12月31日的十大流通股股東持股情況： ```py import datetime as dt from CAL.PyCAL import * data = DataAPI.CCXE.EquMainshFCCXEGet('600000.XSHG', endDateStart='20140930', endDateEnd='20141231') data.head() ``` | | secID | ticker | exchangeCD | secShortName | secShortNameEn | endDate | shNum | shRank | shName | holdVol | holdPct | shareCharType | | --- | --- | | 0 | 600000.XSHG | 600000 | XSHG | 浦發銀行 | NaN | 2014-12-31 00:00:00 | 1 | 1 | 上海國際集團有限公司 | 3157513917 | 16.93 | 101 | | 1 | 600000.XSHG | 600000 | XSHG | 浦發銀行 | NaN | 2014-12-31 00:00:00 | 2 | 2 | 上海國際信托有限公司 | 975923794 | 5.23 | 101 | | 2 | 600000.XSHG | 600000 | XSHG | 浦發銀行 | NaN | 2014-12-31 00:00:00 | 3 | 3 | 上海國鑫投資發展有限公司 | 377101999 | 2.02 | 101 | | 3 | 600000.XSHG | 600000 | XSHG | 浦發銀行 | NaN | 2014-12-31 00:00:00 | 4 | 4 | 百聯集團有限公司 | 190083517 | 1.02 | 101 | | 4 | 600000.XSHG | 600000 | XSHG | 浦發銀行 | NaN | 2014-12-31 00:00:00 | 5 | 5 | 雅戈爾集團股份有限公司 | 162000000 | 0.87 | 101 | 我們按照報表日進行合并，并計算前十大流通股股東持股總比例： ```py data.groupby('endDate').sum() ``` 可以看到，2014年年報中流通股集中度是下降的，相對于上一個季報，持股總比例從29.76%降到了29.25%。看來他的大股東沒啥動靜，小散們先按兵不動！ ## 2. 策略思路有一句俗話：不要在一棵樹上吊死！小散們可以“海選PK”，擇優錄取！我們以上證50成分股為例，挑選出滿足以下條件的股票： + 2015年一季度季報中10大流通股股東持股比例相對于去年年報上升10% 這就是我們認定的大股東吸籌碼的標志： ```py from quartz.api import set_universe import datetime as dt universe = set_universe('SH50') for stock in universe: try: data = DataAPI.CCXE.EquMainshFCCXEGet(stock, endDateStart='20141231', endDateEnd='20150331') except: continue res = data.groupby('endDate').sum()[-2:] if len(res.index) == 2 and res.index[1] == '2015-03-31 00:00:00': chg = res['holdPct'].values[1] / res['holdPct'].values[0] - 1.0 if chg > 0.1: print '%s: %.4f' % (stock, chg) ``` 選出來有三只股票滿足：`601169.XSHG`, `600887.XSHG`, `600703.XSHG` 下面的股價走勢圖來看，這樣的股票總體還是上升的。但是按照這樣投錢真的靠譜嗎？ ```py import pandas as pd stock1 = DataAPI.MktEqudAdjGet(secID=['601169.XSHG'], beginDate='20150331', endDate='20150429', field = ['closePrice', 'tradeDate']) stock2 = DataAPI.MktEqudAdjGet(secID=['600887.XSHG'], beginDate='20150331', endDate='20150429', field = ['closePrice', 'tradeDate']) stock3 = DataAPI.MktEqudAdjGet(secID=['600703.XSHG'], beginDate='20150331', endDate='20150429', field = ['closePrice', 'tradeDate']) ``` ```py import seaborn as sns sns.set_style('white') total = pd.DataFrame({'601169.XSHG':stock1.closePrice.values, '600887.XSHG':stock2.closePrice.values, '600703.XSHG':stock3.closePrice.values}) total.index = stock1.tradeDate.apply(lambda x: dt.datetime.strptime(x, '%Y-%m-%d')) total.plot(subplots=True, figsize=(12,8)) array([<matplotlib.axes.AxesSubplot object at 0x5543d10>, <matplotlib.axes.AxesSubplot object at 0x5572850>, <matplotlib.axes.AxesSubplot object at 0x56a62d0>], dtype=object) ``` ![](https://box.kancloud.cn/2016-07-30_579cbac17f300.png) ## 3. 完整策略我們來吧上面的想法系統化，來看這個策略效率： + 投資域：上證50成分股 + 業績基準：上證50指數 + 調倉頻率：3個月 + 調倉日期：每年的2月28日，5月31日，8月30日，11月30日，遇到節假日的話向后順延 + 開倉信號：十大流通股股東持股比例集中度上升10% + 清倉信號：每個調倉日前一個工作日，清空當前倉位 + 買入方式：等比例買入 + 回測周期：2006年1月1日至2015年4月28日這里的調倉日期的設置，是滿足每期報表結束日后的兩個月，這樣我們有比較大的把握，可以確實拿到當前的報表數據。 ```py import datetime as dt start = '2006-01-01' # 回測起始時間 end = '2012-12-31' # 回測結束時間 benchmark = 'SH50' # 策略參考標準 universe = set_universe('SH50') # 證券池，支持股票和基金 capital_base = 100000 # 起始資金 longest_history = 1 # handle_data 函數中可以使用的歷史數據最長窗口長度 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 account.reportingPair = [('0930', '1231'), ('1231', '0331'), ('0331', '0630'), ('0630', '0930')] def handle_data(account): # 每個交易日的買入賣出指令 hist = account.get_history(longest_history) today = account.current_date year = today.year rebalance_dates = [dt.datetime(year, 2, 28), dt.datetime(year, 5,31), dt.datetime(year, 8, 30), dt.datetime(year, 11,30)] cal = Calendar('China.SSE') rebalance_dates = [cal.adjustDate(d, BizDayConvention.Following) for d in rebalance_dates] rebalanceFlag = False period = -1 for i, d in enumerate(rebalance_dates): # 判斷是否是調倉日 if today == d.toDateTime(): rebalanceFlag = True period = i break # 調倉日前一個交易日，清空所有的倉位 elif today == cal.advanceDate(d, '-1B').toDateTime(): for stock in account.valid_secpos: order_to(stock,0) if rebalanceFlag: if period == 0: year -= 1 # 確定當前調倉日對應需要查詢的報表日期 if account.reportingPair[period][0] < account.reportingPair[period][1]: endDateStart = str(year) + account.reportingPair[period][0] else: endDateStart = str(year-1) + account.reportingPair[period][0] endDateEnd = str(year) + account.reportingPair[period][1] buyList = [] # 確定哪些股票滿足“籌碼”集中要求 for stock in account.universe: try: data = DataAPI.CCXE.EquMainshFCCXEGet(stock, endDateStart=endDateStart, endDateEnd=endDateEnd) except: continue res = data.groupby('endDate').sum()[-2:] tmp = account.reportingPair[period][1] if len(res.index) == 2 and res.index[1] == str(year) + '-' + tmp[:2] + '-' + tmp[2:]+ ' 00:00:00': chg = res['holdPct'].values[1] / res['holdPct'].values[0] - 1.0 if chg > 0.1: buyList.append(stock) print u"%s 買入 : %s" % (today, buyList) # 等權重買入 if len(buyList) != 0: singleCash = account.cash / len(buyList) for stock in buyList: approximationAmount = int(singleCash / hist[stock]['closePrice'][-1]/100.0) * 100 order(stock, approximationAmount) ``` ![](https://box.kancloud.cn/2016-07-30_579cbac19545d.jpg) ``` 2006-02-28 00:00:00 買入 : ['600050.XSHG', '600893.XSHG', '600016.XSHG', '600104.XSHG', '600010.XSHG', '600518.XSHG', '600030.XSHG', '600150.XSHG'] 2006-05-31 00:00:00 買入 : ['600036.XSHG', '600111.XSHG', '600104.XSHG', '600010.XSHG', '600030.XSHG'] 2006-08-30 00:00:00 買入 : ['600050.XSHG', '600893.XSHG', '600000.XSHG', '600104.XSHG', '600637.XSHG', '600837.XSHG', '600150.XSHG'] 2006-11-30 00:00:00 買入 : ['600050.XSHG', '600795.XSHG', '600036.XSHG', '600000.XSHG', '600111.XSHG', '600519.XSHG', '600016.XSHG', '600518.XSHG', '601988.XSHG', '600030.XSHG'] 2007-02-28 00:00:00 買入 : ['600000.XSHG', '600111.XSHG', '601006.XSHG', '600048.XSHG', '600015.XSHG', '600518.XSHG', '600887.XSHG', '600150.XSHG'] 2007-05-31 00:00:00 買入 : ['600795.XSHG', '600111.XSHG', '601166.XSHG', '600104.XSHG', '600015.XSHG', '600637.XSHG', '600837.XSHG'] 2007-08-30 00:00:00 買入 : ['600000.XSHG', '600519.XSHG', '601166.XSHG', '600015.XSHG', '600109.XSHG', '600887.XSHG', '601318.XSHG'] 2007-11-30 00:00:00 買入 : ['600050.XSHG', '600795.XSHG', '600111.XSHG', '601006.XSHG', '600048.XSHG', '600104.XSHG', '600015.XSHG', '600837.XSHG', '601988.XSHG', '600030.XSHG'] 2008-02-28 00:00:00 買入 : ['601328.XSHG', '600050.XSHG', '600795.XSHG', '600000.XSHG', '600018.XSHG', '600016.XSHG', '601006.XSHG', '600104.XSHG', '600028.XSHG', '600518.XSHG', '600837.XSHG', '601169.XSHG', '601988.XSHG', '601398.XSHG'] 2008-06-02 00:00:00 買入 : ['601006.XSHG', '601166.XSHG', '600010.XSHG', '600518.XSHG', '601318.XSHG'] 2008-09-01 00:00:00 買入 : ['601328.XSHG', '600050.XSHG', '601601.XSHG', '600036.XSHG', '600000.XSHG', '600519.XSHG', '600016.XSHG', '601998.XSHG', '600015.XSHG', '600637.XSHG', '600150.XSHG'] 2008-12-01 00:00:00 買入 : ['601601.XSHG', '600795.XSHG', '600104.XSHG', '600837.XSHG', '601169.XSHG', '600030.XSHG'] 2009-03-02 00:00:00 買入 : ['601601.XSHG', '601390.XSHG', '600104.XSHG', '600028.XSHG', '600518.XSHG', '600887.XSHG', '600837.XSHG', '601988.XSHG'] 2009-06-01 00:00:00 買入 : ['600893.XSHG', '600036.XSHG', '600111.XSHG', '600585.XSHG', '600048.XSHG', '600109.XSHG', '600887.XSHG', '601988.XSHG'] 2009-08-31 00:00:00 買入 : ['600050.XSHG', '600893.XSHG', '600000.XSHG', '600111.XSHG', '600519.XSHG', '600015.XSHG', '600010.XSHG', '600887.XSHG', '601766.XSHG', '601398.XSHG', '600150.XSHG'] 2009-11-30 00:00:00 買入 : ['600795.XSHG', '600893.XSHG', '600016.XSHG', '601006.XSHG', '600048.XSHG', '600887.XSHG', '601988.XSHG'] 2010-03-01 00:00:00 買入 : ['601601.XSHG', '600893.XSHG', '600018.XSHG', '600016.XSHG', '601668.XSHG', '600585.XSHG', '601998.XSHG', '600104.XSHG', '600028.XSHG', '601398.XSHG'] 2010-05-31 00:00:00 買入 : ['600111.XSHG', '600999.XSHG', '601628.XSHG', '601318.XSHG'] 2010-08-30 00:00:00 買入 : ['601328.XSHG', '600893.XSHG', '600111.XSHG', '600585.XSHG', '601998.XSHG', '601688.XSHG', '600999.XSHG', '600109.XSHG', '601989.XSHG', '600837.XSHG'] 2010-11-30 00:00:00 買入 : ['600010.XSHG', '601989.XSHG', '601169.XSHG', '600150.XSHG'] 2011-02-28 00:00:00 買入 : ['601601.XSHG', '601857.XSHG', '601390.XSHG', '601288.XSHG', '601668.XSHG', '601088.XSHG', '600999.XSHG', '601989.XSHG', '600837.XSHG'] 2011-05-31 00:00:00 買入 : ['600893.XSHG', '601668.XSHG', '601688.XSHG', '600010.XSHG', '600109.XSHG'] 2011-08-30 00:00:00 買入 : ['600010.XSHG', '600887.XSHG'] 2011-11-30 00:00:00 買入 : ['601288.XSHG', '601818.XSHG', '601766.XSHG'] 2012-02-28 00:00:00 買入 : ['600893.XSHG', '600015.XSHG', '600030.XSHG', '601669.XSHG', '601901.XSHG'] 2012-05-31 00:00:00 買入 : ['601336.XSHG', '601989.XSHG', '601669.XSHG'] 2012-08-30 00:00:00 買入 : ['601336.XSHG', '600837.XSHG', '601901.XSHG'] 2012-11-30 00:00:00 買入 : ['601668.XSHG', '601901.XSHG'] ```