from jqdata import *
import jqdata
import numpy as np
import pandas as pd
import datetime
#先加载画图需要用的包
import matplotlib as mpl
import matplotlib.pyplot as plt
mpl.style.use('ggplot')
import seaborn as sns
打板追涨研究¶
获取数据的函数¶
def get_data(stock_list, date1,date2): #date1为当前日期 date2 为X天前
close = get_price(stock_list, end_date=date1,
frequency='daily', fields='close', count=1)['close']
close_pre = get_price(stock_list, end_date=date2 ,
frequency='daily', fields='close', count=1)['close']
#price_now.columns = ['price']
#price_20days = get_price(stock_list, end_date=date2, frequency='daily', fields='close', count=66)['close'].T
#price_20days.columns = ['price']
high = get_price(stock_list, end_date=date1,
frequency='daily', fields='high', count=1)['high']
close = close.T
close.columns = ['price']
close_pre = close_pre.T
close_pre.columns = ['price']
high = high.T
high.columns = ['price']
close_pct = close/close_pre - 1
high_pct = high/close_pre-1
close_pct.columns = ['Close_price_change']
high_pct.columns = ['High_price_change']
Price = pd.concat([close_pct,high_pct],axis = 1)
Price = Price.dropna()
Price = Price[Price['High_price_change']>=0.08]
return Price
获取深次新股成分股800天的当天最高涨幅大于8%的数据¶
看看获取的数据长什么样
stock_list1 = get_index_stocks('399678.XSHE')
today = datetime.datetime.now()
StartDate = today-datetime.timedelta(800) #800天前
EndDate = today - datetime.timedelta(1) #一天前
trading_dates = jqdata.get_trade_days(start_date=StartDate, end_date=EndDate)
stock_data = get_data(stock_list1,trading_dates[-1]
,trading_dates[-2])
stock_data.head(10)
for i in range(2,len(trading_dates)-2):
temp_data = get_data(stock_list1,trading_dates[-i],trading_dates[-(i+1)])
stock_data = stock_data.append(temp_data)
a = len(stock_data)
b = len(stock_data[stock_data['Close_price_change']>=0.099])
print(stock_data.head(20))
print('涨停板占比%0.3f'%(b/a))
plt.hist(np.array(stock_data['Close_price_change']),color='r')
plt.title('Close_price_change')
plt.show()
获取数据改进版函数¶
考虑了开盘直接涨停买不进去，限制开盘涨幅小于8%\ 即开盘涨小于8%，当日最高价大于8%
#改进版考虑了开盘直接涨停买不进去，限制开盘涨幅小于8%
def get_data2(stock_list, date1,date2): #date1为当前日期 date2 为X天前
close = get_price(stock_list, end_date=date1,
frequency='daily', fields='close', count=1)['close']
close_pre = get_price(stock_list, end_date=date2 ,
frequency='daily', fields='close', count=1)['close']
#price_now.columns = ['price']
#price_20days = get_price(stock_list, end_date=date2, frequency='daily', fields='close', count=66)['close'].T
#price_20days.columns = ['price']
high = get_price(stock_list, end_date=date1,
frequency='daily', fields='high', count=1)['high']
open_price = get_price(stock_list, end_date=date1,
frequency='daily', fields='open', count=1)['open']
open_price = open_price.T
open_price.columns = ['price']
close = close.T
close.columns = ['price']
close_pre = close_pre.T
close_pre.columns = ['price']
high = high.T
high.columns = ['price']
close_pct = close/close_pre - 1
high_pct = high/close_pre-1
open_pct = open_price/close_pre - 1
close_pct.columns = ['Close_price_change']
high_pct.columns = ['High_price_change']
open_pct.columns = ['Open_price_change']
Price = pd.concat([close_pct,high_pct,open_pct],axis = 1)
Price = Price.dropna()
Price = Price[Price['High_price_change']>=0.08]
Price = Price[Price['Open_price_change']<0.08]
return Price
重新获取数据来测试，发现收盘涨停板比例小了很多，但占比还是很高的
stock_data = get_data2(stock_list1,trading_dates[-1]
,trading_dates[-2])
for i in range(2,len(trading_dates)-2):
temp_data = get_data2(stock_list1,trading_dates[-i],trading_dates[-(i+1)])
stock_data = stock_data.append(temp_data)
a = len(stock_data)
b = len(stock_data[stock_data['Close_price_change']>=0.099])
print(stock_data.head(20))
print('涨停板占比%0.3f'%(b/a))
plt.hist(np.array(stock_data['Close_price_change']),color='r')
plt.title('Close_price_change')
plt.show()
¶
stock_data = get_data2(stock_list1,trading_dates[-1]
,trading_dates[-2])
stock_data
改进版考虑了开盘直接涨停买不进去，限制开盘涨幅小于9%¶
#改进版考虑了开盘直接涨停买不进去，限制开盘涨幅小于9%
def get_data3(stock_list, date1,date2): #date1为当前日期 date2 为X天前
close = get_price(stock_list, end_date=date1,
frequency='daily', fields='close', count=1)['close']
close_pre = get_price(stock_list, end_date=date2 ,
frequency='daily', fields='close', count=1)['close']
#price_now.columns = ['price']
#price_20days = get_price(stock_list, end_date=date2, frequency='daily', fields='close', count=66)['close'].T
#price_20days.columns = ['price']
high = get_price(stock_list, end_date=date1,
frequency='daily', fields='high', count=1)['high']
open_price = get_price(stock_list, end_date=date1,
frequency='daily', fields='open', count=1)['open']
open_price = open_price.T
open_price.columns = ['price']
close = close.T
close.columns = ['price']
close_pre = close_pre.T
close_pre.columns = ['price']
high = high.T
high.columns = ['price']
close_pct = close/close_pre - 1
high_pct = high/close_pre-1
open_pct = open_price/close_pre - 1
close_pct.columns = ['Close_price_change']
high_pct.columns = ['High_price_change']
open_pct.columns = ['Open_price_change']
Price = pd.concat([close_pct,high_pct,open_pct],axis = 1)
Price = Price.dropna()
Price = Price[Price['High_price_change']>=0.09]
Price = Price[Price['Open_price_change']<0.09]
return Price
最高价大于9%，开盘小于9%¶
收盘涨停的比例大于0.42，即在9%的时候追有42%的概率当天封板
stock_list1 = get_index_stocks('000002.XSHG')
stock_data = get_data3(stock_list1,trading_dates[-1]
,trading_dates[-2])
for i in range(2,len(trading_dates)-2):
temp_data = get_data3(stock_list1,trading_dates[-i],trading_dates[-(i+1)])
stock_data = stock_data.append(temp_data)
a = len(stock_data)
b = len(stock_data[stock_data['Close_price_change']>=0.099])
print(stock_data.head(20))
print('涨停板占比%0.3f'%(b/a))
plt.hist(np.array(stock_data['Close_price_change']),color='r')
plt.title('Close_price_change')
plt.show()
stock_list1 = get_index_stocks('000002.XSHG')
stock_data = get_data2(stock_list1,trading_dates[-1]
,trading_dates[-2])
for i in range(2,len(trading_dates)-2):
temp_data = get_data2(stock_list1,trading_dates[-i],trading_dates[-(i+1)])
stock_data = stock_data.append(temp_data)
a = len(stock_data)
b = len(stock_data[stock_data['Close_price_change']>=0.099])
print(stock_data.head(20))
print('涨停板占比%0.3f'%(b/a))
plt.hist(np.array(stock_data['Close_price_change']),bins = 20,color='r')
plt.title('Close_price_change')
plt.show()
plt.hist(np.array(stock_data['Close_price_change']),bins = 40,color='r')
plt.title('Close_price_change')
plt.show()
完善函数，计算股价在超过9%瞬间买入，第二天收盘卖出的收益率earning¶
#改进版考虑了开盘直接涨停买不进去，限制开盘涨幅小于9%
def get_data4(stock_list, date1,date2,date3): #date1为当前日期 date2 为1天前 date3为后一天
close = get_price(stock_list, end_date=date1,
frequency='daily', fields='close', count=1)['close']
close_pre = get_price(stock_list, end_date=date2 ,
frequency='daily', fields='close', count=1)['close']
close_after = get_price(stock_list, end_date=date3 ,
frequency='daily', fields='high', count=1)['high']
#price_now.columns = ['price']
#price_20days = get_price(stock_list, end_date=date2, frequency='daily', fields='close', count=66)['close'].T
#price_20days.columns = ['price']
high = get_price(stock_list, end_date=date1,
frequency='daily', fields='high', count=1)['high']
open_price = get_price(stock_list, end_date=date1,
frequency='daily', fields='open', count=1)['open']
open_price = open_price.T
open_price.columns = ['price']
close = close.T
close.columns = ['price']
close_pre = close_pre.T
close_pre.columns = ['price']
close_after = close_after.T
close_after.columns = ['price']
high = high.T
high.columns = ['price']
close_pct = close/close_pre - 1
high_pct = high/close_pre-1
open_pct = open_price/close_pre - 1
earning_next_day = close_after/(1.09 * close_pre) -1 #按照头一天涨幅9%的时候买入来算earning
close_pct.columns = ['Close_price_change']
high_pct.columns = ['High_price_change']
open_pct.columns = ['Open_price_change']
earning_next_day.columns = ['Earning']
Price = pd.concat([close_pct,high_pct,open_pct,earning_next_day],axis = 1)
Price = Price.dropna()
Price = Price[Price['High_price_change']>=0.09]
Price = Price[Price['Open_price_change']<0.09]
return Price
滚动获取数据，并打印出数据来观察
¶
计算满足条件买入后第二日收盘价卖出收益率
stock_list1 = get_index_stocks('399678.XSHE')
stock_data = get_data4(stock_list1,trading_dates[-2],trading_dates[-3],trading_dates[-1])
for i in range(2,len(trading_dates)-3):
temp_data = get_data4(stock_list1,trading_dates[-(i+1)],trading_dates[-(i+2)],trading_dates[-i])
stock_data = stock_data.append(temp_data)
a = len(stock_data)
b = len(stock_data[stock_data['Close_price_change']>=0.099])
print(stock_data.head(20))
print('涨停板占比%0.3f'%(b/a))
zhuanqian = len(stock_data[stock_data['Earning']>0])
print('赚钱占比%0.3f'%(zhuanqian/a))
plt.hist(np.array(stock_data['Earning']),bins = 20,color='r')
plt.title('Earning_next_day')
plt.show()
计算出最终盈利的占比为65.1%¶
len(stock_data[stock_data['Earning']>0])
计算此策略平均每次盈利¶
stock_data['Earning'].mean()
证明此追涨打板策略是有一定效果的，¶
$ t-statistics =（μ-0)/σ $
可以用t检验证明收益率是显著的¶
