3.4 熔断机制 • 股海拾贝之 [熔断错杀股]

来源：https://uqer.io/community/share/568e4c78228e5b18e0ba2962

新年伊始，本是普天同庆之时，A股门前却风声鹤唳；熔断机制推出4天，触发-7%熔断两次；千古跌停中，有不少标的遭遇恐慌性抛售，有人称之为熔断机制的磁石效应！

本文中，我们试图以简单的逻辑，来找到今天即7日有可能遭遇恐慌性抛售的熔断错杀股。

import pandas as pd
import numpy as np
from matplotlib import pylab
import matplotlib.pyplot as plt
import seaborn

1. 今日错杀股

from quartz.api import set_universe

univ = set_universe('A')
univ.remove('002778.XSHE')  # 删除2016-01-06上市新股

以最浅显地理解，错杀股票可能存在于以下情况中：

• 今日开盘涨势不错，到首次熔断时仍为红盘，首次熔断结束后杀跌；
• 今日开盘处于跌势，首次熔断之前跌的少，而首次熔断到二次熔断之间杀跌；
• 首次熔断结束后开始交易，先杀跌，然后有明显拉升迹象，使得二次熔断时价格高于首次熔断时价格

三种情况下的走势如下图所示：

# 沪深300指数今日走势
hs300PreClose = DataAPI.MktIdxdGet(tradeDate='20160106',ticker='399300').closeIndex.values[0]
hs300 = DataAPI.MktBarRTIntraDayGet(securityID='399300.XSHE',startTime='09:30',endTime='10:30')
hs300['hs300 index'] = hs300.closePrice/hs300PreClose - 1
hs300 = hs300[['barTime','hs300 index']]

# 第一种情况，例如‘山东黄金’
case1PrePrice = DataAPI.MktEqudAdjGet(tradeDate='20160106',ticker='600547').closePrice.values[0]
case1 = DataAPI.MktBarRTIntraDayGet(securityID='600547.XSHG',startTime='09:30',endTime='10:30')
case1['600547.XSHG'] = case1.closePrice/case1PrePrice - 1
case1 = case1[['barTime','600547.XSHG']]

# 第二种情况，例如‘红豆股份’
case2PrePrice = DataAPI.MktEqudAdjGet(tradeDate='20160106',ticker='600400').closePrice.values[0]
case2 = DataAPI.MktBarRTIntraDayGet(securityID='600400.XSHG',startTime='09:30',endTime='10:30')
case2['600400.XSHG'] = case2.closePrice/case2PrePrice - 1
case2 = case2[['barTime','600400.XSHG']]

# 第三种情况，例如‘天宝股份’
case3PrePrice = DataAPI.MktEqudAdjGet(tradeDate='20160106',ticker='002220').closePrice.values[0]
case3 = DataAPI.MktBarRTIntraDayGet(securityID='002220.XSHE',startTime='09:30',endTime='10:30')
case3['002220.XSHE'] = case3.closePrice/case3PrePrice - 1
case3 = case3[['barTime','002220.XSHE']]

for case in (case1,case2,case3):
    hs300 = pd.merge(hs300,case)

hs300 = hs300.set_index('barTime')
hs300.plot(figsize=(10,6))

<matplotlib.axes.AxesSubplot at 0x6e01210>

根据以上分析，我们拿取今日行情数据，对全A股中的股票做分析，得到表格如下，其中：

• preClose： 昨日收盘价
• 1stPrice： 首次熔断时价格
• 2ndPrice： 二次熔断时价格
• 1stCollapse： 首次熔断时股价跌幅
• 2ndCollapse： 首次熔断结束，之后股价继续下跌的幅度，即首次熔断后今日股价又下跌了这么多
• collapseRatio：首次熔断之后的跌幅和首次熔断之前的跌幅的比例
• lowBetCollapse：首次熔断和二次熔断之间的股价低点
• closeToLow： 二次熔断时的价格和lowBetCollapse的比例

cols = ['preClose','1stPrice','2ndPrice','1stCollapse','2ndCollapse','collapseRatio','lowBetCollapse','closeToLow']
collapse0107 = pd.DataFrame(0.0,index=univ,columns=cols)

# 股票前收盘价
collapse0107['preClose'] = DataAPI.MktEqudGet(tradeDate='20160106',secID=univ,field='secID,closePrice',pandas='1').set_index('secID')

# 股票在-5%和-7%熔断时候的价格
for stk in collapse0107.index:
    price = DataAPI.MktBarRTIntraDayGet(securityID=stk,startTime='09:56',endTime='10:03').closePrice.values
    collapse0107['1stPrice'][stk] = price[0]
    collapse0107['2ndPrice'][stk] = price[-1]
    collapse0107['lowBetCollapse'][stk] = np.min(price)

# 两次熔断前的股票跌幅
collapse0107['1stCollapse'] = 1-collapse0107['1stPrice']/collapse0107['preClose']
collapse0107['2ndCollapse'] = 1-collapse0107['2ndPrice']/collapse0107['preClose'] - collapse0107['1stCollapse']

# 二次熔断跌幅和一次熔断跌幅比
collapse0107['collapseRatio'] = collapse0107['2ndCollapse']/collapse0107['1stCollapse']

# 第二次熔断时价格与两次熔断之间的最低价的比值
collapse0107['closeToLow'] = collapse0107['2ndPrice']/collapse0107['lowBetCollapse']

collapse0107.sort(columns='collapseRatio',inplace=True)
collapse0107 = collapse0107[~np.isnan(collapse0107.collapseRatio)]

collapse0107.tail()

	preClose	1stPrice	2ndPrice	1stCollapse	2ndCollapse	collapseRatio	lowBetCollapse	closeToLow
603398.XSHG	123.01	121.00	110.71	0.016340	0.083652	5.119403	110.71	1
002750.XSHE	40.59	40.05	37.00	0.013304	0.075142	5.648148	37.00	1
600569.XSHG	3.58	3.55	3.31	0.008380	0.067039	8.000000	3.31	1
002768.XSHE	71.17	71.00	66.00	0.002389	0.070254	29.411765	66.00	1
600782.XSHG	5.81	5.80	5.31	0.001721	0.084337	49.000000	5.31	1

按照以上数据，综合前面的简单逻辑，我们利用以下条件来选择错杀股：

• collapseRatio > 1.5
• collapseRatio < -0.9
• closeToLow > 1.02

得到约40只股票如下

good = collapse0107[(collapse0107.collapseRatio>1.5) | (collapse0107.collapseRatio<-0.9) | (collapse0107.closeToLow>1.02)].index
good_stks = DataAPI.MktEqudGet(secID=good,tradeDate='20160107',field='secID,secShortName,tradeDate,preClosePrice,closePrice')
good_stks

	secID	secShortName	tradeDate	preClosePrice	closePrice
0	000726.XSHE	鲁泰A	2016-01-07	13.67	13.05
1	000766.XSHE	通化金马	2016-01-07	14.04	12.68
2	000838.XSHE	财信发展	2016-01-07	53.34	49.92
3	000856.XSHE	冀东装备	2016-01-07	12.91	12.00
4	000937.XSHE	冀中能源	2016-01-07	5.37	5.14
5	002155.XSHE	湖南黄金	2016-01-07	9.39	9.14
6	002220.XSHE	天宝股份	2016-01-07	16.46	15.15
7	002251.XSHE	步步高	2016-01-07	14.62	13.62
8	002283.XSHE	天润曲轴	2016-01-07	19.47	18.35
9	002291.XSHE	星期六	2016-01-07	14.24	13.03
10	002355.XSHE	兴民钢圈	2016-01-07	18.83	16.95
11	002444.XSHE	巨星科技	2016-01-07	19.44	17.85
12	002506.XSHE	协鑫集成	2016-01-07	9.85	9.66
13	002517.XSHE	泰亚股份	2016-01-07	57.98	52.93
14	002575.XSHE	群兴玩具	2016-01-07	17.67	16.87
15	002588.XSHE	史丹利	2016-01-07	31.15	28.13
16	002615.XSHE	哈尔斯	2016-01-07	29.16	26.64
17	002617.XSHE	露笑科技	2016-01-07	23.17	21.12
18	002621.XSHE	三垒股份	2016-01-07	20.91	20.00
19	002640.XSHE	跨境通	2016-01-07	31.68	31.42
20	002702.XSHE	海欣食品	2016-01-07	24.59	23.05
21	002750.XSHE	龙津药业	2016-01-07	40.59	37.96
22	002768.XSHE	国恩股份	2016-01-07	71.17	67.03
23	002779.XSHE	中坚科技	2016-01-07	91.20	82.62
24	300013.XSHE	新宁物流	2016-01-07	17.50	16.38
25	300148.XSHE	天舟文化	2016-01-07	23.43	21.48
26	300179.XSHE	四方达	2016-01-07	9.79	9.18
27	300320.XSHE	海达股份	2016-01-07	12.88	11.95
28	600005.XSHG	武钢股份	2016-01-07	3.66	3.30
29	600057.XSHG	象屿股份	2016-01-07	12.01	11.25
30	600262.XSHG	北方股份	2016-01-07	35.69	33.50
31	600265.XSHG	ST景谷	2016-01-07	29.81	28.36
32	600291.XSHG	西水股份	2016-01-07	26.11	24.16
33	600298.XSHG	安琪酵母	2016-01-07	30.30	29.47
34	600328.XSHG	兰太实业	2016-01-07	13.63	12.54
35	600395.XSHG	盘江股份	2016-01-07	9.10	8.19
36	600448.XSHG	华纺股份	2016-01-07	9.01	8.60
37	600547.XSHG	山东黄金	2016-01-07	21.89	21.10
38	600569.XSHG	安阳钢铁	2016-01-07	3.58	3.38
39	600671.XSHG	天目药业	2016-01-07	32.98	32.82
40	600732.XSHG	*ST新梅	2016-01-07	8.51	8.46
41	600782.XSHG	新钢股份	2016-01-07	5.81	5.54
42	600874.XSHG	创业环保	2016-01-07	9.63	9.67
43	601001.XSHG	大同煤业	2016-01-07	5.78	5.75
44	603398.XSHG	邦宝益智	2016-01-07	123.01	111.39

我们还想看一下上述股票在过去的05、06两个交易日的表现：

fig = plt.figure(figsize=(10,8))

ax = fig.add_subplot(211)
fullA = DataAPI.MktEqudAdjGet(secID=univ, beginDate='20160104', endDate='20160106', field='secID,tradeDate,closePrice', pandas='1')
fullA = pd.DataFrame(fullA.groupby('secID').last().closePrice/fullA.groupby('secID').first().closePrice - 1)
ax = pylab.hist(fullA.closePrice,bins=50,histtype='stepfilled',range=(-0.22,0.22))
pylab.xlabel("(01-05 to 01-06) 2 Days' Returns")
pylab.ylabel('Number of stocks')

ax = fig.add_subplot(212)
good_stk_data = DataAPI.MktEqudAdjGet(secID=good, beginDate='20160104', endDate='20160106', field='secID,tradeDate,closePrice', pandas='1')
good_stk_data = pd.DataFrame(good_stk_data.groupby('secID').last().closePrice/good_stk_data.groupby('secID').first().closePrice - 1)
ax = pylab.hist(good_stk_data.closePrice,bins=50,histtype='stepfilled',range=(-0.22,0.22))
pylab.xlabel("(01-05 to 01-06) 2 Days' Returns")
pylab.ylabel('Number of stocks')

<matplotlib.text.Text at 0x7335050>

图中，上图为全A股在过去的5、6日两个交易日收益表现分布；下图为我们选出来的今日错杀股在5、6两个交易日的收益表现分布；

明显地，我们看出选出来的股票在过去两天表现比较出色；当然，过去两天的表现好不代表它们今天被错杀

2. 上次熔断时的4日被错杀股

利用上节中的选股条件，我们选出来4日熔断错杀股，来验证我们的逻辑

cols = ['preClose','1stPrice','2ndPrice','1stCollapse','2ndCollapse','collapseRatio','lowBetCollapse','closeToLow']
collapse0104 = pd.DataFrame(0.0,index=univ,columns=cols)

# 股票前收盘价
collapse0104['preClose'] = DataAPI.MktEqudGet(tradeDate='20151231',secID=univ,field='secID,closePrice',pandas='1').set_index('secID')

# 股票在-5%和-7%熔断时候的价格
for stk in collapse0104.index:
    price = DataAPI.MktBarHistOneDayGet(securityID=stk,date='20160104',startTime='13:25',endTime='13:40').closePrice.values
    collapse0104['1stPrice'][stk] = price[0]
    collapse0104['2ndPrice'][stk] = price[-1]
    collapse0104['lowBetCollapse'][stk] = np.min(price)

# 两次熔断前的股票跌幅
collapse0104['1stCollapse'] = 1-collapse0104['1stPrice']/collapse0104['preClose']
collapse0104['2ndCollapse'] = 1-collapse0104['2ndPrice']/collapse0104['preClose'] - collapse0104['1stCollapse']

collapse0104['closeToLow'] = collapse0104['2ndPrice']/collapse0104['lowBetCollapse']

# 二次熔断跌幅和一次熔断跌幅比
collapse0104['collapseRatio'] = collapse0104['2ndCollapse']/collapse0104['1stCollapse']

collapse0104.sort(columns='collapseRatio',inplace=True)
collapse0104 = collapse0104[~np.isnan(collapse0104.collapseRatio)]

collapse0104.tail()

	preClose	1stPrice	2ndPrice	1stCollapse	2ndCollapse	collapseRatio	lowBetCollapse	closeToLow
600836.XSHG	31.31	31.28	30.37	0.000958	0.029064	30.333333	29.47	1.030540
600178.XSHG	11.23	11.23	10.60	0.000000	0.056100	inf	10.48	1.011450
600822.XSHG	15.01	15.01	14.28	0.000000	0.048634	inf	14.11	1.012048
002686.XSHE	17.62	17.62	16.90	0.000000	0.040863	inf	16.18	1.044499
002009.XSHE	20.75	20.75	19.71	0.000000	0.050120	inf	19.20	1.026563

按照之前的选股条件，我们选出来了4日熔断被错杀的股票如下：

good = collapse0104[(collapse0104.collapseRatio>1.5) | (collapse0104.collapseRatio<-0.9) | (collapse0104.closeToLow>1.02)].index
good_stks = DataAPI.MktEqudGet(secID=good,tradeDate='20160107',field='secID,secShortName,tradeDate,preClosePrice,closePrice')
good_stks

	secID	secShortName	tradeDate	preClosePrice	closePrice
0	000040.XSHE	宝安地产	2016-01-07	15.45	13.91
1	000048.XSHE	康达尔	2016-01-07	44.68	40.21
2	000517.XSHE	荣安地产	2016-01-07	6.20	5.58
3	000519.XSHE	江南红箭	2016-01-07	19.20	17.28
4	000520.XSHE	长航凤凰	2016-01-07	13.11	11.80
5	000547.XSHE	航天发展	2016-01-07	18.81	16.94
6	000552.XSHE	靖远煤电	2016-01-07	9.85	8.87
7	000597.XSHE	东北制药	2016-01-07	11.50	10.36
8	000667.XSHE	美好集团	2016-01-07	5.62	5.06
9	000708.XSHE	大冶特钢	2016-01-07	13.20	11.88
10	000709.XSHE	河北钢铁	2016-01-07	3.67	3.30
11	000723.XSHE	美锦能源	2016-01-07	14.19	12.77
12	000757.XSHE	浩物股份	2016-01-07	10.80	9.72
13	000767.XSHE	漳泽电力	2016-01-07	6.29	5.66
14	000795.XSHE	太原刚玉	2016-01-07	17.12	15.41
15	000801.XSHE	四川九洲	2016-01-07	30.26	27.24
16	000898.XSHE	鞍钢股份	2016-01-07	5.19	4.68
17	000932.XSHE	华菱钢铁	2016-01-07	3.85	4.14
18	000952.XSHE	广济药业	2016-01-07	21.11	19.00
19	000990.XSHE	诚志股份	2016-01-07	24.59	22.18
20	002009.XSHE	天奇股份	2016-01-07	22.85	20.57
21	002013.XSHE	中航机电	2016-01-07	24.02	21.62
22	002025.XSHE	航天电器	2016-01-07	25.42	22.89
23	002045.XSHE	国光电器	2016-01-07	18.10	16.29
24	002149.XSHE	西部材料	2016-01-07	29.70	26.73
25	002157.XSHE	正邦科技	2016-01-07	21.13	19.30
26	002179.XSHE	中航光电	2016-01-07	37.56	34.11
27	002191.XSHE	劲嘉股份	2016-01-07	16.71	15.04
28	002200.XSHE	云投生态	2016-01-07	27.99	25.19
29	002220.XSHE	天宝股份	2016-01-07	16.46	15.15
...	...	...	...	...	...
104	600655.XSHG	豫园商城	2016-01-07	15.44	14.20
105	600662.XSHG	强生控股	2016-01-07	17.13	15.42
106	600663.XSHG	陆家嘴	2016-01-07	50.94	45.93
107	600685.XSHG	中船防务	2016-01-07	41.41	37.27
108	600734.XSHG	实达集团	2016-01-07	24.90	22.41
109	600755.XSHG	厦门国贸	2016-01-07	8.24	7.43
110	600760.XSHG	中航黑豹	2016-01-07	14.60	13.17
111	600774.XSHG	汉商集团	2016-01-07	29.00	26.10
112	600822.XSHG	上海物贸	2016-01-07	15.97	14.37
113	600826.XSHG	兰生股份	2016-01-07	36.16	32.54
114	600833.XSHG	第一医药	2016-01-07	18.55	16.71
115	600834.XSHG	申通地铁	2016-01-07	20.30	18.27
116	600836.XSHG	界龙实业	2016-01-07	34.32	30.89
117	600841.XSHG	上柴股份	2016-01-07	18.28	16.45
118	600855.XSHG	航天长峰	2016-01-07	45.30	40.77
119	600860.XSHG	京城股份	2016-01-07	11.38	10.24
120	600868.XSHG	梅雁吉祥	2016-01-07	7.61	6.89
121	600879.XSHG	航天电子	2016-01-07	18.94	17.05
122	600893.XSHG	中航动力	2016-01-07	42.87	38.58
123	600965.XSHG	福成五丰	2016-01-07	14.85	13.66
124	601069.XSHG	西部黄金	2016-01-07	23.71	21.97
125	601233.XSHG	桐昆股份	2016-01-07	12.37	11.16
126	601636.XSHG	旗滨集团	2016-01-07	5.18	4.70
127	601700.XSHG	风范股份	2016-01-07	10.80	9.72
128	601888.XSHG	中国国旅	2016-01-07	56.49	51.97
129	601890.XSHG	亚星锚链	2016-01-07	13.43	12.09
130	601989.XSHG	中国重工	2016-01-07	9.40	8.48
131	601998.XSHG	中信银行	2016-01-07	6.84	6.39
132	603696.XSHG	安记食品	2016-01-07	64.11	57.70
133	603901.XSHG	永创智能	2016-01-07	40.84	36.76

134 rows × 5 columns

我们选出来的4日熔断被错杀的股票，在后面的5、6两日的表现究竟如何呢？请看下图

fig = plt.figure(figsize=(10,8))

ax = fig.add_subplot(211)
fullA = DataAPI.MktEqudAdjGet(secID=univ, beginDate='20160104', endDate='20160106', field='secID,tradeDate,closePrice', pandas='1')
fullA = pd.DataFrame(fullA.groupby('secID').last().closePrice/fullA.groupby('secID').first().closePrice - 1)
ax = pylab.hist(fullA.closePrice,bins=50,histtype='stepfilled',range=(-0.22,0.22))
pylab.xlabel("(01-05 to 01-06) 2 Days' Returns")
pylab.ylabel('Number of stocks')

ax = fig.add_subplot(212)
good_stk_data = DataAPI.MktEqudAdjGet(secID=good, beginDate='20160104', endDate='20160106', field='secID,tradeDate,closePrice', pandas='1')
good_stk_data = pd.DataFrame(good_stk_data.groupby('secID').last().closePrice/good_stk_data.groupby('secID').first().closePrice - 1)
ax = pylab.hist(good_stk_data.closePrice,bins=50,histtype='stepfilled',range=(-0.22,0.22))
pylab.xlabel("(01-05 to 01-06) 2 Days' Returns")
pylab.ylabel('Number of stocks')

<matplotlib.text.Text at 0x74df810>

图中，上图为全A股在过去的5、6日两个交易日收益表现分布；下图为我们选出来的4日熔断错杀股在5、6两个交易日的收益表现分布，可以看出选出的错杀股在过去两天均有10%左右的涨幅；

明显地，我们看出选出来的4日熔断错杀股在后面的两天表现出色

3. 结论

2节中对于4日熔断错杀股在5、6两个交易日的数据，似乎能够支持我们在第1节中的错杀股选股逻辑；对于1节中选出来的今天即7日错杀股，搬个板凳看后面走势究竟如何

PS：股市风险大，投资需谨慎；本文仅是研究之用，不构成任何荐股观点

---

我们一直在努力

apachecn/AiLearning

---