Introduction to Auction Theory and Auction Failure
As per the market auction theory, auction continues in one direction till the winner of the auction is decided. However, if auction fails to continue in one direction and re-starts on other direction abruptly, we term it as auction failure.
In the market profile theory, it is stated that the initial trading period of the session decides the direction for the day’s trade. You might have seen that the market in the opening first bar experiences gap ups and gap downs. Market in the first 15 mins is mostly clueless about the direction and may move in either direction to find buyer and sellers to support the trade. After about an hour, market sets its own direction and generally as per the auction theory the market should continue in this direction. If however, the market changes its direction, then it can be termed as market failure.
To make it more objective, consider the following rules :
Graphical Representation of failed auction
In the above example, failure happened at the lowest point of session-1 and the reaction was obtained in the session-3.
The above is another example, here auction failure occurred at the lowest point of session-1 and retest happened on the session three , but from the below. Total reaction noted was about 2.5%.
The lowest point of session -1 is outside the low of initial 1 hour price range (15 min of 4 bar = 1 hour). However, the close of the session is inside the initial range setup in the first hour. Thus the lowest point can be considered as auction failure. It may be noted that the reaction line has been drawn below the auction failure point at a distance of about 0.3%. This is because at this point of auction failure, we expect that there must be some more buyers/sellers below this point and the re-test will be done at this price level.
Generally the reversal is of the range of 0.4% to 1.5%. It is suggested that the full or partial profit should be booked at 0.4% profit.
The python code - Auction Failure Screening
|
import datetime as datetime import pandas as pd from dateutil import parser import os def readData02(csvFileName): df = pd.read_csv(csvFileName,names=['Script','DateData','timeData','Open','High','Low','Close','Volume','OI']) df['Datetime'] = pd.to_datetime(df['DateData'].astype(str)+' '+df['timeData'].astype(str),format='%Y%m%d %H:%M') d = {'Open': 'first', 'High': 'max', 'Low': 'min', 'Close': 'last', 'Volume': 'sum'} df = df.resample('5T', on='Datetime').agg(d) df = df[df['Open'].notna()] df = df[df['Volume'].notna()] df1 = df.reset_index() df1.to_csv("resultHDFC.csv") return df1 # return five minutes df l=os.listdir("..//..//Google Drive//oneminutedata//2022//FEB/22FEB//22FEB") print(l) l.remove("desktop.ini") for item in l: item1 = "..//..//Google Drive//oneminutedata//2022//FEB/22FEB//22FEB//" + item df = readData02(item1) fromDate = '2022-02-22' toDate = '2022-02-22' myDate0 = parser.parse(fromDate) myDate1 = parser.parse(toDate) + datetime.timedelta(days=1) i=0 while(myDate0 + datetime.timedelta(days=i) < myDate1): df1 = df[df['Datetime'] < myDate0 + datetime.timedelta(days=i+1)][df[df['Datetime'] < myDate1]['Datetime'] > myDate0 + datetime.timedelta(days=i)] # print(df1) df2 = df1[df1['Datetime'] < myDate0 + datetime.timedelta(days=i, hours=10,minutes=20)] myDate = myDate0 + datetime.timedelta(days=i) ADHigh = df2['High'].max() ADLow = df2['Low'].min() df3 = df1[df1['Datetime'] < myDate0 + datetime.timedelta(days=i, hours=15,minutes=35)] df4 = df3[df3['Datetime'] > myDate0 + datetime.timedelta(days=i, hours=15,minutes=25)].reset_index(drop=True) myClose = df4['Close'].min() df5 = df1[df1['Datetime'] > myDate0 + datetime.timedelta(days=i, hours=10,minutes=15)] myHigh = df5['High'].max() myLow = df5['Low'].min() if myLow < ADLow and myClose > ADLow: print(item,myDate, myLow,"Auction Failure at Low") if myHigh > ADHigh and myClose < ADHigh: print(item,myDate, myHigh,"Auction Failure at High") i = i + 1 |
The format of input File
Mine file was not having any headers, it was a 1 min data. Code for reading the CSV file converts the 1 min data into 5 min data. Further, I have daily files in a single folder, so I have used OS module to list the files. desktop.ini is already there is each folder, so to remove that file , I have pop it out.
The data looks like as under :
DLF,20220222,09:08,331.00,331.00,331.00,331.00,29341,0
DLF,20220222,09:16,332.95,333.95,331.60,332.35,209224,0
wherein first col is Scrip name
Second is date
Third is time
fourth - fifth -sixth -seventh are open/high/low /close
eight is volume
and the last one is IO, which is zero for equity
