Pair Trading - Exploring The Low Risk Statistical Arbitrage Trading Concepts

ncube

Well-Known Member
#71
Yes something wrong done in excell I replaced it with your new file...now seem ok but still p value is different

View attachment 26981
Ok now the issue is consistent for you, @20.sum1 & @cloudTrader , you guys are getting extremely low pValue, need to analyze if statmodel library version is correct, can you post the output for the following commands:
import statsmodels.api as sm
sm.version.version
 

VJAY

Well-Known Member
#72
Ok now the issue is consistent for you, @20.sum1 & @cloudTrader , you guys are getting extremely low pValue, need to analyze if statmodel library version is correct, can you post the output for the following commands:
import statsmodels.api as sm
sm.version.version
please give the command only ..I not getting from above post
 

ncube

Well-Known Member
#73
Yes something wrong done in excell I replaced it with your new file...now seem ok but still p value is different

View attachment 26981
Ok now the issue is consistent for you, @20.sum1 & @cloudTrader , you guys are getting extremely low pValue, need to analyze if statmodel library version is correct, can you post the output for the following commands:
import statsmodels.api as sm
sm.version.version
please give the command only ..I not getting from above post
1532959343648.png
 

VJAY

Well-Known Member
#74
Ok now the issue is consistent for you, @20.sum1 & @cloudTrader , you guys are getting extremely low pValue, need to analyze if statmodel library version is correct, can you post the output for the following commands:
import statsmodels.api as sm
sm.version.version


View attachment 26983
1532959561173.png

looks like it different with you
 

ncube

Well-Known Member
#75
These are few pairs I was tracking some time back, I will update this list in the evening with few more pairs based on my analysis:

('HCLTECH', 'INFY')
('HCLTECH', 'TCS')
('HCLTECH', 'TECHM')
('HCLTECH', 'WIPRO')
('INFY', 'TCS')
('INFY', 'TECHM')
('INFY', 'WIPRO')
('TCS', 'TECHM')
('TCS', 'WIPRO')
('TECHM', 'WIPRO')
('ALBK', 'ANDHRABANK')
('ALBK', 'ORIENTBANK')
('ALBK', 'SYNDIBANK')
('ANDHRABANK', 'ORIENTBANK')
('ANDHRABANK', 'SYNDIBANK')
('ORIENTBANK', 'SYNDIBANK')
('TATAMTRDVR', 'TATAMOTORS')
('RELCAPITAL', 'RELINFRA')
('PFC', 'RECLTD')
('CIPLA', 'GLENMARK')
('BANKINDIA', 'CANBK')
('APLLTD', 'RAJESHEXPO')
Pairs I am currently tracking:

('ALBK', 'ANDHRABANK')
('ALBK', 'BANKBARODA')
('ALBK', 'BANKINDIA')
('ALBK', 'CANBK')
('ALBK', 'ORIENTBANK')
('ALBK', 'SBIN')
('ALBK', 'UNIONBANK')
('ANDHRABANK', 'BANKBARODA')
('ANDHRABANK', 'BANKINDIA')
('ANDHRABANK', 'CANBK')
('ANDHRABANK', 'ORIENTBANK')
('ANDHRABANK', 'SBIN')
('ANDHRABANK', 'SYNDIBANK')
('ANDHRABANK', 'UNIONBANK')
('BANKINDIA', 'CANBK')
('BANKINDIA', 'SBIN')
('BANKINDIA', 'SYNDIBANK')
('BANKINDIA', 'UNIONBANK')
('CANBK', 'ORIENTBANK')
('CANBK', 'SBIN')
('CANBK', 'SYNDIBANK')
('CANBK', 'UNIONBANK')
('ORIENTBANK', 'SBIN')
('ORIENTBANK', 'SYNDIBANK')
('ORIENTBANK', 'UNIONBANK')
('SBIN', 'SYNDIBANK')
('SBIN', 'UNIONBANK')
('SYNDIBANK', 'UNIONBANK')
('APOLLOTYRE', 'CEATLTD')
('APOLLOTYRE', 'MRF')
('APOLLOTYRE', 'TVSSRICHAK')
('MRF', 'TVSSRICHAK')
('RELCAPITAL', 'RELINFRA')
('PFC', 'RECLTD')
('IOB', 'UCOBANK')
('IDFCBANK', 'IDFC')
('DHFL', 'IBULHSGFIN')
('CIPLA', 'LUPIN')
('AUROPHARMA', 'GLENMARK')
 

VJAY

Well-Known Member
#77
No I meant can you run those 2 commands in a new cell and show me the output. I want to check if the statsmodel version that you have installed is correct.
commands not working :(
 

VJAY

Well-Known Member
#78
Untitled.png


I think its 0.9.0
 

ncube

Well-Known Member
#79
Yes, thats correct, the latest version for statsmodels package now is 0.9.0, I was using 0.8.0 and on upgrading to 0.9.0 I am also getting the same pValue as you guys.

Checked the 0.9.0 documentation and found that they have changed the default value for autolag = 'aic", this need to be changed to autolag = None.

Please add a parameter to the coint function as shown in the below image, it will work fine.

pvalue = coint(S1,S2,autolag=None)[1]

1532961838099.png


Or you can download the file from my google drive, I have uploaded the latest file there.
 

VJAY

Well-Known Member
#80
Yes, thats correct, the latest version for statsmodels package now is 0.9.0, I was using 0.8.0 and on upgrading to 0.9.0 I am also getting the same pValue as you guys.

Checked the 0.9.0 documentation and found that they have changed the default value for autolag = 'aic", this need to be changed to autolag = None.

Please add a parameter to the coint function as shown in the below image, it will work fine.

pvalue = coint(S1,S2,autolag=None)[1]

View attachment 26987

Or you can download the file from my google drive, I have uploaded the latest file there.
Ok now i think its okey ...is this pvalue is same with yours?
Untitled.png



Why here your last check point 11 hours ago showung?

Untitled1.png