Pair Trading Strategy Journal

Closed one trade today;

Sold COL @ 31.78
Covered ITW @ 30.84

Opened 1 new trade;

Long STM @ 6.10
Short MXIM @ 11.81

 

Ive been away for the last week so time to update the journal

Closed 3 trades since last posting;

Sold APA @ 65.50
Covered EOG @ 68.79

Sold AZN @ 37.34
Covered DGX @ 43.57

Sold STM @ 6.28
Covered MXIM @ 11.03

Opened 2 new trades Friday;

Long APA @ 62.61
Short OXY @ 46.19

Long AGU @ 24.45
Short CF @ 47.05

 

Exited one trade today;

Sold AGU @ 30.59
Covered CF @ 52.30

 

Hello Jonny
If it is not too much trouble, could you add the % P&L on the trade and the Annual % P&L. (not the $ amount that you made).

I wish you continued success.

Gabe

 

Jonny,

Thanks for taking the time to maintain this thread.


I see that you gave PairTrade Finder a good review in the ET products and services section.


Here is my question.

Let's take a spread like XOM/CVX , where the 3 year correlation and the 1 year correlation are both high (3 yr: .96 / 1yr: .86 - based on spdrindex.com correlations ) and they are obviously in the same business with comparable macro issues.

When PairTrade Finder does a backtest - how profitable is this pair?


Now, let's compare it to GE/PFE, where the 3 year correlation is decent (.75) and the 1 year correlation is (.89). They are not, however, in the same business.

When PairTrade Finder does its backtest - how profitable is this pair?

Just interested in seeing if the buy/sell signal methodology is clever enough to exploit a relatively high correlation for short term trades regardless of longer-term macro exposures.

 

Hey jonny, I did a little research on my own about pairs trading and ended up finding a bunch of references to "cointegration". Earlier in this thread someone else brought this up as well.

Do you use cointegration for choosing pairs, or something else (i.e. correlation)?

 

Strangely enough over the last year GE/PFE has made more than XOM/CVX, although I don't trade either pair.

 

Yes this has been discussed at length, I don't use cointegration I use correlation and only trade pairs over 80% correlation, mostly over 90%, some use cointegration and thats fine, whatever works for the trader I guess.

 

Thanks for looking into that, Jonny. Interesting outcome. Strangely enough ... as you said.


If you don't mind, what was the P/L for the two spreads and over what time period?

 

Exited one trade today;

Sold BIDU @ 132.15
Covered GOOG @ 324.78