In this algorithm I use a pairs trading strategy with Chevron and Exxon. Whenever they diverge I short sell the over-performer and go long on the under-performer. Once they converge I sell everything. I'm not getting great results at the moment. Is it that I didn't choose a good pair? Did I mess up my negative signs and accidentally short sell the under-performer? All constructive feedback is welcome.
Even assuming your algorithm is absolutely perfect, I'd be surprised if you find any significant returns with this kind of strategy. Simply because if there were significant returns with such a simple strategy, people would capitalize on this correlation and it would disappear. I think that's what you are seeing here-- on these timescales (days/minutes) this pair is weakly correlated.
i think your ratio calculation is a bit strange, for example during the week of nov 1, 2010 your ratio is lower than your lowerbound. and the week of dec 13th 2010 it's higher than your highbound, yet the spread is aproximatly the same.
it seems like you should be doing your calculation based on the spread instead. for example, entering the position when the spread is in a high percentile and exiting when the spread is in a low percentile.
a rough test strat of that here.... but i might have my math wrong, so double-check.
EDIT: don't use this backtest, there's a bug in that i'm using the wrong side of the numpy array. use the example from my last post instead.
sorry my previous example is wrong because i was using the wrong side of the numpy array (i thought new items were added in front)
here's a fix of my example. i see that in the last year there seems to be a regime change thus performance degrades
Hey Jordan. You're calculating your ratios as A / B, where A and B are prices. The 'ratio' that you really want should be solved by using a simple linear regression, of the form Y = bX, where Y and X are your two price histories.
maybe you could post an example? i'm a little weak on my math so seeing code for doing either linear regression or cointegration to execute this strategy would be appreciated.
I think he means that Sum(Y)/Sum(X) = b is more useful (and simpler) than solving for b in Y = bX with some sort of pseudoinverse best fit
I OTOH know the math, but am new to the site, and haven't even looked at the prog lang. Pretty sure you can find code samples on google. Pair (or group) trading is about finding a mean-reverting linear combination of securities. That is the essence of cointegration. Has been done since the 90s, so I wouldn't put money on blue chip stocks like those. Once you go down on cap, you'll have to be more attentive to liquidity obviously. Then the plot thickens ;)
Please sign in or
join Quantopian to
post a reply.
Need help?
Contact support.
Company
Learn & Support
Notices
Follow Us
The material on this website is provided for informational purposes only and does not constitute an offer to sell, a solicitation to buy, or a recommendation or endorsement for any security or strategy, nor does it constitute an offer to provide investment advisory services by Quantopian.
In addition, the material offers no opinion with respect to the suitability of any security or specific investment. No information contained herein should be regarded as a suggestion to engage in or refrain from any investment-related course of action as none of Quantopian nor any of its affiliates is undertaking to provide investment advice, act as an adviser to any plan or entity subject to the Employee Retirement Income Security Act of 1974, as amended, individual retirement account or individual retirement annuity, or give advice in a fiduciary capacity with respect to the materials presented herein. If you are an individual retirement or other investor, contact your financial advisor or other fiduciary unrelated to Quantopian about whether any given investment idea, strategy, product or service described herein may be appropriate for your circumstances. All investments involve risk, including loss of principal. Quantopian makes no guarantees as to the accuracy or completeness of the views expressed in the website. The views are subject to change, and may have become unreliable for various reasons, including changes in market conditions or economic circumstances.