More about Learning Spread Trades
Thursday, February 24, 2011 at 2:32PMWritten by Norm Winer. Follow me on Twitter and StockTwits.
Norm trades spreads for Bigger Capital. Norm comments on Chris Cachat's post about spread trades follow:
We enjoyed your post on spread trading. It’s nice to hear we inspired you to learn more about spreads, and it’s always helpful to hear how other traders approach things.
Reading your post, I see you are grappling with some of the same issues we did, and still are, and we thought we’d share some of our thoughts and experiences with you (and by we, I mean @biggercapital, @biggercapitalnw, and @fledlingtrader, who collaborates with us on trades).
It looks like you are basing your calculations on a 30 day time period. We have found that, while that approach works well with some spreads such as GLD-GDX, it often pays to look at longer time periods as well. We are now experimenting with looking at Cointegration and Z scores (standard deviations from the mean) based on five years, two years, one year and sometimes six months of data. We have found that something that looks cheap over a 90 day period, or even a two year period, can look a little bit pricey over a five year period. Just look at where NFX-BRY was trading a week ago and where it’s trading today.
Regarding when to unwind or when to cut, we only recently started giving serious thought to this. Because we prefer to unwind quickly also, we have probably been cutting winners too early. We are now paying closer attention to whether a spread still looks cheap or expensive before we unwind, even if we have already made a good return.
Lastly, definitely look more closely at half-life, especially if you’d rather not hold positions for long periods. Good luck and let us know if you discover anything interesting!
P.S. Micheal's book How Traders Achieve Creative Flow is now available on Amazon.com. If you want to use your creativity to innovate your trading edge, this book is for you.
Reader Comments (3)
Stat arb is a tough trade to do with one-off trades. In other words pairs trading is about taking lots and lots of small bets with high probability but an inverse risk to reward. Or in other words your avg loss is greater then your average gain. You can work to increase your profitability by focusing on the initial probability (cointegration, fundamental analysis, volatility analysis, etc) but in the long run I feel the odds and payout will revert back to the mean.
Pairs are also tough because there is absolutely no reason they can't deviate significantly from the mean. I've seen related securities (common & warrants) deviate enough to blow traders out of trade. Ultimately they're proven right but can't stick around long enough to profit from it. Try taking just any 2 securities and there's even less reason for these to revert. That being said, there are a handful of pairs I tend to "like" due to their long term reversion. GLD vs GDX is a good example.
This is not meant to discourage anyone from trying to design a better mousetrap. By all means, if you discover something you can profit handsomely. However its my personal opinion that execution, portfolio management, and position sizing is where the money is.
One more thought regarding unwinding of trades, a suggestion is to back-test various schemes over a portfolio of pairs and do something mechanical. Then it becomes just a matter of following the rules. Let the long term odds play in your favor. My humble .02 cents!
Thanks Largecaptrader.
Very informative comments. I totally agree with your points. Keep them small and many is the way to go. IMO.
Thanks largecaptrader. We've also had some success with the GLD-GDX pair. As for stat arb we are still feeling our way around that one. As you said we are trying to come up with some rules (for entering and exiting trades) that we can apply to lots of small trades.