Over the past 6 months my favorite pair NUE/X has sucked. What are some of the pairs out there that people trade.
Why does it suck now? Did it stop reverting? Did it stop moving at all? or did 1 of the components started to behave differently-a specialist change perhaps?
they quit moving together in a recognizable pattern. As did most of the steels as they started hitting highs.
I don't trade pairs anymore, but when I did this website was a good resource for correlated stocks - www.impactopia.com Click on the link under USA Equities and you can enter a symbol and find what stocks may trade well with it.
Just out of curiosity, since I have never pairs traded, what correlation coefficient would you guys consider appropriate for "correlated pairs" and what divergence percentage or other criteria do you use for determining setups?
Interesting link too. I'm curioius what data they ran their correlation on though. I ran some on CHE versus CHTT based off based off 2 year EOD and trailing 12 months EOD on the adjusted close, and I got a different coefficient than they did, but I didn't see anywhere what data they use. I got -.33 for last 12 months. who knows, maybe im doing it wrong. anybody else?
I feel like I should respond since I posted the link, but I can't help you on how they go about their calculations. Basically I just used the site to find names to run through the model where I was working.
How do you guys pair trade? For me it was buy one when the other went up. I didn't do too good at it.
I would short one, then buy the other. My theory was that similar assets should be priced similarily. Worked great for a while, lately, not as great as it use to be.
The pairs trading group I worked with for a short time would look at similar stocks and the price ratio of them over the last few years. Buy / sells would be placed based on this ratio reaching extremes. The hold time was for weeks so it was more investing than trading. I've tried applying (through testing) the same system to much shorter time frames with no success.