I did a linear regression on a TON of stocks to find the r value, and found many stocks correlated. Most of them are in the same industry. However sometimes I get BGC and DO, a cable company, and a diamond drilling company that are correlated. Could this be true that they trade the same way? or was this just a mear coincidence?
I'm not sure how far back your study goes, but lots of companies to well together in a bull market. Finding correlation is only one part of pairs trading, you still have to do a lot of work to find suitable pairs after that.
It's common to find pairs that are correlated for no reason. If you search enough pairs your guaranteed to find some pairs that were correlated in the past simply by coincidence.
Two words: "Spurious Correlation" I can't be arsed explaining this so look it up - here is a start: http://www.gummy-stuff.org/perfect-correlation.htm http://www.gummy-stuff.org/cointegration.htm Be careful with either "tool" and have "fun" with Stab Art (sic).
Simple correlation is not enough to evaluate a good pair. We go very deeply into fundamentals for each stock, being extremely sure that we have the closest possible sub-industry groups. There is a lot of daily updating and research from our pairs guys who have developed super databases of hundreds of pairs at various ratio's...the work never stops, well worth it, but it never stops. P.S. edit: check your PM, please. Don
Don, Not sure if this is part of your Bright's family trading secret, but I was wondering if you could tell me a couple of the fundamental factors used to select your pairs Thanks
We have about 30 columns in our spreadsheet that covers everything from price to book to number of employees. We do have several internal ranking ratio's as well. Most of the info can be found under "key statistics" on yahoo finance. Don