Rolling Correlations

Discussions about the testing and simulation of mechanical trading systems using historical data and other methods. Trading Blox Customers should post Trading Blox specific questions in the Customer Support forum.
Post Reply
sunyata
Roundtable Fellow
Roundtable Fellow
Posts: 94
Joined: Sat Jun 06, 2009 4:27 pm
Location: Charlottesville, VA

Rolling Correlations

Post by sunyata » Wed Oct 03, 2012 4:40 pm

Have been researching correlations and how they change over time. Took soybean and soybean oil price data from 2000 - present and calculated rolling correlations for 40 days, 120 days, and 1 year. The resultant graphs are attached. The 40-day correlations are very erratic as you see and longer calculation periods smooth the results, just as with a moving average. Given the small sample size, the margin of error on the 40-day correlations is plus-minus 13%, although they are statistically significant.

I am newbie so I do not know at this point what calculation period is most useful for trading. I thought I'd share in case anyone was interested and to generate dialogue.
Attachments
40 day.JPG
40 day.JPG (90.43 KiB) Viewed 2101 times
120 day.JPG
120 day.JPG (72.86 KiB) Viewed 2103 times
1yr.JPG
1yr.JPG (66.05 KiB) Viewed 2102 times

LeviF
Roundtable Knight
Roundtable Knight
Posts: 1428
Joined: Mon Dec 22, 2003 12:24 pm
Location: Des Moines, IA
Contact:

Post by LeviF » Wed Oct 03, 2012 6:43 pm

Looking at correlations on a weekly basis may help to smooth out some of your spikes.

sunyata
Roundtable Fellow
Roundtable Fellow
Posts: 94
Joined: Sat Jun 06, 2009 4:27 pm
Location: Charlottesville, VA

Post by sunyata » Wed Oct 03, 2012 7:35 pm

Thanks for the tip. I'll look into that.

babelproofreader
Roundtable Knight
Roundtable Knight
Posts: 138
Joined: Wed Nov 10, 2004 4:36 pm

Post by babelproofreader » Thu Oct 04, 2012 6:55 am

You might also find the ideas here useful.

Post Reply