Quantopian's community platform is shutting down. Please read this post for more information and download your code.
Back to Community
Question about stationarity

Hi,

First of all, thank you Delaney Granizo-Mackenzie and Maxwell Margenot for your excellent lecture on Integration, Cointegration, and Stationarity. I am an old floor trader with almost no math background...and I'm trying to pair trade futures contracts intraday . I do well for a while but then get blown out when the spread suddenly moves to a new level and I'm caught with the wrong position.

I believe my problem is my time series moves from a state of stationarity to non-stationarity.

My question is: Can I anticipate this move before it happens with the suggested Dickey-Fuller test........or will said test just tell me after the fact what just happened as with many other technical indicators.

I could hire a programmer to set up the test so that it is running in the background, in real time ,as I trade if it has any predictive qualities.

If this isn't the answer to keep me out of the market during these shifts....can anyone suggest something else?

Thank you for your patience,
mitch