WFA vs Alternate Data

Strategies might be attractive on the surface but fail moving forward. Why? Here we discuss the qualities of effective trading strategies, and the many traps one should avoid.

WFA vs Alternate Data

Postby rjay » Sun Sep 11, 2011 12:04 pm

Currently I don't use Walk Forward Analysis: I just include the out-of-sample data as an Alternate Data source and reject any strategy that is less than 75% of the APR and Drawdown figures for the in-sample data.

Now that I am going to use WFA, is there any point in my including the OOS data as an Alternate Data filter ? If I don't include it, then I'll end up with a lot more strategies so it will be more work to decide which ones to do a WFA on, but aren't I essentially repeating myself if I do both Alternate Data scans and a WFA on the same OOS data ?

Hope that makes sense ....
rjay
 
Posts: 116
Joined: Wed Jul 26, 2006 6:51 am

Re: WFA vs Alternate Data

Postby Overload » Mon Sep 12, 2011 8:11 am

I think it depends how you set it up. But a Walk-Forward Analysis on alternate data is still different than a single test on alternate data. In a WFA, you are testing different parameter sets, while in a single alternate data test you are testing just one for the entire period.

In addition, you should take not of the checkbox on the OneClick Setups, Dynamic Strategies tab, named "Force Alternate Data Evaluations to Use Same Evaluation Periods as Main Search". When checked, the alternate data tests will use the same parameter sets AND out-of-sample time intervals as the underlying search itself. This is much different than a single out-of-sample test.

Pete
Overload
 
Posts: 2246
Joined: Wed Nov 30, 2005 12:14 pm


Return to Curve-Fitting and Other Pitfalls

cron