fbpx

You have from from hundreds to trillions of potential parameter sets to choose from. How do you choose a set or sets to go to production with?

(Last Updated On: August 28, 2011)

You have from from hundreds to trillions of potential parameter sets to choose from. How do you choose a set or sets to go to production with?

What metrics do you use and how do you apply them? The industry at large has the Sharpe Ratio, Han is looking at Entropy and I have progressed from my old Q ratio to a variation of the Omega ratio. What else is out there?

—-

,

Are you speaking from the perspective of a genetic optimization fitness function? As an aside, I am not stalking you; we just seem to follow the same threads.

—–
Genetic optimization functions are fine, but to me the first question is what metric(s) to optimize on.

Most of big boys are simultaneously running multiple parameter sets in each model, but still the first question is what metrics are we using to create those sets.

I fell in love with cumulative frequency ogives of return streams and drawdowns etc year ago, but all the packaged systems seem to offer are things like Arithmetic Return (Profit) over Worst Draw (in $).

Of course the thread can easily go off in many directions and I welcome that too.

—–

For my evaluation metric in genetic opt, I use the area under the curve of the receiver operator curve (AUC of ROC). It’s a pretty good metric of how much your model is actually differentiating.

Great article:
http://matlabdatamining.blogspot.com/2007/06/roc-curves-and-auc.html

 

NOTE I now post my TRADING ALERTS into my personal FACEBOOK ACCOUNT and TWITTER. Don't worry as I don't post stupid cat videos or what I eat!
Don't miss out!
Get my TRADE LIKE A BOSS 2 PDF Books

You will received instantly the download links.

Invalid email address
Give it a try. You can unsubscribe at any time.

NOTE!

Check NEW site on stock forex and ETF analysis and automation

Scroll to Top