"How should I handle outliers in my backtest?"
By Dave Mabe
Several traders replied to my email yesterday about handling bad data, asking about how to handle outliers that aren't the result of bad data.
Outliers are something that should keep you up at night a little.
Here are some quick thoughts on outliers:
Make sure they are true outliers and not just bad data
Are the outliers all in your direction? (i.e. highly profitable trades) If so, that's better.
If they're all in your direction, is there still a chance of big losing outliers that haven't appeared yet? Or is there something fundamental about your strategy that avoids them? (Which is good!)
If there are big losing outliers, then it's time to pause and give it more thought. Is there a way to limit them?
These are important questions because they're often the reason you have trouble using larger size with your strategy.
One of the first questions I ask a new trader I coach is:
Are you trading your strategy with full size? If not, why not?
There are usually good, healthy reasons why they're not.
Sometimes the reason is fear of the potential for a large losing trade - an outlier.
So understanding outliers and gathering more data about them can unlock a path to confidence - and trading with larger size.
Look for more on this topic in the coming days...
-Dave
P.S. Do you wish you had a column library that would tell YOU how to make your strategy profitable? My column library is now included with MabeKit.