Momentum Testing Equity Jump

[RobGiles]

Still testing various momentum approaches on the ASX. Have been testing from June ’07 to Sept ’18 on the back of Nick’s advice to select a period where the system has a chance to perform, but is also subject to some challenges. Got the following stats:
CAR 19.26%
MaxDD -30.61%
CAR/MaxDD 0.63
Profit Factor 2.17

Whilst performance was fine, I was unhappy with the max DD I was getting so tested using Julian’s ADX(dur) > 30 (on the individual stock, not the Index). This changed the performance stats to:

CAR 14.39%
MaxDD -22.41%
CAR/MaxDD 0.64
Profit Factor 1.97
The thing that I can’t get my head around is the large jump in equity (i.e. a vertical jump) in June 2013. Looking through the trade data, there was a 869% profit on 1 stock (FGX). Have others seen a large impact from 1 stock like this in their testing and is it something to be wary of?

Any other ideas on how to keep MaxDD at around -20% but kick the CAR would be appreciated.

[Nick Radge]

Run a backtest zoomed into that period, then sort the trade list and look at the highest profit trades. Should stand out.

[RobGiles]

So as it turns out, liquidity filters matter on large universes (this system was tested on the All Ords). The 55.7% equity jump in June 2013 was due to the b-test taking a position in a 6c stock with a daily volume of only 2.6M shares. I would have added a significant amount to the daily volume so the b-test is unrealistic. only problem is liquidity filter kills the system so back to the drawing board. Definitely hasn’t been a case of use my rules for my existing US MOM system and simply apply them to the ASX.

[Nick Radge]

Try adding that symbol to the Exclude List and run the test again. See if any others stand out. If not, might be okay.

In terms of liquidity with longer term systems you may be able to ease into a position over the course of a day or two. I do that with my Growth Portfolio positions sometimes.

[Author]

Posts