In a recent post I previewed a neat service coming soon from my colleague Jeff Pietsch that ranks various indicators by their rolling 252 period Sortino ratio.
I raised the possibility of being able use these rolling Sortinos to switch between systems or manage position size.
It’s taken a while to get the models built but they are showing some initial promise.
The Rolling Sortinos
Despite sounding like a traveling circus act, this metric does appear to have some use as a filtering tool.
For the first study I’ve turned to the DVB indicator once again. As I’ve blogged about recently, this formerly standout indicator has not only stopped working since August, it has actually become profitable to trade the opposite rules.
The equity curve below tells its own story.
When we view the rolling 252 period Sortino ratio, we can see that its power has been on the wane since the heady days of 2008 and after a late return to form in 2010, the trading the indicator using the basic settings has been in negative territory for some time.
Could have used the Sortino ratio to forewarn us
I believe so yes, but there are many ways this could have been done such as comparing the Sortino or looking at whether the Sortino is dropping or rising etc.
The one method I have explored with good results is to measure the 50 period rate of change in the 252 rolling sortino then bound this by performing a percent ranking of this rate of change.
The Sortino Rolling Rate of Change Rank
As a summary here is the methodology:
- Trade the DVB indicator long below 0.5, short above 0.5.
- Take the rolling 252 period Sortino of the compound equity curve.
- Calculate the 50 period rate of change of this Sortino.
- Make a 252 period percent rank of this rate of change.
The idea is to capture periods when the Sortino is not only dropping, but dropping quickly.
Filter:
- Trade the method when the rolling rate of change rank is 0.25 or above and the rolling Sortino is above 1.
In Parallel I performed the filter on the DVB with the opposite basic rules (Buy above 0.5 sell below 0.5). This way the filter will switch you between mean reversion and trend following.
Results:
We don’t get a super smooth equity curve, but in reality you can never get this with a system filter without the benefit of hind sight bias. What we do get is a much improved risk reward.
The basic DVB method now has a 42% max draw down with the 250 day CAGR at -18%. By contrast the Sortino ROC rank filter (which also switches on the opposite rules) has an 18% 250 day CAGR. Of more interest is the long term MAR (CAGR compared to the draw down) which is 1.52 for the filtered method and 0.62 for the basic method.
There is certainly potential with this.
