Was in the shower a moment ago and was thinking on how to enhance a typical trend following strategy.
Your typical Trend system usually is designed to catch the big moves in the market; this is pretty much the essence of the trend following philosophy. One reason many people feel uncomfortable with trend following is the low win rate. A typical system will usually have a 40% average win rate; if you reveal this to the typical investor… they will just walk away before you even get to explain yourself.
Lets take a step back and analyze the other 60% of the time when your typical trend following system looses. Also lets suppose we are using the typical dual moving average cross over as our system. If you are a trader you will probably know that moving averages have absolutely no predictive value but thats not my point here. Just like any other trend following systems, there are times when the market consolidates which usually brings forth a string of losses. These usually are reflected by the drawdowns shown in the backtests (ie look at post “Developmental Backtest, January 29th, 2011).
Drawdowns. This is the main thing that has been nagging me lately (also during my showers). I understand that risk and reward is inversely correlated and implementing more filters will simply decrease my return %. (I currently have a system that is giving me 50% draw in backtest, so I am guessing it will turn to 60% if I bring it live)
One idea I have come up with is to implement mean reversion strategies in intraday time frame. This is to complement my core trend following systems so as to smooth returns. My hypothesis is that if I implement a trend following strategy plus another mean reversion strategy I will be able to reduce draw down and smooth out my returns. My major assumption here is that these two strategies have very low correlations as I would be double betting on a particular asset if they were highly correlated.
I still putting ideas on the drawing board but I am hoping to engineer a system that is able to produce returns that are…
1. minimal risk/drawdown
2. uncorrelated with market performance
Anyways, mechanica is on its way to my doors but I got 3 exams this week…(love/hate)