This is my first post, so please forgive if it resembles a run-on sentence... If my delivery seems familiar, you might know me by my blog, MKTSTK.com. If so, you know I generally DGAF about dropping some truly valuable information on the community just for fun, so without further ado...
Today I wanted to present a little strategy that's been causing some buzz amongst the prop traders and hedge fundos I've been talking to so far. In case you were wondering if Quantopian gave you an edge, this should be proof positive of their commitment to attracting top-notch datasets. Please note that although this strat uses the fetcher API, soon you should be able to use Psychsignal's datasets I've developed directly in Quantopian in addition to the PsychSignal Trader Mood API which is already available for free for backtesting and live trading.
The strategy is driven by a daily datafeed I created called the HIVE-MIND. The Hive, as its called for brevity's sake, is made of two distinct components: 1) the Hive-Bot and 2) the Hive-Net.
The Hive-Bot measures the activity of a symbol with respect to the social media landscape. The Hive-Bot transforms the multidimensional social message flow into a simple scale between 0 and 1.0, called the Social Anomaly Score (SAS). At the high end, 1.0 represents a frenzy level of activity related to a symbol. In the middle, 0.5 is meant to signify a normal social pattern (i.e. what is expected given the historical profile of the symbol over time). At the other extreme, a reading near 0.0 represents a low amount of interest.
My research has shown that the Hive-Bot's SAS is predictive of volatility and correlation. Thus it makes sense to use it as a market timing mechanism. There are many possible forms for this to take within a real trading strategy. One such conceivable usage is to switch between mean reverting and momentum strategies. Despite many idiosyncrasies, trading strategies often break-down into simplistic categories of being levered to momentum or mean reversion. If one could differentiate, a priori, between mean reverting and momentum periods in the market one could make a fortune... but how might you construct such a strategy?
The following presents a model which combines both mean reverting and momentum based strategies. While it is not levered by default, the strategy can choose to employ leverage when it is advantageous. The strategy uses the Hive-Bot's SAS to sense when to switch between mean reverting and momentum regimes. The strategy uses two different look-back windows to gauge momentum and trades around 50 equity ETF's.
Since I'm new to Q, any feedback would be much appreciated. Happy Trading
|Returns||1 Month||3 Month||6 Month||12 Month|
|Alpha||1 Month||3 Month||6 Month||12 Month|
|Beta||1 Month||3 Month||6 Month||12 Month|
|Sharpe||1 Month||3 Month||6 Month||12 Month|
|Sortino||1 Month||3 Month||6 Month||12 Month|
|Volatility||1 Month||3 Month||6 Month||12 Month|
|Max Drawdown||1 Month||3 Month||6 Month||12 Month|