In my quest to develop a better method of generating returns, I am making a change to my monthly model.  I am replacing the paired index model with a style/sector model.  This model will also hold only two positions like the paired index model, but it will have access to a broader array of instruments including bond ETFs as well as international ETFs.  The stop on this model will be 10% and it will comprise 30% of my overall model, which leaves total portfolio risk from this model at 3%.

Since this monthly model is so simple and it only updates monthly, I wanted a piece that could withstand heightened volatility.  I have also lengthened the lookback periods in this model to enhance its ability to hold winning positions longer.  This model will also not hold any inverse positions, and will also hold more individual sectors such as financials, energy, etc.  In combination with the sector model, this will allow overweighting to sectors that may be outperforming.  Backtesting has produced the following returns:

Year	Return
2002	3.25%
2003	29.70%
2004	5.52%
2005	21.84%
2006	12.62%
2007	-3.38%
2008	-9.35%
Thru July 2009	26.69%
Annualized	14.04%
Std Dev	15.81%

Notice how the model held up very well in the poor market years of 2002 and 2008, and is up over 26% so far in 2009.  Again, this is with no inverse positions.  Implementation of this new model means that there is one change for the overall model.  Buy 15% EEM in place of 15% DIA.  The QQQQ position remains the same.